what to do with DSD ? [Archive]

Scott Dorsey

April 12th 07, 11:58 PM

Harry Lavo > wrote:
>
>Don't get discouraged by the "PCM uber alles" posturing of the Arny Kruegers
>of the world. Michael Bishop revealed in another forum the other day that
>he recorded using the Korg as a backup during a session, and was astounded
>at the quality of the 128Fs DSD. His only complaint was the usual one for
>lower cost gear...the preamps were a bit noisy...but using the line in from
>his normal high quality outboard pre's took care of that.

The truth is, though, that PCM can sound very good as well, and we are now
at the level where the recording system itself is no longer a serious
bottleneck in the recording chain. This is a wonderful place to be and it
took a more than a century to get here.

I have heard some wonderful sounding DSD recordings, but the need to convert
to a PCM representation to do anything with the DSD datastream is kind of a
big issue.

>I'd look at it this way...you may have to convert to high bit rate PCM to do
>anything practical with it now, but you've got an extreme high quality
>archive, and if DSD continues to catch on for that purpose, soon their will
>be other options for playback, possibility including PCI cards, outboard
>converters, or units like the Teac that can serve as external decoders.
>Then you've got the best of both worlds.

This is true. And it _is_ easier to make good DSD converters than good
PCM converters right now. This is a minor plus on the DSD side. But,
a typical sigma-delta converter can be looked on as a DSD converter with
additional logic to convert the DSD datastream to PCM... so in fact splitting
those into two seperate units and saving the DSD part may not be such an
improvement anyway.

The folks who are saying SACD is dead in the water and a dying format are
probably right, but for that matter you could say the same thing about
standard PCM formats like the compact disk. The mass market doesn't want
quality. The question is whether SACD and CD can survive as niche players
in the future, the way LP has hung on. I can't answer that one.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

William Sommerwerck

April 13th 07, 12:15 AM

> ...the consumer implementation, SACD, was not a success
> in the market.

How do you define "success"? I have over 200 SACDs.

ChristopheRonald

April 13th 07, 03:18 AM

wow arny. your a asshat. your soo nice in your responses (sarcasm
intended)

On Apr 12, 7:31 am, "Arny Krueger" > wrote:
> "Nick's Picks" > wrote in message
>
> ups.com
>
> > I'm in need of direction.
>
> I guess.
>
> > I recently started mastering with the new Korg DSD
> > recorders.
>
> Congratulations on buying a recorder for which almost nobody has a player to
> match.
>
> > They come bundled w/software for converting
> > the .dff files to PCM at any resolution.
>
> Begging the question why not just use a PCNM recorder.
>
> > now, what my question is...., how can I build a library
> > of the masters that I can actually listen to?
>
> Listen to them on the Korg.
>
> > I've got a good reference system, and playing the masters
> > off of my recording deck is stunning...yet not very
> > practical.
>
> As you say.
>
> > Is there a computer based solution? Some sort of
> > soundcard / software that will play these files and send
> > a HQ analog signal out?
>
> Why not just convert them to PCM?
>
> Of course that begs the question, why not just record in PCM in the first
> place?
>
> > I was thinking perhaps one of the Tascam rack mount jobs
> > (the dvra1000 that does DSD), and w/the hard drive
> > option, using that as my "playback deck". Might be the
> > cheapest solution?
>
> Supposedly the Korg cost you about $1200.
>
> The Tascam runs about $1300.
>
> Intesesting economic theory you seem to be pursuing there...

Arny Krueger

April 13th 07, 12:02 PM

"William Sommerwerck" > wrote in
message

>> ...the consumer implementation, SACD, was not a success
>> in the market.

> How do you define "success"?

More to the point, it is helpful to be able to define failure.

The whole high-resolution audio distribution format initiative of around
Y2K, including both SACD and DVD-A, was simply ineffective.

(1) There was never any reliable real-world evidence that the two hi-rez
actually and practically enhanced sound quality. If they were as good as was
claimed it should be excruciatingly easy to do a good unbiased listening
test showing a distinct advantage over the CD format. Lots of people tried,
but it never worked out. Been there, done that.

(2) The so-called hi-rez media and equipment languished in the mainstream
retail channels. Again, if it really sounded that much better, it should
have flown off of retail store shelves. I daresay, were there not an
internet and its ability to foster niche products, both mediums would be
even far deader than they are today.

(3) Some recording industry execuitives who bet their careers on so-called
hi rez distribution media, had short tenures.

(4) The sales of both SACD and DVD-A equipment and media never really made
any dent in the mainstream, and both formats are declining rapidly. At one
point when SACD sales (arguably the more popular of the two) was said by its
advocates to be peaking, it was still being outsold by vinyl.

(5) It is readily demonstrable that the primary limits to the real-world
dynamic range and bandpass of audio recordings is not PCM recording formats.
Furthermore, even the now-ancient CD format is not the primary limit to the
dynamic range of musical recordings.

(6) The market's wide acceptance of newer distribution formats that vastly
underperform the admitted limitations of the CD format shows that for all
practical purposes, it is a technical overkill format for most music lovers.

> I have over 200 SACDs.

I have about a dozen SACDs and DVD-As, and a player for them. Nothing I
heard made me want to invest any further.

Arny Krueger

April 13th 07, 12:04 PM

"hank alrich" > wrote in message

> Arny Krueger > wrote:
>
>> Supposedly the Korg cost you about $1200.
>
> Depends which one: $699
>
> http://www.midi-store.com/KORG-MR-1-DIGITAL-RECORDER-p-16691.html

Interesting. I'll keep my Microtrack.

Arny Krueger

April 13th 07, 12:07 PM

"Harry Lavo" > wrote in message

> "Nick's Picks" > wrote in message
> oups.com...
>> Of you could buy a Sonoma system and make
>>> SACDs.
>>
>> Sonoma system eh?
>> time to google...
>> thank you Mike.
>>
>
> Don't get discouraged by the "PCM uber alles" posturing
> of the Arny Kruegers of the world. Michael Bishop
> revealed in another forum the other day that he recorded
> using the Korg as a backup during a session, and was
> astounded at the quality of the 128Fs DSD.

As in:

http://www.audioasylum.com/scripts/t.pl?f=hirez&m=234367

Clearly one of those sighted, non-time-synched, non-level-matched
evaluations that you seem to find to be definitive, Harry. Why is it that
experimental rigor and real-world recording and playback techniques (No MRI)
is the natural enemy of so-called hi-rez formats?

Arny Krueger

April 13th 07, 12:09 PM

"Mike Rivers" > wrote in message
oups.com
> On Apr 12, 5:34 pm, "Harry Lavo" > wrote:
>
>> I'd look at it this way...you may have to convert to
>> high bit rate PCM to do anything practical with it now,
>> but you've got an extreme high quality archive, and if
>> DSD continues to catch on for that purpose, soon their
>> will be other options for playback

> DSD was originally implemented by Sony as a high quality
> means of archiving their masters using an A/D converter
> that was relatively easy to build and didn't have the
> hardware problems of a 196 kHz PCM converter.

One practical problem with Sony's DSD strategy is that the so-called
hardware problems of 196 KHz PCM converters were solved so thorougly that
they can even be found in $39.95 DVD players.

Mike Rivers

April 13th 07, 12:55 PM

On Apr 13, 7:09 am, "Arny Krueger" > wrote:

> One practical problem with Sony's DSD strategy is that the so-called
> hardware problems of 196 KHz PCM converters were solved so thorougly that
> they can even be found in $39.95 DVD players.

You may not agree with Dan Lavry, but he doesn't think so. On the
other hand, he doesn't make $39.95 DVD players, so what people do with
those doesn't matter a lot fo him.

Arny Krueger

April 13th 07, 01:09 PM

"Mike Rivers" > wrote in message
ps.com

> On Apr 13, 7:09 am, "Arny Krueger" >
> wrote:

>> One practical problem with Sony's DSD strategy is that
>> the so-called hardware problems of 196 KHz PCM
>> converters were solved so thorougly that they can even
>> be found in $39.95 DVD players.

> You may not agree with Dan Lavry, but he doesn't think
> so.

I read what Dan said here about them a little differently. And, if he
thought the issues through from scratch today, he might say things a little
differently than he did the last time he was posting a lot here.

The converter market is still very dynamic. For example, it very much looks
like someone is now selling the moral equivalent of the Lynx L22 for under
$200.

> On the other hand, he doesn't make $39.95 DVD
> players, so what people do with those doesn't matter a
> lot to him.

I basically like Dan and I like what he's trying to do, but I don't see
what he is about right now as having a lot of legs.

[email protected]

April 13th 07, 01:21 PM

On Apr 13, 8:09 am, "Arny Krueger" > wrote:

> I basically like Dan and I like what he's trying to do, but I don't see
> what he is about right now as having a lot of legs.

what he says does not need legs,
it needs ears.... and a 39.95 blue light special is not about ears
is is about the wal mart get it cheap
mentality

get some good ears to go with your cheap dogma...

Scott Dorsey

April 13th 07, 02:25 PM

In article m>,
Mike Rivers > wrote:
>On Apr 13, 7:09 am, "Arny Krueger" > wrote:
>
>> One practical problem with Sony's DSD strategy is that the so-called
>> hardware problems of 196 KHz PCM converters were solved so thorougly that
>> they can even be found in $39.95 DVD players.
>
>You may not agree with Dan Lavry, but he doesn't think so. On the
>other hand, he doesn't make $39.95 DVD players, so what people do with
>those doesn't matter a lot fo him.

Well, part of the problem is that on the $39.95 DVD players, many of
the problems of early PCM systems (like clock timing errors) have been
replaced with new and different problems (like shoddy analogue sections
with noise and linearity issues).
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Harry Lavo

April 13th 07, 02:53 PM

"Chel van Gennip" > wrote in message
...
> On Fri, 13 Apr 2007 01:15:16 +0200, William Sommerwerck wrote:
>
>>> ...the consumer implementation, SACD, was not a success in the market.
>>
>> How do you define "success"? I have over 200 SACDs.
>
> I think that a market share dropping from 0.15% to 0.1% a few years after
> introduction is NOT a success in the market. And, as I wrote before, by
> now there are much better technical ways to distribute audio available.
>

I think you will find its importance in the classical music market is alive
and getting stronger. It is no coincidence that SACD is stronger in Europe
and Asia than it is in the US....sales of classical music and its
appreciation by the population in performance are also much higher in those
areas.

I can tell you that in a five-channel system consisting of first rate
electronics, five full-range Thiel speakers, and a Sony SACD player with
direct DSD decoding and no PCM bass manipulation, NOTHING comes closer to
recreating a seat in the concert hall with all that implies.

And as Brothers in Arms, Dark Side, and Love show spectacularly, hi-rez pop
can be creative and gorgeous as well.

It is just too bad the industry and media didn't get off it's ass and better
support these new formats. Let's hope the newer hi-rez video mediums can be
more successfully adopted, and that SACD continues to expand.

Harry Lavo

April 13th 07, 03:04 PM

"Arny Krueger" > wrote in message
...
> "Harry Lavo" > wrote in message
>
>> "Nick's Picks" > wrote in message
>> oups.com...
>>> Of you could buy a Sonoma system and make
>>>> SACDs.
>>>
>>> Sonoma system eh?
>>> time to google...
>>> thank you Mike.
>>>
>>
>> Don't get discouraged by the "PCM uber alles" posturing
>> of the Arny Kruegers of the world. Michael Bishop
>> revealed in another forum the other day that he recorded
>> using the Korg as a backup during a session, and was
>> astounded at the quality of the 128Fs DSD.
>
> As in:
>
> http://www.audioasylum.com/scripts/t.pl?f=hirez&m=234367
>
> Clearly one of those sighted, non-time-synched, non-level-matched
> evaluations that you seem to find to be definitive, Harry. Why is it that
> experimental rigor and real-world recording and playback techniques (No
> MRI) is the natural enemy of so-called hi-rez formats?
>

Once again, Arny, you completely discount the "ears" of a very talented
recording and production engineer....simply because he hasn't yet done a
rigorous, scientific test. Know what, Arny, most decisions aren't made that
way. But some people are experienced and critical listeners who
nevertheless seem somehow to arrive at top quality decisions without
scientific tests.

I didn't offer Michel Bishop's post as "proof" of anything; nor would he
claim it was "proof" of anything. Your committment to an anti-hi-rez stance
years ago, before you had ever heard hi-rez does you no credit, despite the
pose of asking for "science".

Arny Krueger

April 13th 07, 03:47 PM

"Harry Lavo" > wrote in message

> "Chel van Gennip" > wrote in
> message ...
>> On Fri, 13 Apr 2007 01:15:16 +0200, William Sommerwerck
>> wrote:
>>>> ...the consumer implementation, SACD, was not a
>>>> success in the market.
>>>
>>> How do you define "success"? I have over 200 SACDs.
>>
>> I think that a market share dropping from 0.15% to 0.1%
>> a few years after introduction is NOT a success in the
>> market. And, as I wrote before, by now there are much
>> better technical ways to distribute audio available.
>
> I think you will find its importance in the classical
> music market is alive and getting stronger.

Proving your claims Harry, is not something that we should be tasked with.
IOW we're not supposed to find reliable information about the importance of
Hi Rez formats to classical sales, that's your job. It's up to you to
provide evidence to support your many speculations. At this point you
obviously have no evidence to support this, or many other of your claims.

> I can tell you that in a five-channel system consisting of first rate
> electronics, five full-range Thiel speakers, and a Sony SACD player with
> direct DSD decoding and no PCM bass manipulation, NOTHING comes closer to
> recreating a seat in the concert hall with all that implies.

....so can any number of high end salesmen with storerooms full of unsold
so-called Hi-Rez players for easily-deceived consumers to buy for inflated
prices.

Your thesis Harry, which nets out to the magical idea that DSD has sonic
advantages over PCM with comparable bandwidth and dynamic range, simply
violates the laws of physics as we know them, and as we have observed them
to work for more than 50 years.

The fact of the matter is that SACD recordings have far less dynamic range
and bandwidth than even 24/96 PCM, due to the degraded SNR above 22 KHz.
Furthermore there is considerable evidence that this is a moot point, due to
the dearth of recordings that even exploit the dynamic range of the CD
format.

If the Hi-Rez formats offer that much better sound quality, it should be
absolutely trivial to demonstrate it in a unbiased evaluation.

After all, it was very easy for just about every music lover in the world to
hear that the CD format had many sonic advantages compared to LP, 8-Track,
cassette and the like. And, it is often possible to demonstrate that the
compressed formats that are coming into vogue can have reduced sound
quality. Not so with the so-called hi-rez formats.

Mike Rivers

April 13th 07, 03:48 PM

On Apr 13, 8:09 am, "Arny Krueger" > wrote:

> I read what Dan said here about them a little differently. And, if he
> thought the issues through from scratch today, he might say things a little
> differently than he did the last time he was posting a lot here.

I'm not going to fight Lavry's battle for him, but I suspect that it
would be hard to find a 192 kHz converter that sounds as good as his
96 kHz converter. SOUNDS being the operative word here. Here's a
relevant portion from his 2004 Sampling Rate paper:

===============
There are reports of better sound with higher sampling rates. No
doubt, the folks that like the "sound of a 192KHz" converter hear
something. Clearly it has nothing to do with more bandwidth: the
instruments make next to no 96KHz sound, the microphones don't respond
to it, the speakers don't produce it, and the ear can not hear it.

Moreover, we hear some reports about "some of that special quality
captured by that 192KHz is retained when down sampling to 44.1KHz.
Such reports neglect the fact that a 44.1KHz sampled material can not
contain above 22.05KHz of audio. Some claim that that 192K is closer
to the audio tape. That same tape that typically contains "only" 20KHz
of audio gets converted to digital by a 192K AD, than stripped out of
all possible content above 22KHz (down sample to CD).

"If you hear it, there is something there" is an artistic statement.
If you like it and want to use it, go ahead. But whatever you hear is
not due to energy above audio. All is contained within the "lower
band". It could be certain type of distortions that sound good to you.
Can it be that someone made a real good 192KHz device, and even after
down sampling it has fewer
distortions? Not likely. The same converter architecture can be
optimized for slower rates and with more time to process it should be
more accurate (less distortions).

The danger here is that people who hear something they like may
associate better sound with faster sampling, wider bandwidth, and
higher accuracy. This indirectly implies that lower rates are
inferior. Whatever one hears on a 192KHz system can be introduced into
a 96KHz system, and much of it into lower sampling rates. That
includes any distortions associated with 192KHz gear, much of which is
due to insufficient time to achieve the level of accuracy of slower
sampling.

Conclusion:
There is an inescapable tradeoff between faster sampling on one hand
and a loss of accuracy, increased data size and much additional
processing requirement on the other hand. AD converter designers can
not generate 20 bits at MHz speeds, yet they often utilize a circuit
yielding a few bits at MHz speeds as a step towards making many bits
at lower speeds.The compromise between speed and accuracy is a
permanent engineering and scientific reality.

Sampling audio signals at 192KHz is about 3 times faster than the
optimal rate. It compromises the accuracy which ends up as audio
distortions.
While there is no up side to operation at excessive speeds, there are
further disadvantages:

1. The increased speed causes larger amount of data (impacting data
storage and data transmission speed requirements).

2. Operating at 192KHz causes a very significant increase in the
required processing power, resulting in very costly gear and/or
further compromise in audio quality.

The optimal sample rate should be largely based on the required signal
bandwidth. Audio industry salesman have been promoting faster than
optimal rates. The promotion of such ideasis based on the fallacy that
faster rates yield more accuracy and/or more detail. Weather motivated
by profit or ignorance, the promoters, leading the industry in the
wrong direction, are stating the opposite of what is true.

Copyright Dan Lavry, Lavry Engineering, Inc, 2004
====================

The "Further disadvantages" above are probably of less concern today
than in 2004 when he wrote the paper, but the rest of it makes pretty
good sense. The whole paper is at:
http://www.lavryengineering.com/documents/Sampling_Theory.pdf

> The converter market is still very dynamic. For example, it very much looks
> like someone is now selling the moral equivalent of the Lynx L22 for under
> $200.

Who might that be? I could use another decent sound card.

Arny Krueger

April 13th 07, 03:48 PM

> wrote in message
ups.com
> On Apr 13, 8:09 am, "Arny Krueger" >
> wrote:
>
>> I basically like Dan and I like what he's trying to do,
>> but I don't see what he is about right now as having a
>> lot of legs.
>
> what he says does not need legs,
> it needs ears.... and a 39.95 blue light special is not
> about ears is is about the wal mart get it cheap
> mentality

I'm sorry that you are in denial about modern technology.

> get some good ears to go with your cheap dogma...

Been there done that - the difference is that I dare to listen without
having my eyes tell me what the *right* answer should be.

Arny Krueger

April 13th 07, 03:52 PM

"Harry Lavo" > wrote in message

> "Arny Krueger" > wrote in message
> ...
>> "Harry Lavo" > wrote in message
>>
>>> "Nick's Picks" > wrote in message
>>> oups.com...
>>>> Of you could buy a Sonoma system and make
>>>>> SACDs.
>>>>
>>>> Sonoma system eh?
>>>> time to google...
>>>> thank you Mike.
>>>>
>>>
>>> Don't get discouraged by the "PCM uber alles" posturing
>>> of the Arny Kruegers of the world. Michael Bishop
>>> revealed in another forum the other day that he recorded
>>> using the Korg as a backup during a session, and was
>>> astounded at the quality of the 128Fs DSD.

>> As in:

>> http://www.audioasylum.com/scripts/t.pl?f=hirez&m=234367

>> Clearly one of those sighted, non-time-synched,
>> non-level-matched evaluations that you seem to find to
>> be definitive, Harry. Why is it that experimental rigor
>> and real-world recording and playback techniques (No
>> MRI) is the natural enemy of so-called hi-rez formats?

> Once again, Arny, you completely discount the "ears" of a
> very talented recording and production engineer....simply
> because he hasn't yet done a rigorous, scientific test.

Pardon me for suggesting that he's susceptible to human bias.

> Know what, Arny, most decisions aren't made that way.

Most decisions in recording involve differences that are obviously audible
when put to a similar test.

> But some people are experienced and critical listeners
> who nevertheless seem somehow to arrive at top quality
> decisions without scientific tests.

That's because most decisions in recording relate to differences that when
measured, turn out to be well above JNDs.

> I didn't offer Michel Bishop's post as "proof" of
> anything; nor would he claim it was "proof" of anything.

I don't think that Bishop needs you to speak for him, Harry.

> Your committment to an anti-hi-rez stance years ago,
> before you had ever heard hi-rez does you no credit,

That would be a false claim, Harry. One that it seems impossible to correct
you about. It's part of your dogma - everybody who disagrees with you can't
or won't hear.

> despite the pose of asking for "science".

No pose Harry, just how things are around here.

William Sommerwerck

April 13th 07, 04:41 PM

>> I can tell you that in a five-channel system consisting of first rate
>> electronics, five full-range Thiel speakers, and a Sony SACD player
>> with direct DSD decoding and no PCM bass manipulation, NOTHING
>> comes closer to recreating a seat in the concert hall with all that
implies.

> ...so can any number of high end salesmen with storerooms full of unsold
> so-called Hi-Rez players for easily-deceived consumers to buy for inflated
> prices.

I paid about $2000 for my multi-ch Sony SACD player. (That was a discount
price from Oade Brothers.) Considering its build quality, and apparent sound
quality, I don't consider that unreasonable or "inflated".

By the way, SACD and DVD-A players give you something you don't (easily) get
from CD -- surround sound.

> Your thesis Harry, which nets out to the magical idea that DSD has
> sonic advantages over PCM with comparable bandwidth and dynamic
> range, simply violates the laws of physics as we know them, and as
> we have observed them to work for more than 50 years.

WHAT laws of physics do we "know", Arny? That if product A has measured
specs comparable to product B, that they necessarily sound alike (or very
nearly so), EVEN THOUGH THOSE SPECS HAVE NEVER BEEN CORRELATED WITH
LISTENING QUALITY?

We don't listen to specifications. Nor are bandwidth and dynamic range
necessarily the only or principal things we perceive when listening to
reproduced sound.

I don't understand how someone who appears to be intelligent and
well-educated can believe that controlled (and likely over-controlled)
listening tests that bear little relationship to the way people actually
listen are the last word on sound differences or quality. They aren't, and
never were.

When someone is willing to run long-term blind or double-blind listening
tests, in which listeners are free to listen in the usual way, when and how
they feel like it, without being forced to sit down and make comparisons or
decisions, then I will start taking listening tests seriously. But that will
never happen, not only because it's difficult to do properly, but because
both sides in this "issue" are desperately convinced that such tests will
disprove what they "know", beyond any shadow of doubt, to be true.

William Sommerwerck

April 13th 07, 04:54 PM

> Been there done that - the difference is that I dare to listen without
> having my eyes tell me what the *right* answer should be.

No, you have your "science" telling your ears what they ought to hear.

"The proof of the pudding is in the eating." Arny, if you know so much...
Tell me why the Crown K1 amplifier -- which has excellent specs, and was
designed by one of the "industry greats", Gerald Stanley -- sounds so
utterly awful. Then tell me why a Parasound A21 sounds so completely
different from a Krell KSA-250.

I make the following claim, and I'd be willing to prove it... Given
recordings I'm familiar with, I will tell you which of these amplifiers is
playing, without having to compare it with any of the others. That's how
grossly different they sound.

I do not believe in magic. I do not believe the reasons for sonic
differences are incomprehensible or unmeasurable. I merely claim that nobody
has done much to legitimately qualify and quantify such differences. All we
get from Arny, et al., is if there are no measurable differences, then there
are no sonic differences -- which is plainly intellectual bilge.

I want Mr. Arny "I know the secrets of the universe" Krueger to tell me what
the sonically significant measurements are, and _why_ they correlate with
one's subjective reactions.

Scott Dorsey

April 13th 07, 05:59 PM

In article >,
William Sommerwerck > wrote:
>>> I can tell you that in a five-channel system consisting of first rate
>>> electronics, five full-range Thiel speakers, and a Sony SACD player
>>> with direct DSD decoding and no PCM bass manipulation, NOTHING
>>> comes closer to recreating a seat in the concert hall with all that
>implies.
>
>> ...so can any number of high end salesmen with storerooms full of unsold
>> so-called Hi-Rez players for easily-deceived consumers to buy for inflated
>> prices.
>
>I paid about $2000 for my multi-ch Sony SACD player. (That was a discount
>price from Oade Brothers.) Considering its build quality, and apparent sound
>quality, I don't consider that unreasonable or "inflated".
>
>By the way, SACD and DVD-A players give you something you don't (easily) get
>from CD -- surround sound.

There was actually a four-channel standard format built into the Red Book.
As far as I know, it was never actually implemented in a single product. But
it would have been nice to have seen anyway.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Arny Krueger

April 13th 07, 06:10 PM

"William Sommerwerck" > wrote in
message

>> Your thesis Harry, which nets out to the magical idea
>> that DSD has
>> sonic advantages over PCM with comparable bandwidth and
>> dynamic
>> range, simply violates the laws of physics as we know
>> them, and as
>> we have observed them to work for more than 50 years.

> WHAT laws of physics do we "know", Arny?

If you have to ask... ;-(

> That if product
> A has measured specs comparable to product B, that they
> necessarily sound alike (or very nearly so), EVEN THOUGH
> THOSE SPECS HAVE NEVER BEEN CORRELATED WITH LISTENING
> QUALITY?

Straw man argument.

In fact, *all* of the things that are commonly specified for audio gear are
known to have some correlation with listening quality.

For example, one common spec is frequency response. Now seriously, are you
going to assert that there is no correlation between good frequency response
and listening quality? I would seriously hope not!

> We don't listen to specifications.

Where did I say that we did listen to specifications? Everybody with a
brain knows that specifications are abstractions of things that are known to
cause sonic difficulties.

> Nor are bandwidth and
> dynamic range necessarily the only or principal things we
> perceive when listening to reproduced sound.

In fact bandwidth and dynamic range summarize everything that we hear wrong,
when we hear something wrong. They include every kind of linear and
nonlinear distortion, in the sense that any known kind of linear or
nonlinear distortion degrades bandwidth and/or dynamic range.

> I don't understand how someone who appears to be
> intelligent and well-educated can believe that controlled
> (and likely over-controlled) listening tests that bear
> little relationship to the way people actually listen are
> the last word on sound differences or quality. They
> aren't, and never were.

That's another myth. In fact most real-world listening is done blind. I
don't do a visual inspection of the entire record/reproduce chain when I
listen to music. Nobody does. In fact much of it is unknowable most of the
time.

Furthermore, the sound quality of audio chains tend to be dominated by the
weakest links. These days, the weakest links are likely to not be the
recorded formats used for production or distribution.

Finally, many parts of the audio chain that tend to dominate SQ are things
whose usuage details tends to dominate over the inherent properties of the
device. For example, I can know the on-axis frequency response of a certain
microphone at a certain distance, but I also know that the mic's frequency
response is highly dependent on distance and angle, which I geneally can't
very easily know for other people's recordings.

Therefore when I listen to other people's recordings, it is mostly a blind
evaluation.

Similarly, most people listen mostly blind most of the time.

That's one very strong reason why blind tests are the ones that provide the
results that correlate best with the broadest-based experience. To wit,
blind tests show no audible benefits for so-called Hi-rez formats, and
Hi-rez formats basically tanked in the marketplace.

> When someone is willing to run long-term blind or
> double-blind listening tests, in which listeners are free
> to listen in the usual way, when and how they feel like
> it, without being forced to sit down and make comparisons
> or decisions, then I will start taking listening tests
> seriously.

That all happened decades ago. No joy. It happened again decades later,
still no joy.

> But that will never happen, not only because
> it's difficult to do properly, but because both sides in
> this "issue" are desperately convinced that such tests
> will disprove what they "know", beyond any shadow of
> doubt, to be true.

Wrong every way possible.

The people who are most desperately convinced that they already know the
right answers and don't need any more stinkin' tests to convince them of
anything, are rather logically and obviously the people who have done the
fewest tests.

The people with the most open minds are the people who did and continute to
put their beliefs (which were not as one-sided as many would like to
believe) to the test the most often and the most recently.

Arny Krueger

April 13th 07, 06:18 PM

"William Sommerwerck" > wrote in
message
>> Been there done that - the difference is that I dare to
>> listen without having my eyes tell me what the *right*
>> answer should be.
>
> No, you have your "science" telling your ears what they
> ought to hear.

Doesn't matter. Besides, enough listening is done by people who are ignorant
of the science so that if there was some hidden truth, it would out.

> "The proof of the pudding is in the eating." Arny, if you
> know so much... Tell me why the Crown K1 amplifier --
> which has excellent specs, and was designed by one of the
> "industry greats", Gerald Stanley -- sounds so utterly
> awful.

Never heard one, never tested one on the bench. I plead ignornace.

> Then tell me why a Parasound A21 sounds so
> completely different from a Krell KSA-250.

Same basic answer.

> I make the following claim, and I'd be willing to prove
> it... Given recordings I'm familiar with, I will tell you
> which of these amplifiers is playing, without having to
> compare it with any of the others. That's how grossly
> different they sound.

This makes no difference to me. I imagine that if it were convenient enough,
I'd put forth the effort to do a proper set of listening tests and
measurements and try to explain the correlation between measurements and
listening tests. But, sighted casual evaluations are so fraught with
difficulites, I won't even try if they are all there is.

> I do not believe in magic.

Do you believe that SACD sounds better or even different from PCM, given the
same or at least similar dynamic range and bandwidth?

> I do not believe the reasons
> for sonic differences are incomprehensible or
> unmeasurable.

I'm unconvinced.

>I merely claim that nobody has done much to
> legitimately qualify and quantify such differences.

There's a legion of serious researchers that would disagree with that.

> All we get from Arny, et al., is if there are no measurable
> differences, then there are no sonic differences -- which
> is plainly intellectual bilge.

No, what we find repeatedly is that if there are no audible differences,
then the measurements agree that there should be no audible differences.

> I want Mr. Arny "I know the secrets of the universe"
> Krueger to tell me what the sonically significant
> measurements are, and _why_ they correlate with one's
> subjective reactions.

My idea of a complete set of measurement of a power amp is here:

http://www.pcavtech.com/pwramp/macrot-5000VZ/index.htm

Here are some brief audio samples that demonstrate some of the audible
differences that can be attributed to this amp:

http://www.pcabx.com/product/macrot-5000VZ/index.htm

[email protected]

April 13th 07, 06:52 PM

On Apr 13, 10:48 am, "Arny Krueger" > wrote:
> > wrote in message
>
> ups.com
>
> > On Apr 13, 8:09 am, "Arny Krueger" >
> > wrote:
>
> >> I basically like Dan and I like what he's trying to do,
> >> but I don't see what he is about right now as having a
> >> lot of legs.
>
> > what he says does not need legs,
> > it needs ears.... and a 39.95 blue light special is not
> > about ears is is about the wal mart get it cheap
> > mentality
>
> I'm sorry that you are in denial about modern technology.
>
> > get some good ears to go with your cheap dogma...
>
> Been there done that - the difference is that I dare to listen without
> having my eyes tell me what the *right* answer should be.

yes you do remind me of the sales people at the big box stores...
"look at the specs.... ya can't hear that god damn good ..."

but you can hear that good!!!

but then listening is an art...

and it requires practice... too

< http://www.tnt-audio.com/topics/realstereo_e.html >

Arny Krueger

April 13th 07, 07:24 PM

> wrote in message
ups.com
> On Apr 13, 10:48 am, "Arny Krueger" >
> wrote:
>> > wrote in message
>>
>> ups.com
>>
>>> On Apr 13, 8:09 am, "Arny Krueger" >
>>> wrote:
>>
>>>> I basically like Dan and I like what he's trying to
>>>> do, but I don't see what he is about right now as
>>>> having a lot of legs.
>>
>>> what he says does not need legs,
>>> it needs ears.... and a 39.95 blue light special is not
>>> about ears is is about the wal mart get it cheap
>>> mentality
>>
>> I'm sorry that you are in denial about modern technology.
>>
>>> get some good ears to go with your cheap dogma...

>> Been there done that - the difference is that I dare to
>> listen without having my eyes tell me what the *right*
>> answer should be.

> yes you do remind me of the sales people at the big box
> stores... "look at the specs.... ya can't hear that god
> damn good ..."

You're quoting me saying something that I never said. So much for the
truth.

> but you can hear that good!!!

I do a lot of recording and live sound. Both are arts that are rather
dependent on one's ability to listen good sound into existence.

> but then listening is an art...

You're preaching to the choir.

> and it requires practice... too

How many hours a week do you listen to live music with a console at hand?
I'm setting no records, but for me it averages about 6-8 hours a week. IOW,
your negative comments about my lack of interest in listening are completely
bogus inventions of your mind.

William Sommerwerck

April 13th 07, 07:38 PM

> There was actually a four-channel standard format built into
> the Red Book. As far as I know, it was never actually
> implemented in a single product. But it would have been nice
> to have seen anyway.

Yes. It required double-speed playback, which halved the recording time.
This, and the fact the disks were incompatible with stereo-only playback,
meant that no quad CDs were produced.

Les Cargill

April 13th 07, 07:40 PM

William Sommerwerck wrote:

> It's possible a DSD recording converted to PCM will be a more accurate
> rendition of the original than one made directly in PCM. But I don't know
> whether anyone has tried to prove or disprove this.
>
> And (as AK will be quick to point out), PCM is subjectively perfect, anyway,
> so what difference does it make?
>
>

If there's a difference, how do we account for it?

The Nyquist Theorem limits of PCM are well understood - which
constraint is improved on by DSD?

All Lavry is restating is that the Shannon Channel
Theorem is still holding sway, and that the additional
bandwidth can be a Problem.

Has anybody A/B'd a DSD signal with it's PCM copy
in a vetted test yet? :)

--
Les Cargill

Arny Krueger

April 13th 07, 07:44 PM

"Les Cargill" > wrote in message

> William Sommerwerck wrote:
>
>> It's possible a DSD recording converted to PCM will be a
>> more accurate rendition of the original than one made
>> directly in PCM. But I don't know whether anyone has
>> tried to prove or disprove this. And (as AK will be quick to point out),
>> PCM is
>> subjectively perfect, anyway, so what difference does it
>> make?

> If there's a difference, how do we account for it?

Tough question.

> The Nyquist Theorem limits of PCM are well understood -
> which constraint is improved on by DSD?

None.

> All Lavry is restating is that the Shannon Channel
> Theorem is still holding sway, and that the additional
> bandwidth can be a Problem.

If only a waste of resources.

> Has anybody A/B'd a DSD signal with it's PCM copy
> in a vetted test yet? :)

If there were any clear positive results, the DSD advocates would be
shouting it from the rooftops.

The vendors have tried to keep the naive consumers all a-twitter by
remastering just about every SACD release.

Scott Dorsey

April 13th 07, 07:55 PM

Les Cargill > wrote:
>
>If there's a difference, how do we account for it?

If there is a difference, it's almost certainly due to implementation
issues and not to the theory itself. Don't put that down, though...
implementation issues are _important_.

>The Nyquist Theorem limits of PCM are well understood - which
>constraint is improved on by DSD?

None, but the performance of real-world converters may not equal that
of a theoretically perfect converter.

>All Lavry is restating is that the Shannon Channel
>Theorem is still holding sway, and that the additional
>bandwidth can be a Problem.

That's true. BUT, the good news about DSD is that it has power
density issues that make it hard to do abusive processing and get
away with it. That's a huge plus in my book.

>Has anybody A/B'd a DSD signal with it's PCM copy
>in a vetted test yet? :)

Not one that I would trust yet.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

William Sommerwerck

April 13th 07, 08:05 PM

>> That if product
>> A has measured specs comparable to product B, that they
>> necessarily sound alike (or very nearly so), EVEN THOUGH
>> THOSE SPECS HAVE NEVER BEEN CORRELATED WITH
>> LISTENING QUALITY?

> Straw man argument.

Not at all. It's the fundamental issue here. Measurements and specs are
largely useless correlated with subjective impression.

> In fact, *all* of the things that are commonly specified for audio gear
> are known to have some correlation with listening quality.

Counter-example... Some years ago I performed a comparative review of ReVox
and Otari tape decks. The ReVox measured better in every respect -- but
there was no obvious subjective difference between them. (I recorded full
orchestra onto both machines at the same time.)

What about the measurements that _aren't_ performed? When was the last time
you checked an amplifier for its level of self-modulation? Or ran a
comb-filtered audio signal through it to see what "fell into the cracks"?

> For example, one common spec is frequency response.
> Now seriously, are you going to assert that there is no
> correlation between good frequency response and
> listening quality? I would seriously hope not!

I'm even willing to grant that variations in frequency response (followed by
IM distortion) are the two most-signifcant factors in sound quality. But I
"know" they're not the only ones.

>> We don't listen to specifications.

> Where did I say that we did listen to specifications? Everybody
> with a brain knows that specifications are abstractions of things
> that are known to cause sonic difficulties.

Okay. So we know IM distortion is A Bad Thing, right? How much is audible?
And how does the order of the distortion affect its audibility? If you can't
answer those questions, what is the point of measuring IM distortion (other
than to confirm the amp is working as designed)?

>> Nor are bandwidth and
>> dynamic range necessarily the only or principal things we
>> perceive when listening to reproduced sound.

> In fact bandwidth and dynamic range summarize everything that
> we hear wrong, when we hear something wrong. They include
> every kind of linear and nonlinear distortion, in the sense that any
> known kind of linear or nonlinear distortion degrades bandwidth
> and/or dynamic range.

From my perspective, you've just leaped over the edge. You've got to be
kidding.

Here's a counter-example. If I use an expander to increase dynamic range, I
added IM to the signal.

>> I don't understand how someone who appears to be
>> intelligent and well-educated can believe that controlled
>> (and likely over-controlled) listening tests that bear
>> little relationship to the way people actually listen are
>> the last word on sound differences or quality. They
>> aren't, and never were.

> That's another myth. In fact most real-world listening is done blind.
> I don't do a visual inspection of the entire record/reproduce chain
> when I listen to music. Nobody does. In fact much of it is unknowable
> most of the time.

Arny, I don't like "getting personal", but you really need to talk with
someone, and I don't mean a psychiatrist. Your point of view on matters
audio is beginning to harden rather badly. You're playing Humpty Dumpty too
often.

You are badly twisting the accepted purpose of blind auditions, which has
nothing to do with judging the quality of recordings, or more than one
component in the reproduction chain.

> Furthermore, the sound quality of audio chains tend to be dominated by
> the weakest links. These days, the weakest links are likely to not be the
> recorded formats used for production or distribution.

The formats might not be the weak links, but the recordings themselves are.
If you want to hear truly atrocious sound, find the CD of Murray Perahia
playing Scarlatti and Handel. Absolutely atrocious sound. I can't believe he
authorized its release.

> Therefore when I listen to other people's recordings, it is mostly
> a blind evaluation.

This is silly. I've heard -- and recorded -- live music often enough to have
a good idea of what it should sound like.

> That's one very strong reason why blind tests are the ones that provide
the
> results that correlate best with the broadest-based experience. To wit,
> blind tests show no audible benefits for so-called Hi-rez formats, and
> Hi-rez formats basically tanked in the marketplace.

Non-blind tests showed that Beta was superior, and it failed, too. There are
many reasons products fail, and quality (or lack thereof) is not always the
principal factor.

One of the reasons the public has not rushed to embraced hi-res video is
that most viewers find DVD more than adequate -- despite the _fact_ that
Blu-ray and HD DVD are visibly superior.

>> When someone is willing to run long-term blind or
>> double-blind listening tests, in which listeners are free
>> to listen in the usual way, when and how they feel like
>> it, without being forced to sit down and make comparisons
>> or decisions, then I will start taking listening tests
>> seriously.

> That all happened decades ago. No joy. It happened again
> decades later, still no joy.

I've never heard of such testing. Who did it?

>> But that will never happen, not only because
>> it's difficult to do properly, but because both sides in
>> this "issue" are desperately convinced that such tests
>> will disprove what they "know", beyond any shadow of
>> doubt, to be true.

> Wrong every way possible.

No, you're afraid that long-term listening would show repeatable and
significant differences that do not show up (or show up only weakly) in ABX
testing.

> The people who are most desperately convinced that they already know the
> right answers and don't need any more stinkin' tests to convince them of
> anything, are rather logically and obviously the people who have done the
> fewest tests.

> The people with the most open minds are the people who did and continute
to
> put their beliefs (which were not as one-sided as many would like to
> believe) to the test the most often and the most recently.

Then do what I ask, Arny. Set up long-term listening tests. Why are you
afraid they'll reveal something you don't want to acknowledge?

I've said it before, and I'll say it again... The essence of good science is
not looking for "fact", but asking the right questions.

William Sommerwerck

April 13th 07, 08:10 PM

Give me a few days on this one. I want to get back to Dr. Barclay and ask
him a thing or three.

William Sommerwerck

April 13th 07, 08:19 PM

> Has anybody A/B'd a DSD signal with its PCM copy [sic]
> in a vetted test yet? :)

Actually, a better test would be to compare the DSD layer of an SACD with
its CD layer. This itself would include two comparisons -- when the CD layer
is derived from the DSD stream (eg, a RCA Living Stereo SACD), and when the
CD layer is a direct PCM transfer (eg, a Mercury Living Presence SACD).

I'm slightly embarrassed to say that, though I've had an SACD player for
several years, I haven't made the comparison. It's clumsy to do well,
because the player cannot simultaneously output both signals. Perhaps I'll
find time over the weekend...

Arny Krueger

April 13th 07, 08:45 PM

"William Sommerwerck" > wrote in
message

>> Straw man argument.

> Not at all. It's the fundamental issue here. Measurements
> and specs are largely useless correlated with subjective
> impression.

Specs didn't come from outer space or pure theory. They were developed
progressively as people tried to obtain objective measures of sound quality.

>> In fact, *all* of the things that are commonly specified
>> for audio gear are known to have some correlation with
>> listening quality.

> Counter-example... Some years ago I performed a
> comparative review of ReVox and Otari tape decks. The
> ReVox measured better in every respect -- but there was
> no obvious subjective difference between them. (I
> recorded full orchestra onto both machines at the same
> time.)

Not an example at all. An analog tape machine is a tremendously complex
device to relevantly characterize with numbers, what with all the frequency
and amplitude dependencies. You picked an example that tends to be
self-obfuscating.

> What about the measurements that _aren't_ performed?

They seem to fit into two categories, either not sonically relevant or so
sonically relevant that their omission obfuscates practical sound quality
considerations.

> When
> was the last time you checked an amplifier for its level
> of self-modulation?

That's known as nonlinear distortion. IM and/or HD measurements can
characterize it.

> Or ran a comb-filtered audio signal
> through it to see what "fell into the cracks"?

Actually, I've probably published about as many comb filter tests as anybody
in modern times. Check these two pages out, there's almost 100 comb filter
tests of mine between them:

http://www.pcavtech.com/soundcards/compare/index.htm

http://www.pcavtech.com/play-rec/summary/index.htm

>> For example, one common spec is frequency response.
>> Now seriously, are you going to assert that there is no
>> correlation between good frequency response and
>> listening quality? I would seriously hope not!

> I'm even willing to grant that variations in frequency
> response (followed by IM distortion) are the two
> most-signifcant factors in sound quality. But I "know"
> they're not the only ones.

The choir hears and approves. ;-)

>>> We don't listen to specifications.
>
>> Where did I say that we did listen to specifications?
>> Everybody
>> with a brain knows that specifications are abstractions
>> of things
>> that are known to cause sonic difficulties.

> Okay. So we know IM distortion is A Bad Thing, right? How
> much is audible?

It depends on what sort of IM and what you listen to. I assembled what is
arguably perceptually a worst case for source material and found that 0.1%
IM is someplace around JND.

> And how does the order of the distortion
> affect its audibility?

IM is one of those places where low order distortion is about as sonically
deadly as anything.

> If you can't answer those
> questions, what is the point of measuring IM distortion
> (other than to confirm the amp is working as designed)?

My approach is to use knowlege about psychoacoustics - IOW what sort of
sounds produce the least-likely-to-be-masked distortion products, and what
sorts of equipment faults produce the least-likely-to-be-masked distortion
products, and test for JND with those.

>>> Nor are bandwidth and
>>> dynamic range necessarily the only or principal things
>>> we perceive when listening to reproduced sound.

>> In fact bandwidth and dynamic range summarize everything that
>> we hear wrong, when we hear something wrong. They include
>> every kind of linear and nonlinear distortion, in the
>> sense that any known kind of linear or nonlinear
>> distortion degrades bandwidth
>> and/or dynamic range.

> From my perspective, you've just leaped over the edge.

I'm not worried.

> You've got to be kidding.

I'm on solid ground, whether you know it or not.

> Here's a counter-example. If I use an expander to
> increase dynamic range, I added IM to the signal.

Of course. That's why those products are properly called dynamics expanders,
not dynamic range expanders. They can expand dynamics but the IM that comes
with them means that dynamic range is not expanded.

>>> I don't understand how someone who appears to be
>>> intelligent and well-educated can believe that
>>> controlled (and likely over-controlled) listening tests
>>> that bear little relationship to the way people
>>> actually listen are the last word on sound differences
>>> or quality. They aren't, and never were.

>> That's another myth. In fact most real-world listening
>> is done blind.
>> I don't do a visual inspection of the entire
>> record/reproduce chain
>> when I listen to music. Nobody does. In fact much of it
>> is unknowable most of the time.

> Arny, I don't like "getting personal", but you really
> need to talk with someone, and I don't mean a
> psychiatrist. Your point of view on matters audio is
> beginning to harden rather badly. You're playing Humpty
> Dumpty too often.

If you've got something relevant to share besides your frustration, you can
resume sharing it at your earliest convenience.

>You are badly twisting the accepted purpose of blind auditions,

No, I'm making a point that you probably never thought of.

> which has nothing to do with judging the quality of recordings,

That's rarely if ever done blind.

> or more than one component in the reproduction chain.

Since the partitioning of processing functions into components has a certain
amount of arbitraryness, what constitutes "one component" is one of those
YMMV things.

>> Furthermore, the sound quality of audio chains tend to
>> be dominated by the weakest links. These days, the
>> weakest links are likely to not be the recorded formats
>> used for production or distribution.

> The formats might not be the weak links, but the
> recordings themselves are. If you want to hear truly
> atrocious sound, find the CD of Murray Perahia playing
> Scarlatti and Handel. Absolutely atrocious sound. I can't
> believe he authorized its release.

Stuff happens. That's one reason why I try to restrict my comments to the
properties of formats, and not get mixed up with such highly specific
implementations of them.

>> Therefore when I listen to other people's recordings, it
>> is mostly a blind evaluation.

> This is silly. I've heard -- and recorded -- live music
> often enough to have a good idea of what it should sound
> like.

Everybody seems to say that, and since I've made about a 1,000 recordings
of live performances in the past 5 years, more than 100 recordings of live
performances in just the past 60 days, I get to say it with as much vigor as
most. ;-) Now that I've said it, I could write a paper about what's wrong
with that idea, no? The executive summary is that there is a lot wrong with
that idea.

>> That's one very strong reason why blind tests are the
>> ones that provide the results that correlate best with
>> the broadest-based experience. To wit, blind tests show
>> no audible benefits for so-called Hi-rez formats, and
>> Hi-rez formats basically tanked in the marketplace.

> Non-blind tests showed that Beta was superior, and it
> failed, too.

Not nearly as badly or quickly. And, there weren't two independent
implementations of Beta that went to market concurrently, sponsored by two
separate consortiums of manufacturers, etc., etc. Not a good comparison at
all.

> There are many reasons products fail, and
> quality (or lack thereof) is not always the principal factor.

That's one of the differences - the problem was not that SACD or DVD-A
lacked quality. What they lacked is a distinct quality advantage that
justified the hassle that went with them, which really wasn't that much.

> One of the reasons the public has not rushed to embraced
> hi-res video is that most viewers find DVD more than
> adequate -- despite the _fact_ that Blu-ray and HD DVD
> are visibly superior.

The jury for Blu-Ray and HD DVD hasn't even been picked yet. ;-)

>>> When someone is willing to run long-term blind or
>>> double-blind listening tests, in which listeners are
>>> free to listen in the usual way, when and how they feel
>>> like it, without being forced to sit down and make
>>> comparisons or decisions, then I will start taking
>>> listening tests seriously.

>> That all happened decades ago. No joy. It happened again
>> decades later, still no joy.

> I've never heard of such testing. Who did it?

One example is Nousaine's Audio article "Flying Blind".

Nousaine, Thomas, "Flying Blind: The Case Against Long Term Testing", Audio,
pp. 26-30, Vol. 81 No. 3 (March 1997)

There are over 60 relevant google hits, so people have been talking about it
quite a bit.

>>> But that will never happen, not only because
>>> it's difficult to do properly, but because both sides in
>>> this "issue" are desperately convinced that such tests
>>> will disprove what they "know", beyond any shadow of
>>> doubt, to be true.

>> Wrong every way possible.

> No, you're afraid that long-term listening would show
> repeatable and significant differences that do not show
> up (or show up only weakly) in ABX testing.

I'm eagerly waiting for documentation of such a test. I've been waiting now
for several decades.

>> The people who are most desperately convinced that they
>> already know the right answers and don't need any more
>> stinkin' tests to convince them of anything, are rather
>> logically and obviously the people who have done the
>> fewest tests.

>> The people with the most open minds are the people who
>> did and continute to put their beliefs (which were not
>> as one-sided as many would like to believe) to the test
>> the most often and the most recently.

> Then do what I ask, Arny. Set up long-term listening
> tests. Why are you afraid they'll reveal something you
> don't want to acknowledge?

I keep telling you that this is old news. And now I've illustrated that
with an article that was published in a well-known popular journal that you
seem to know nothing about.

> I've said it before, and I'll say it again... The essence
> of good science is not looking for "fact", but asking the
> right questions.

A lot of this is really old news. It is just that you don't know about it
for some reason.

Arny Krueger

April 13th 07, 08:47 PM

"William Sommerwerck" > wrote in
message
>> Has anybody A/B'd a DSD signal with its PCM copy [sic]
>> in a vetted test yet? :)
>
> Actually, a better test would be to compare the DSD layer
> of an SACD with its CD layer. This itself would include
> two comparisons -- when the CD layer is derived from the
> DSD stream (eg, a RCA Living Stereo SACD), and when the
> CD layer is a direct PCM transfer (eg, a Mercury Living
> Presence SACD).
>
> I'm slightly embarrassed to say that, though I've had an
> SACD player for several years, I haven't made the
> comparison. It's clumsy to do well, because the player
> cannot simultaneously output both signals. Perhaps I'll
> find time over the weekend...

The obvious comparison is to transcribe a SACD to a CD with a good audio
interface, and them compare the two - level-matched, time-synched, and
blind. The hard part is the time-synched part - you have to keep the two
discs synched within a fraction of second, or a quick switch will easily
identify which is which even if they are actually identical.

Les Cargill

April 13th 07, 09:07 PM

Scott Dorsey wrote:

> Les Cargill > wrote:
>
>>If there's a difference, how do we account for it?
>
>
> If there is a difference, it's almost certainly due to implementation
> issues and not to the theory itself. Don't put that down, though...
> implementation issues are _important_.
>
>

That's kind of where I was leaning - if it's an "AK47
versus cuftomme English fowling peice" thing, that's
great. Cheap stuff will be better.

>>The Nyquist Theorem limits of PCM are well understood - which
>>constraint is improved on by DSD?
>
>
> None, but the performance of real-world converters may not equal that
> of a theoretically perfect converter.
>
>

But how far off are they *really*?

>>All Lavry is restating is that the Shannon Channel
>>Theorem is still holding sway, and that the additional
>>bandwidth can be a Problem.
>
>
> That's true. BUT, the good news about DSD is that it has power
> density issues that make it hard to do abusive processing and get
> away with it. That's a huge plus in my book.
>

It's not like somebody can't rig an 1176 between the
preamp and the DSD recorder. If that's not abusive...

>
>>Has anybody A/B'd a DSD signal with it's PCM copy
>>in a vetted test yet? :)
>
>
> Not one that I would trust yet.
> --scott

Geez, it's been years...

--
Les Cargill

William Sommerwerck

April 13th 07, 10:19 PM

>> Not at all. It's the fundamental issue here. Measurements
>> and specs are largely useless correlated with subjective
>> impression.

> Specs didn't come from outer space or pure theory. They were developed
> progressively as people tried to obtain objective measures of sound
quality.

If I realize that an amplifier produces harmonic distortion or
intermodualtion products, I can decided to measure them -- regardless of
whether I have a qualitative and quantitative understanderstanding of how
they affect subjective quality.

> > What about the measurements that _aren't_ performed?

> They seem to fit into two categories, either not sonically relevant or so
> sonically relevant that their omission obfuscates practical sound quality
> considerations.

How do YOU know, God?

I don't know why I argue with Arny. He knows nothing at all -- including how
to ask the right questions -- but thinks he understands everything.

William Sommerwerck

April 13th 07, 10:22 PM

> The obvious comparison is to transcribe a SACD to a CD with a good audio
> interface, and them compare the two - level-matched, time-synched, and
> blind. The hard part is the time-synched part - you have to keep the two
> discs synched within a fraction of second, or a quick switch will easily
> identify which is which even if they are actually identical.

That's only one possible test, and the least critical and useful, in my
opinion.

It's easy enough to do with an ADC followed by a DAC, such as those in a
DAT. (I could do it with my Sony.) It does not, of course, introduce much
jitter or other waveform errors that might conceivably alter the sound.

Harry Lavo

April 13th 07, 10:48 PM

"Arny Krueger" > wrote in message
. ..
> "William Sommerwerck" > wrote in
> message
>
>>> Your thesis Harry, which nets out to the magical idea
>>> that DSD has
>>> sonic advantages over PCM with comparable bandwidth and
>>> dynamic
>>> range, simply violates the laws of physics as we know
>>> them, and as
>>> we have observed them to work for more than 50 years.
>
>> WHAT laws of physics do we "know", Arny?
>
> If you have to ask... ;-(
>
>> That if product
>> A has measured specs comparable to product B, that they
>> necessarily sound alike (or very nearly so), EVEN THOUGH
>> THOSE SPECS HAVE NEVER BEEN CORRELATED WITH LISTENING
>> QUALITY?
>
> Straw man argument.
>
> In fact, *all* of the things that are commonly specified for audio gear
> are known to have some correlation with listening quality.
>
> For example, one common spec is frequency response. Now seriously, are you
> going to assert that there is no correlation between good frequency
> response and listening quality? I would seriously hope not!
>
>> We don't listen to specifications.
>
> Where did I say that we did listen to specifications? Everybody with a
> brain knows that specifications are abstractions of things that are known
> to cause sonic difficulties.
>
>> Nor are bandwidth and
>> dynamic range necessarily the only or principal things we
>> perceive when listening to reproduced sound.
>
> In fact bandwidth and dynamic range summarize everything that we hear
> wrong, when we hear something wrong. They include every kind of linear
> and nonlinear distortion, in the sense that any known kind of linear or
> nonlinear distortion degrades bandwidth and/or dynamic range.
>
>> I don't understand how someone who appears to be
>> intelligent and well-educated can believe that controlled
>> (and likely over-controlled) listening tests that bear
>> little relationship to the way people actually listen are
>> the last word on sound differences or quality. They
>> aren't, and never were.
>
> That's another myth. In fact most real-world listening is done blind. I
> don't do a visual inspection of the entire record/reproduce chain when I
> listen to music. Nobody does. In fact much of it is unknowable most of the
> time.
>
> Furthermore, the sound quality of audio chains tend to be dominated by the
> weakest links. These days, the weakest links are likely to not be the
> recorded formats used for production or distribution.
>
> Finally, many parts of the audio chain that tend to dominate SQ are things
> whose usuage details tends to dominate over the inherent properties of the
> device. For example, I can know the on-axis frequency response of a
> certain microphone at a certain distance, but I also know that the mic's
> frequency response is highly dependent on distance and angle, which I
> geneally can't very easily know for other people's recordings.
>
> Therefore when I listen to other people's recordings, it is mostly a
> blind evaluation.
>
> Similarly, most people listen mostly blind most of the time.
>
> That's one very strong reason why blind tests are the ones that provide
> the results that correlate best with the broadest-based experience. To
> wit, blind tests show no audible benefits for so-called Hi-rez formats,
> and Hi-rez formats basically tanked in the marketplace.
>
>> When someone is willing to run long-term blind or
>> double-blind listening tests, in which listeners are free
>> to listen in the usual way, when and how they feel like
>> it, without being forced to sit down and make comparisons
>> or decisions, then I will start taking listening tests
>> seriously.
>
> That all happened decades ago. No joy. It happened again decades later,
> still no joy.
>
>> But that will never happen, not only because
>> it's difficult to do properly, but because both sides in
>> this "issue" are desperately convinced that such tests
>> will disprove what they "know", beyond any shadow of
>> doubt, to be true.
>
> Wrong every way possible.
>
> The people who are most desperately convinced that they already know the
> right answers and don't need any more stinkin' tests to convince them of
> anything, are rather logically and obviously the people who have done the
> fewest tests.
>
> The people with the most open minds are the people who did and continute
> to put their beliefs (which were not as one-sided as many would like to
> believe) to the test the most often and the most recently.

Arny, you simply don't know what you don't know. You might profitably spend
some time monitoring the discussions JJ is having on the Audio Pro
newslist.....rather than spouting off here about the "known knowns". You
know the "known knowns" but you don't know the "known unknowns" much less
the "unknown unknowns" and the "unknown knowns"). (Yes, that's that damned
Rumsfeld, but it is basically a restating of the Johari Window, so I am
allowed :-) ).

[email protected]

April 13th 07, 11:12 PM

On Apr 13, 2:24 pm, "Arny Krueger" > wrote:

> >>> On Apr 13, 8:09 am, "Arny Krueger" >
> >>> wrote:
>
> >>>> I basically like Dan and I like what he's trying to
> >>>> do, but I don't see what he is about right now as
> >>>> having a lot of legs.
>
> You're quoting me saying something that I never said. So much for the
> truth.

bull ****, my qouting was done using the google interface, ... you
said it!!!

> How many hours a week do you listen to live music with a console at hand?
> I'm setting no records, but for me it averages about 6-8 hours a week. IOW,
> your negative comments about my lack of interest in listening are completely
> bogus inventions of your mind.

four to six hrs a day here no console, no comb filtering, 96/24.
IOW using a 39.95 sacd player as a reference is not serious listening,
the spdiff cable I use cost more than that cheap piece of crap
and it is a stock item not a tweakoid cable.

[email protected]

April 13th 07, 11:12 PM

Scott Dorsey

April 13th 07, 11:13 PM

William Sommerwerck > wrote:
>> Has anybody A/B'd a DSD signal with its PCM copy [sic]
>> in a vetted test yet? :)
>
>Actually, a better test would be to compare the DSD layer of an SACD with
>its CD layer. This itself would include two comparisons -- when the CD layer
>is derived from the DSD stream (eg, a RCA Living Stereo SACD), and when the
>CD layer is a direct PCM transfer (eg, a Mercury Living Presence SACD).

This WOULD be a good test if there were not some SACD issues where the
CD layer was somewhat degraded (or maybe enhanced depending on how you
look at it) by different processing. If you can be assured that the
processing is the same and you can match levels, it would be a very
interesting comparison.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Harry Lavo

April 13th 07, 11:31 PM

"Mike Rivers" > wrote in message
ups.com...
> On Apr 13, 8:09 am, "Arny Krueger" > wrote:
>
>> I read what Dan said here about them a little differently. And, if he
>> thought the issues through from scratch today, he might say things a
>> little
>> differently than he did the last time he was posting a lot here.
>
> I'm not going to fight Lavry's battle for him, but I suspect that it
> would be hard to find a 192 kHz converter that sounds as good as his
> 96 kHz converter. SOUNDS being the operative word here. Here's a
> relevant portion from his 2004 Sampling Rate paper:

>snip Lavry's report, irrelevant to my comment below<

He and JJ and others have been having a very interesting discussion on the
Pro Audio listserve since before March 1 on other aspects of the problem
(they agree, btw, that from a practical standpoint, 64khz/20bits is an
adequate standard for utter transparency in final reproduction....basically
the same conclusion as reached by the hi-rez industry group in Japan back in
the mid-90's.) Their current discussion focuses on how the sliding time
window of our hearing sensitivity wrecks havoc with the filter
characteristics of even the most benign filter at 20khz...offering yet
another potential distortion that could be and clearly may be audible when
it comes to the conventional CD cutoff. They've also discussed the
pre-ringing of PCM filters, and JJ at least, feels that under some
circumstances it cleary could be audible.

>> The converter market is still very dynamic. For example, it very much
>> looks
>> like someone is now selling the moral equivalent of the Lynx L22 for
>> under
>> $200.
>
> Who might that be? I could use another decent sound card.
>

What the hell is the "moral" equivalent of a decoder?

Harry Lavo

April 13th 07, 11:48 PM

"Arny Krueger" > wrote in message
...
> "William Sommerwerck" > wrote in
> message
>>> Been there done that - the difference is that I dare to
>>> listen without having my eyes tell me what the *right*
>>> answer should be.
>>
>> No, you have your "science" telling your ears what they
>> ought to hear.
>
> Doesn't matter. Besides, enough listening is done by people who are
> ignorant of the science so that if there was some hidden truth, it would
> out.

Arny has said elsewhere that it doesn't matter that ABX has no controls
against "false negatives" because he "knows" he always gives it his best
shot to differentiate. LOL! But of course, any of us doing sighted tests
can't possibly steel ourselves against bias.

>
>> "The proof of the pudding is in the eating." Arny, if you
>> know so much... Tell me why the Crown K1 amplifier --
>> which has excellent specs, and was designed by one of the
>> "industry greats", Gerald Stanley -- sounds so utterly
>> awful.
>
> Never heard one, never tested one on the bench. I plead ignornace.
>
>> Then tell me why a Parasound A21 sounds so
>> completely different from a Krell KSA-250.
>
> Same basic answer.

Now look up the number of times Krueger has said in so many words "all
competent amplifiers not exceeding their design limits" are
undistinquishable from one another.

Apparently "science" allows him to reach that conclusion without actually
listening to amplifiers.

>
>> I make the following claim, and I'd be willing to prove
>> it... Given recordings I'm familiar with, I will tell you
>> which of these amplifiers is playing, without having to
>> compare it with any of the others. That's how grossly
>> different they sound.
>
> This makes no difference to me. I imagine that if it were convenient
> enough, I'd put forth the effort to do a proper set of listening tests and
> measurements and try to explain the correlation between measurements and
> listening tests. But, sighted casual evaluations are so fraught with
> difficulites, I won't even try if they are all there is.

In other words, don't bother me with no stupid long term listening.

>
>> I do not believe in magic.
>
> Do you believe that SACD sounds better or even different from PCM, given
> the same or at least similar dynamic range and bandwidth?
>
>> I do not believe the reasons
>> for sonic differences are incomprehensible or
>> unmeasurable.
>
> I'm unconvinced.

But where there are measureable differences, between the two, such as SACD's
lack of pre-ringing and superior asymetrical impulse response (and thus more
lifelike transient response) you simply dismiss the possibility.
Fortunately, true engineers such as JJ and Dan Lavry don't....they speculate
and investigate as they are doing now.

>
>>I merely claim that nobody has done much to
>> legitimately qualify and quantify such differences.
>
> There's a legion of serious researchers that would disagree with that.

And a legion more who would agree with that.

>
>> All we get from Arny, et al., is if there are no measurable
>> differences, then there are no sonic differences -- which
>> is plainly intellectual bilge.
>
> No, what we find repeatedly is that if there are no audible differences,
> then the measurements agree that there should be no audible differences.

That's not how you preach the gospel, my friend.

>
>> I want Mr. Arny "I know the secrets of the universe"
>> Krueger to tell me what the sonically significant
>> measurements are, and _why_ they correlate with one's
>> subjective reactions.

Without looking below, I'll bet he doesn't answer your question.

>
> My idea of a complete set of measurement of a power amp is here:
>
> http://www.pcavtech.com/pwramp/macrot-5000VZ/index.htm
>
> Here are some brief audio samples that demonstrate some of the audible
> differences that can be attributed to this amp:
>
> http://www.pcabx.com/product/macrot-5000VZ/index.htm

Bingo! Essentially "what I know about and can measure I can hear". But
don't ask me to hear what I haven't yet measured.

Harry Lavo

April 13th 07, 11:56 PM

"Arny Krueger" > wrote in message
...
> > wrote in message
> ups.com
>> On Apr 13, 10:48 am, "Arny Krueger" >
>> wrote:
>>> > wrote in message
>>>
>>> ups.com
>>>
>>>> On Apr 13, 8:09 am, "Arny Krueger" >
>>>> wrote:

>snip, irrelevant to what follows<

>
>> but you can hear that good!!!
>
> I do a lot of recording and live sound. Both are arts that are rather
> dependent on one's ability to listen good sound into existence.

On the other hand, that is not a courtesy you extend to Michael Bishop.

>> but then listening is an art...
>
> You're preaching to the choir.

The church choir I do most of my recordings with (he says) ;-)

>
>> and it requires practice... too
>
> How many hours a week do you listen to live music with a console at hand?
> I'm setting no records, but for me it averages about 6-8 hours a week.
> IOW, your negative comments about my lack of interest in listening are
> completely bogus inventions of your mind.

When you work your way up to producing successful commercial CD's like
Michael Bishop, perhaps you should drop him a note and ask whether you could
listen to DSD along with him, as you can't hear a difference from 44.1/16
and perhaps he could point it out.

Harry Lavo

April 14th 07, 12:05 AM

"Chel van Gennip" > wrote in message
...
> On Fri, 13 Apr 2007 15:53:39 +0200, Harry Lavo wrote:
>
>
>> "Chel van Gennip" > wrote in message
>> ...
>>> On Fri, 13 Apr 2007 01:15:16 +0200, William Sommerwerck wrote:
>>>
>>>>> ...the consumer implementation, SACD, was not a success in the
>>>>> market.
>>>>
>>>> How do you define "success"? I have over 200 SACDs.
>>>
>>> I think that a market share dropping from 0.15% to 0.1% a few years
>>> after introduction is NOT a success in the market. And, as I wrote
>>> before, by now there are much better technical ways to distribute audio
>>> available.
>>>
>>>
>> I think you will find its importance in the classical music market is
>> alive and getting stronger.
>
> The discussion is about market success. Do you have actual sales figures?
> In the Dutch (NVPI)figures I see a 50% drop in SACD sales from 2003 to
> 2004, in 2005 SACD is not mentioned anymore.
>
>> Let's hope the newer hi-rez video mediums can be more successfully
>> adopted, and that SACD continues to expand.
>
> "SACD continues to expand" has nothing to do with reality, sales are
> dropping fast for a number of years now.

I don't know the source you are quoting, but it is important that it include
web sources, since virtually all SACD sales and most DVD-A sales have moved
to the web. You also are no doubt aware that Sony's leaving the market
sharply reduced manufacturers shipments for a period of about 1-1/2 years,
coinciding with your time period I believe.

I'd have to dig back in the archives, but over on RAHE we have this
discussion every year and I believe last time around we calculated that in
the US, overall classical sales were dropping as fast/faster than SACD
indicating a probable continued (although not very large) gain in share.
But even those numbers are not very clear, as the RIAA is not totally
forthcoming about what manufacturers are counted or not in sales, especially
web sales. For example, Channel Classics, Pentatone, BIS, and Naxos are all
companies issuing SACD after SACD...the only company in the US doing so is
Telarc....along with the SFO and Philadelphia Orchestra. And Hong Kong and
Japanese companies are issuing SACDs hand-over-fist, including major label
reissues that are not releasing those SACDs in the US. So it is a very
fluid and dynamic market and hard to judge. But the list of titles keeps
growing with about 80 new issues each month.

Mike Rivers

April 14th 07, 12:14 AM

On Apr 13, 6:31 pm, "Harry Lavo" > wrote:

> He and JJ and others have been having a very interesting discussion on the
> Pro Audio listserve since before March 1 on other aspects of the problem

I tried to follow that for a while, but it got kind of deep for a
practical-minded guy like me. But as of today, they seem to at least
agree that things are getting better.

> What the hell is the "moral" equivalent of a decoder?

I think he probably means a functional equivalent of the Lynx L22
sound card, which is a mighty good sounding I/O device for a computer.
Even I noticed an immediate difference when I replaced the Turtle
Beach card in my modest studio computer with an L22.

[email protected]

April 14th 07, 12:14 AM

On Apr 13, 6:31 pm, "Harry Lavo" > wrote:

> What the hell is the "moral" equivalent of a decoder?

isn't that the term guitar center sales people use
when the sales spif is better???

Harry Lavo

April 14th 07, 12:27 AM

"Arny Krueger" > wrote in message
...
> "William Sommerwerck" > wrote in
> message
>>> Has anybody A/B'd a DSD signal with its PCM copy [sic]
>>> in a vetted test yet? :)
>>
>> Actually, a better test would be to compare the DSD layer
>> of an SACD with its CD layer. This itself would include
>> two comparisons -- when the CD layer is derived from the
>> DSD stream (eg, a RCA Living Stereo SACD), and when the
>> CD layer is a direct PCM transfer (eg, a Mercury Living
>> Presence SACD).
>>
>> I'm slightly embarrassed to say that, though I've had an
>> SACD player for several years, I haven't made the
>> comparison. It's clumsy to do well, because the player
>> cannot simultaneously output both signals. Perhaps I'll
>> find time over the weekend...
>
> The obvious comparison is to transcribe a SACD to a CD with a good audio
> interface, and them compare the two - level-matched, time-synched, and
> blind. The hard part is the time-synched part - you have to keep the two
> discs synched within a fraction of second, or a quick switch will easily
> identify which is which even if they are actually identical.

If you are doing it blind (especially double blind) why would you know which
was which?

You don't need to re-record.

You need two identical machines that ears and measurements tell you are
identical in both SACD and CD so far as freqency response f rom 30hz-18khz
is concerned and which exhibit no other anomalies.

You need two identical hybrid sacds...ideally recorded directly in DSD with
the same stereo mix fed out of the DSD workstation. Most contemporary
classical SACDs are prepared this way. A separate test using really
high-quality master tape would be excellent as well as yet another source.

You need to time-synch and double-blind test. Ideally it should be a
preference test, rather than an identity test as work at HK suggests this is
easier for people to do.

Ideally you do it across a cross-section of careful and experienced
listeners well attuned to live classical music for a set number of tests
each, allowing each to control the switching.

You do it in a casual and comfortable home-like setting, with amenities and
no pressure and you allow them to stop when fatiqued and resume when they
are rested and have taken a break.

You sum the responses and apply statistical analysis. When you get over
100-150 separate samples from at least 10 people, you look at the
statistics. For scientific "proof" you want 95% reliability (19:1 odds).
Great if you get 99% (99:1). Still interesting if you get 80% (4:1).

If you want a control (and you should) you provide a blind monadic test
individually to 100 similar listeners for each of the two formats, and ask
them to rate the music and musical playback quality on a serious of semantic
differential scales. When finished you then apply statistics to the scalers
between the two groups of 100, and note differences that are statistically
signifcant at high rates. The overall results should support the individual
pairs conclusions if both tests are valid, and the semantic differential
scaling should tell you how and why people perceived one as better than the
other.

A simple little test....don't know why everybody doesn't do it.......

William Sommerwerck

April 14th 07, 01:02 AM

>> I don't know the source you are quoting, but it is important
>> that it include web sources, since virtually all SACD sales
>> and most DVD-A sales have moved to the web.

I recently saw a statement that the total number of SACD sales last year was
about one in a thousand of all optical-media audio sales. That's not so hot.

Of course, it doesn't matter how many are sold, as long as the record
companies can at least break even on the ones they sell.

William Sommerwerck

April 14th 07, 01:04 AM

> If you want a control (and you should) you provide a blind monadic test
> individually to 100 similar listeners for each of the two formats, and ask
> them to rate the music and musical playback quality on a serious of
semantic
> differential scales. When finished you then apply statistics to the
scalers
> between the two groups of 100, and note differences that are statistically
> signifcant at high rates. The overall results should support the
individual
> pairs conclusions if both tests are valid, and the semantic differential
> scaling should tell you how and why people perceived one as better than
the
> other.

Yeah... That, too.

Harry Lavo

April 14th 07, 02:23 AM

"Chel van Gennip" > wrote in message
...
> On Sat, 14 Apr 2007 01:05:53 +0200, Harry Lavo wrote:
>
>>>> Let's hope the newer hi-rez video mediums can be more successfully
>>>> adopted, and that SACD continues to expand.
>>>
>>> "SACD continues to expand" has nothing to do with reality, sales are
>>> dropping fast for a number of years now.
>>
>> I don't know the source you are quoting, but it is important that it
>> include web sources, since virtually all SACD sales and most DVD-A sales
>> have moved to the web.
>
> The US source:
> http://www.riaa.com/news/newsletter/pdf/2005yrEndStats.pdf
> The Dutch source:
> http://www.nvpi.nl/assets/nvpi/NVPI%20MARKTINFORMATIE%20AUDIO%202005.pdf
>
> --
> Chel van Gennip (chel vangennip nl)
> Visit Serg van Gennip's site http://www.serg.vangennip.com

Thanks for the Links. The US RIAA stuff I actually had tucked away on
file...but not the Dutch data. Any idea when these things are
updated....seems like 2006 data should be iminent.

At any rate I did an analysis and am sending you the spreadsheet via email.

A basic summary is as follows:

* Classical sales in Holland are declining as a percent of the total market,
as well as in units and dollars -- from 8% in 2003 to 6% in 2005.
* Comparable sales figures were not included for the US...I used the
oft-quoted figure of 3% of sales for all three years in the comparison.
* Accordingly, classical as a category is about twice as large in Holland as
in the US.
* On top of that, the ratio of hi-rez (DVD-A & SACD) to Classical sales is
more than twice as high in Holland as in the US (about 17% vs. about 7%)
* Thus, hi-rez sales to total market is about four times as high in Holland
as it is in the US (1.0% vs. 0.2%).
* In both markets the ratio stayed the same or increased from 2004 to 2005,
after falling from slightly higher levels 2003 to 2004.
* If classical is falling as a percent of total sales in the United States,
as it is in Holland, (rather than staying at the assumed 3%) then the ratio
of hi-rez to classical sales is actually increasing in the United States (as
it is in Holland) rather than staying steady.
* When all this is added together, in the United States hi-rez dollar sales
fell 8.4% 2005 over 2004, identical to the total market, while in Holland
hi-rez sales actually rose 3.4% while the total market declined 13.4%.
* These numbers clearly show the small size of the niche market, but don't
suggest things are worse today than in 2004 (it will be interesting to see
2006...my guess is that the ratio will increase still further and that total
hi-rez sales will grow slightly).

Again, thanks for the sales charts.

Harry Lavo

April 14th 07, 02:31 AM

"William Sommerwerck" > wrote in message
. ..
>>> I don't know the source you are quoting, but it is important
>>> that it include web sources, since virtually all SACD sales
>>> and most DVD-A sales have moved to the web.
>
> I recently saw a statement that the total number of SACD sales last year
> was
> about one in a thousand of all optical-media audio sales. That's not so
> hot.
>
> Of course, it doesn't matter how many are sold, as long as the record
> companies can at least break even on the ones they sell.
>

Two-tenths of one percent...that's more like one-in-five-hundred. But it
was steady from 2004 to 2005...I'm anxious to see the 2006 numbers.
Incidentally, if you read my other post in response to Chel, you'll see it
is five times higher in Holland, or one-in-a-hundred.

Harry Lavo

April 14th 07, 02:34 AM

"Harry Lavo" > wrote in message
. ..

>snip<

> * Thus, hi-rez sales to total market is about four times as high in
> Holland as it is in the US (1.0% vs. 0.2%).

>snip<

This obviously should be "five times as high", not four.

[email protected]

April 14th 07, 06:45 AM

Harry Lavo > wrote:

> "Mike Rivers" > wrote in message
> ups.com...
> > On Apr 13, 8:09 am, "Arny Krueger" > wrote:
> >
> >> I read what Dan said here about them a little differently. And, if he
> >> thought the issues through from scratch today, he might say things a
> >> little
> >> differently than he did the last time he was posting a lot here.
> >
> > I'm not going to fight Lavry's battle for him, but I suspect that it
> > would be hard to find a 192 kHz converter that sounds as good as his
> > 96 kHz converter. SOUNDS being the operative word here. Here's a
> > relevant portion from his 2004 Sampling Rate paper:

> >snip Lavry's report, irrelevant to my comment below<

> He and JJ and others have been having a very interesting discussion on the
> Pro Audio listserve since before March 1 on other aspects of the problem
> (they agree, btw, that from a practical standpoint, 64khz/20bits is an
> adequate standard for utter transparency in final reproduction....basically
> the same conclusion as reached by the hi-rez industry group in Japan back in
> the mid-90's.) Their current discussion focuses on how the sliding time
> window of our hearing sensitivity wrecks havoc with the filter
> characteristics of even the most benign filter at 20khz...offering yet
> another potential distortion that could be and clearly may be audible when
> it comes to the conventional CD cutoff. They've also discussed the
> pre-ringing of PCM filters, and JJ at least, feels that under some
> circumstances it cleary could be audible.

And those circumstances were clearly and plainly stated to be very narrowly
defined test signals. Yet you said in RAHE that they were "often" audible
in a context that seemed to assume general conditions, since you have
earlier written many times that you think testing with defined test signals is
irrelevent for music listening. If they were "often" audible with music,
they would certainly recomend to use it as a test signal. But they don't!

You wanted to know where you misrepresented what they were talking about and
that is it. Please stop it.

Natalie Drest

April 14th 07, 12:37 PM

"Scott Dorsey" > wrote in message
...
> Harry Lavo > wrote:
> >

>
> The folks who are saying SACD is dead in the water and a dying format are
> probably right, but for that matter you could say the same thing about
> standard PCM formats like the compact disk. The mass market doesn't want
> quality. The question is whether SACD and CD can survive as niche players
> in the future, the way LP has hung on. I can't answer that one.
> --scott
> --
> "C'est un Nagra. C'est suisse, et tres, tres precis."

Would you be so kind as to you answer this one for me?

Here's a couple of points I've wondered about for a while, expressed fairly
concisely in an old article from Electronic Musician- I found it as a result
of reading this (long) thread.

"If you're asking, "Why would I want to record frequencies no one can hear?"
you're not alone. One reason is that frequencies we can't hear as pure tones
may nevertheless audibly influence the color of complex tones."

I'm not sold on this idea...

"In particular, the character of transients may be influenced by components
whose frequencies are above the official range of our hearing."

Mmmm... maybe?
"Another point in favor of high sampling rates is that timing differences
smaller than 1/44,100 of a second may affect our ability to localize sounds.
That suggests that DSD should be capable of superior imaging in stereo and
surround applications, and naturally Sony and Philips claim that that is the
case."

This seems to make sense to me... should I get the books out & try to
understand Mr. Nyquist? What am I missing here?

TIA...

Scott Dorsey

April 14th 07, 01:18 PM

Natalie Drest > wrote:
>"If you're asking, "Why would I want to record frequencies no one can hear?"
>you're not alone. One reason is that frequencies we can't hear as pure tones
>may nevertheless audibly influence the color of complex tones."
>
>I'm not sold on this idea...

It's possible but there is no actual good data for it, in spite of a bunch
of really badly botched papers from a Japanese concern that keep turning
up at AES shows. On the other hand, there isn't any good data to show
this _can't_ take place.

>"In particular, the character of transients may be influenced by components
>whose frequencies are above the official range of our hearing."
>
>Mmmm... maybe?

The idea makes sense intutively.... but there is no actual evidence for
the effect.

>"Another point in favor of high sampling rates is that timing differences
>smaller than 1/44,100 of a second may affect our ability to localize sounds.
>That suggests that DSD should be capable of superior imaging in stereo and
>surround applications, and naturally Sony and Philips claim that that is the
>case."

This is doubtful, however. For this, we _do_ actually have science. Above
around 1KHz or so, we use amplitude differences and not timing differences
to locate sounds. Timing differences are only used at lower frequencies. So
you can in fact get away with pretty substantial timing differences, and
there is a paper by Vanderkooy in the seventies that demonstrates how much
you can get away with before it becomes audible. 1/44,100 is not much.

>This seems to make sense to me... should I get the books out & try to
>understand Mr. Nyquist? What am I missing here?

Absolutely, you should definitely understand the Nyquist theorem if you
don't already. Nobody debates the accuracy of it, but what they do debate
is whether additional frequency response is useful or not. I am inclined
these days to say it isn't, but I wouldn't bet my life on it.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Natalie Drest

April 14th 07, 01:21 PM

"Scott Dorsey" > wrote in message
...
> Natalie Drest > wrote:
> >"If you're asking, "Why would I want to record frequencies no one can
hear?"
> >you're not alone. One reason is that frequencies we can't hear as pure
tones
> >may nevertheless audibly influence the color of complex tones."
> >
> >I'm not sold on this idea...
>
> It's possible but there is no actual good data for it, in spite of a bunch
> of really badly botched papers from a Japanese concern that keep turning
> up at AES shows. On the other hand, there isn't any good data to show
> this _can't_ take place.
>
> >"In particular, the character of transients may be influenced by
components
> >whose frequencies are above the official range of our hearing."
> >
> >Mmmm... maybe?
>
> The idea makes sense intutively.... but there is no actual evidence for
> the effect.
>
> >"Another point in favor of high sampling rates is that timing differences
> >smaller than 1/44,100 of a second may affect our ability to localize
sounds.
> >That suggests that DSD should be capable of superior imaging in stereo
and
> >surround applications, and naturally Sony and Philips claim that that is
the
> >case."
>
> This is doubtful, however. For this, we _do_ actually have science.
Above
> around 1KHz or so, we use amplitude differences and not timing differences
> to locate sounds. Timing differences are only used at lower frequencies.
So
> you can in fact get away with pretty substantial timing differences, and
> there is a paper by Vanderkooy in the seventies that demonstrates how much
> you can get away with before it becomes audible. 1/44,100 is not much.
>
> >This seems to make sense to me... should I get the books out & try to
> >understand Mr. Nyquist? What am I missing here?
>
> Absolutely, you should definitely understand the Nyquist theorem if you
> don't already. Nobody debates the accuracy of it, but what they do debate
> is whether additional frequency response is useful or not. I am inclined
> these days to say it isn't, but I wouldn't bet my life on it.
> --scott
> --
> "C'est un Nagra. C'est suisse, et tres, tres precis."

Thanks Scott. As usual, tres precis.

Mike Rivers

April 14th 07, 01:27 PM

On Apr 14, 7:37 am, "Natalie Drest" >
wrote:

> in an old article from Electronic Musician- I found it as a result
> of reading this (long) thread.

> "If you're asking, "Why would I want to record frequencies no one can hear?"
> you're not alone. One reason is that frequencies we can't hear as pure tones
> may nevertheless audibly influence the color of complex tones."

> I'm not sold on this idea...

Neither are a lot of other people, particularly since transducers
(microphones and loudspeakers) that have useful output above 20 kHz
aren't very common. But they exist, and recordings of certain
instruments have shown that there is considerable energy above the
human range of hearing. That doesn't mean that we should all do it all
the time, but there's something there that might be useful some day.

> "In particular, the character of transients may be influenced by components
> whose frequencies are above the official range of our hearing."

> Mmmm... maybe?

We don't completely understand the human hearing mechanism. While we
know that we can't hear continuous, pure tones above about 20 kHz,
give or take a few, we don't know where this filtering takes place. I
recall a somewhat inconclusive experiment where someone was trying to
show that energy above the normal range of perception causes the
"compressor" in our hearing mechanism to respond, so this could indeed
affect how we hear transients, or anything, really.

> "Another point in favor of high sampling rates is that timing differences
> smaller than 1/44,100 of a second may affect our ability to localize sounds.
> That suggests that DSD should be capable of superior imaging in stereo and
> surround applications, and naturally Sony and Philips claim that that is the
> case."

I've seen some calculations of wavelengths that suggest that this is
baloney, and no tests suggesting that it might actually be true.

> This seems to make sense to me... should I get the books out & try to
> understand Mr. Nyquist?

That certainly won't hurt, but if you want to do some research, you
should be looking at some obscure and little studied aspects human
physiology and how it affects our hearing. There isn't a lot of study
in this area because most of the studies in hearing are about being
able to understand speech and live a normal life, not being able to
improve our resolution of localization. You can improve your chances
of staying alive if you can hear that the bus that's about to run you
over is coming from the left or right, and that's about where the
practical science is.

I did learn something interesting about wide bandwidth recording from
the people who make a monstrous sample library with which they
simulate large pipe organs with multi-channel installations. They
record each pipe individually rather than recording the blended sound
of the organ from a distance, then they play back those pipes and let
them combine naturally in air. What they claim (and can prove by
looking at the frequency content of their recordings) is that there's
ultrasonic energy coming off a pipe, and when two pipes are playing
together, the beating of those ultrasonic frequencies is down in the
audible range. What you hear sounds different with and without that
ultrasonic energy.

My take on the whole thing is that I just don't want to worry about
things that I can't hear. I don't care about what you can hear. I'd
rather take the chance that someone will discover that something is
missing than take the chance that something that I can't hear (but
someone else can) will be annoying, or just plain wrong.

While there's a lot less going on up at the high end than at the low
end, it's easy to understand the concept when you think about bass.
It's often mixed incorrectly because it can't be heard correctly in
the mixing environment. If you can't hear the bass when you're mixing
(or recording) you won't know that it's getting in the way when
listening on a system that can accurately reproduce what's recorded.

William Sommerwerck

April 14th 07, 01:37 PM

> Neither are a lot of other people, particularly since transducers
> (microphones and loudspeakers) that have useful output above
> 20 kHz aren't very common. But they exist, and recordings of certain
> instruments have shown that there is considerable energy above the
> human range of hearing. That doesn't mean that we should all do it all
> the time, but there's something there that might be useful some day.

When designing bridges or skyscrapers, one leaves a substantial margin --
usually 100% -- for error. That is, if you think a structure will need to
support only 10,000 pounds, you nevertheless design it for 20,000.

I see no reason this should shouldn't (sort of) apply to sound reproduction.
I'd expect a tweeter that was "naturally" flat to 30kHz would "sound better"
than one that was flat to 20kHz, then dropped off precipitously.

Arny Krueger

April 14th 07, 03:17 PM

"Harry Lavo" > wrote in message

> "Arny Krueger" > wrote in message
> . ..
>> "William Sommerwerck" > wrote
>> in message
>>
>>>> Your thesis Harry, which nets out to the magical idea
>>>> that DSD has
>>>> sonic advantages over PCM with comparable bandwidth and
>>>> dynamic
>>>> range, simply violates the laws of physics as we know
>>>> them, and as
>>>> we have observed them to work for more than 50 years.
>>
>>> WHAT laws of physics do we "know", Arny?
>>
>> If you have to ask... ;-(
>>
>>> That if product
>>> A has measured specs comparable to product B, that they
>>> necessarily sound alike (or very nearly so), EVEN THOUGH
>>> THOSE SPECS HAVE NEVER BEEN CORRELATED WITH LISTENING
>>> QUALITY?
>>
>> Straw man argument.
>>
>> In fact, *all* of the things that are commonly specified
>> for audio gear are known to have some correlation with
>> listening quality. For example, one common spec is frequency response.
>> Now
>> seriously, are you going to assert that there is no
>> correlation between good frequency response and
>> listening quality? I would seriously hope not!
>>> We don't listen to specifications.
>>
>> Where did I say that we did listen to specifications? Everybody with a
>> brain knows that specifications are
>> abstractions of things that are known to cause sonic
>> difficulties.
>>> Nor are bandwidth and
>>> dynamic range necessarily the only or principal things
>>> we perceive when listening to reproduced sound.
>>
>> In fact bandwidth and dynamic range summarize everything
>> that we hear wrong, when we hear something wrong. They
>> include every kind of linear and nonlinear distortion,
>> in the sense that any known kind of linear or nonlinear
>> distortion degrades bandwidth and/or dynamic range.
>>> I don't understand how someone who appears to be
>>> intelligent and well-educated can believe that
>>> controlled (and likely over-controlled) listening tests
>>> that bear little relationship to the way people
>>> actually listen are the last word on sound differences
>>> or quality. They aren't, and never were.
>>
>> That's another myth. In fact most real-world listening
>> is done blind. I don't do a visual inspection of the
>> entire record/reproduce chain when I listen to music.
>> Nobody does. In fact much of it is unknowable most of
>> the time. Furthermore, the sound quality of audio chains tend to
>> be dominated by the weakest links. These days, the
>> weakest links are likely to not be the recorded formats
>> used for production or distribution. Finally, many parts of the audio
>> chain that tend to
>> dominate SQ are things whose usuage details tends to
>> dominate over the inherent properties of the device. For
>> example, I can know the on-axis frequency response of a
>> certain microphone at a certain distance, but I also
>> know that the mic's frequency response is highly
>> dependent on distance and angle, which I geneally can't
>> very easily know for other people's recordings. Therefore when I listen
>> to other people's recordings, it
>> is mostly a blind evaluation.
>>
>> Similarly, most people listen mostly blind most of the
>> time. That's one very strong reason why blind tests are the
>> ones that provide the results that correlate best with
>> the broadest-based experience. To wit, blind tests show
>> no audible benefits for so-called Hi-rez formats, and
>> Hi-rez formats basically tanked in the marketplace.
>>> When someone is willing to run long-term blind or
>>> double-blind listening tests, in which listeners are
>>> free to listen in the usual way, when and how they feel
>>> like it, without being forced to sit down and make
>>> comparisons or decisions, then I will start taking
>>> listening tests seriously.
>>
>> That all happened decades ago. No joy. It happened
>> again decades later, still no joy.
>>
>>> But that will never happen, not only because
>>> it's difficult to do properly, but because both sides in
>>> this "issue" are desperately convinced that such tests
>>> will disprove what they "know", beyond any shadow of
>>> doubt, to be true.

>> Wrong every way possible.

>> The people who are most desperately convinced that they
>> already know the right answers and don't need any more
>> stinkin' tests to convince them of anything, are rather
>> logically and obviously the people who have done the
>> fewest tests.

>> The people with the most open minds are the people who
>> did and continute to put their beliefs (which were not
>> as one-sided as many would like to believe) to the test
>> the most often and the most recently.

For a good example of people with closed minds, consider the people who
don't need tests to know that they are right, like Harry Lavo.

> Arny, you simply don't know what you don't know.

You seem to know so much less than I Harry, that it would be good for you to
first try to learn what I knew back in 1995.

> You might profitably spend some time monitoring the
> discussions JJ is having on the Audio Pro
> newslist.....rather than spouting off here about the
> "known knowns".

Been there, done that. Frankly, that list often makes me sick. It makes some
of my friends who publish in the JAES sick. There's so much to know about
audio that leads to such clear and obvious flaws with the work that we do,
and here's these guys obsessing over problems that are for most if not all
practically purposes, solved.

> You know the "known knowns" but you
> don't know the "known unknowns" much less the "unknown
> unknowns" and the "unknown knowns"). (Yes, that's that
> damned Rumsfeld, but it is basically a restating of the
> Johari Window, so I am allowed :-) ).

Dr. cure thy self.

William Sommerwerck

April 14th 07, 03:20 PM

>> When designing bridges or skyscrapers, one leaves a substantial
>> margin -- usually 100% -- for error. That is, if you think a structure
>> will need to support only 10,000 pounds, you nevertheless design
>> it for 20,000.

> When I buy lamps, I especially look for the ones with flat output in
> infrared and utraviolet, to see better greens [sic] and reds.

A lamp "flat" into the IR and UV might very well have more-even output in
the visible spectrum.

Do you know anything about color photography or color matching?

>> I see no reason this should shouldn't (sort of) apply to sound
>> reproduction. I'd expect a tweeter that was "naturally" flat to
>> 30kHz would "sound better" than one that was flat to 20kHz,
>> then dropped off precipitously.

> Do you have test reports that show the effects you expect?

I will sidestep that question, and pose you this one... If price weren't an
object, and you could use whatever technology you wanted... Would you design
a tweeter that was flat to 20kHz, or flat to 28kHz?

A driver doesn't roll off for some magical, mysterious reason. The main
reason is that it's hit a resonance. Now, would you rather that the
resonance -- and its attendant side-effects -- fell at the edge of and
within the audible spectrum -- or were above it?

Harry Lavo

April 14th 07, 03:26 PM

> wrote in message
...
> Harry Lavo > wrote:
>
>> "Mike Rivers" > wrote in message
>> ups.com...
>> > On Apr 13, 8:09 am, "Arny Krueger" > wrote:
>> >
>> >> I read what Dan said here about them a little differently. And, if he
>> >> thought the issues through from scratch today, he might say things a
>> >> little
>> >> differently than he did the last time he was posting a lot here.
>> >
>> > I'm not going to fight Lavry's battle for him, but I suspect that it
>> > would be hard to find a 192 kHz converter that sounds as good as his
>> > 96 kHz converter. SOUNDS being the operative word here. Here's a
>> > relevant portion from his 2004 Sampling Rate paper:
>
>> >snip Lavry's report, irrelevant to my comment below<
>
>> He and JJ and others have been having a very interesting discussion on
>> the
>> Pro Audio listserve since before March 1 on other aspects of the problem
>> (they agree, btw, that from a practical standpoint, 64khz/20bits is an
>> adequate standard for utter transparency in final
>> reproduction....basically
>> the same conclusion as reached by the hi-rez industry group in Japan back
>> in
>> the mid-90's.) Their current discussion focuses on how the sliding time
>> window of our hearing sensitivity wrecks havoc with the filter
>> characteristics of even the most benign filter at 20khz...offering yet
>> another potential distortion that could be and clearly may be audible
>> when
>> it comes to the conventional CD cutoff. They've also discussed the
>> pre-ringing of PCM filters, and JJ at least, feels that under some
>> circumstances it cleary could be audible.
>
> And those circumstances were clearly and plainly stated to be very
> narrowly
> defined test signals. Yet you said in RAHE that they were "often" audible
> in a context that seemed to assume general conditions, since you have
> earlier written many times that you think testing with defined test
> signals is
> irrelevent for music listening. If they were "often" audible with music,
> they would certainly recomend to use it as a test signal. But they don't!
>
> You wanted to know where you misrepresented what they were talking about
> and
> that is it. Please stop it.

Stop what? I have accurately summarized what they are discussing and have
concluded. I should not have used the word "often" on RAHE...it was an
honest mistake and overstatement...which I would have acknowledge had you
pointed this out to me over there, rather than simply "leaving" in a huff.
I didn't use it here. Their explorations included test signals...their
conclusions about potential significance were not restricted to test
signals. The fact that it doesn't square with your preconceptions is not my
problem.

Arny Krueger

April 14th 07, 03:26 PM

"Harry Lavo" > wrote in message

> "William Sommerwerck" > wrote
> in message
> . ..
>>>> I don't know the source you are quoting, but it is
>>>> important that it include web sources, since virtually
>>>> all SACD sales and most DVD-A sales have moved to the
>>>> web.
>>
>> I recently saw a statement that the total number of SACD
>> sales last year was
>> about one in a thousand of all optical-media audio
>> sales. That's not so hot.
>>
>> Of course, it doesn't matter how many are sold, as long
>> as the record companies can at least break even on the
>> ones they sell.

> Two-tenths of one percent...that's more like
> one-in-five-hundred. But it was steady from 2004 to
> 2005...I'm anxious to see the 2006 numbers. Incidentally,
> if you read my other post in response to Chel, you'll see
> it is five times higher in Holland, or one-in-a-hundred.

Time to wake up and smell the coffee, Harry. The numbers for the first half
of 2006 have been out for over 6 months.

Announcement:

http://www.riaa.com/news/newsletter/101206.asp

Actual numbers:

http://www.riaa.com/news/newsletter/pdf/2006midYrStats.pdf

SACD > 44% down from previous period.
DVD-A > 35% down from previous period

DVD-A+SACD are now less than 0.1% of CD sales.

Arny Krueger

April 14th 07, 03:32 PM

"Harry Lavo" > wrote in message

> "Arny Krueger" > wrote in message
> ...
>> "William Sommerwerck" > wrote
>> in message
>>>> Has anybody A/B'd a DSD signal with its PCM copy [sic]
>>>> in a vetted test yet? :)
>>>
>>> Actually, a better test would be to compare the DSD
>>> layer of an SACD with its CD layer. This itself would
>>> include two comparisons -- when the CD layer is derived
>>> from the DSD stream (eg, a RCA Living Stereo SACD), and
>>> when the CD layer is a direct PCM transfer (eg, a
>>> Mercury Living Presence SACD).
>>>
>>> I'm slightly embarrassed to say that, though I've had an
>>> SACD player for several years, I haven't made the
>>> comparison. It's clumsy to do well, because the player
>>> cannot simultaneously output both signals. Perhaps I'll
>>> find time over the weekend...
>>
>> The obvious comparison is to transcribe a SACD to a CD
>> with a good audio interface, and them compare the two -
>> level-matched, time-synched, and blind. The hard part
>> is the time-synched part - you have to keep the two
>> discs synched within a fraction of second, or a quick
>> switch will easily identify which is which even if they
>> are actually identical.
>
> If you are doing it blind (especially double blind) why
> would you know which was which?

That's the point - you don't want to. It breaks the blindness of the test.

> You don't need to re-record.

You do if you want to include all of the alleged deliterious effects of the
CD audio format.

However my proposed experiment is very dangerous to you Harry - its actually
something that lot of people can do. You could do it, but of course you
won't.

> You need two identical machines that ears and
> measurements tell you are identical in both SACD and CD
> so far as freqency response f rom 30hz-18khz is concerned
> and which exhibit no other anomalies.

That could work too, if you had recordings that were identical in ever way
but the format.

> You need two identical hybrid sacds...ideally recorded
> directly in DSD with the same stereo mix fed out of the
> DSD workstation.

That's not controllable by most people.

> Most contemporary classical SACDs are
> prepared this way. A separate test using really
> high-quality master tape would be excellent as well as
> yet another source.

That's not available to most people.

> You need to time-synch and double-blind test. Ideally
> it should be a preference test, rather than an identity
> test as work at HK suggests this is easier for people to
> do.
> Ideally you do it across a cross-section of careful and
> experienced listeners well attuned to live classical
> music for a set number of tests each, allowing each to
> control the switching.
> You do it in a casual and comfortable home-like setting,
> with amenities and no pressure and you allow them to stop
> when fatiqued and resume when they are rested and have
> taken a break.
> You sum the responses and apply statistical analysis. When you get over
> 100-150 separate samples from at least
> 10 people, you look at the statistics. For scientific
> "proof" you want 95% reliability (19:1 odds). Great if
> you get 99% (99:1). Still interesting if you get 80%
> (4:1).
> If you want a control (and you should) you provide a
> blind monadic test individually to 100 similar listeners
> for each of the two formats, and ask them to rate the
> music and musical playback quality on a serious of
> semantic differential scales. When finished you then
> apply statistics to the scalers between the two groups of
> 100, and note differences that are statistically
> signifcant at high rates. The overall results should
> support the individual pairs conclusions if both tests
> are valid, and the semantic differential scaling should
> tell you how and why people perceived one as better than
> the other.
> A simple little test....don't know why everybody doesn't
> do it.......

Arny Krueger

April 14th 07, 03:40 PM

"Harry Lavo" > wrote in message

> "Arny Krueger" > wrote in message
> ...
>> "William Sommerwerck" > wrote
>> in message
>>>> Been there done that - the difference is that I dare to
>>>> listen without having my eyes tell me what the *right*
>>>> answer should be.
>>>
>>> No, you have your "science" telling your ears what they
>>> ought to hear.
>>
>> Doesn't matter. Besides, enough listening is done by
>> people who are ignorant of the science so that if there
>> was some hidden truth, it would out.
>
> Arny has said elsewhere that it doesn't matter that ABX
> has no controls against "false negatives" because he
> "knows" he always gives it his best shot to
> differentiate.

That would be yet another one of your petty lies, Harry.

What I did say is that otherwise skeptical people tend to get sucked into
doing their best for listening tests that they might otherwise pooh-pooh.

Arny Krueger

April 14th 07, 03:42 PM

> wrote in message
ups.com
> On Apr 13, 2:24 pm, "Arny Krueger" >
> wrote:
>
>>>>> On Apr 13, 8:09 am, "Arny Krueger" >
>>>>> wrote:
>>
>>>>>> I basically like Dan and I like what he's trying to
>>>>>> do, but I don't see what he is about right now as
>>>>>> having a lot of legs.
>>
>> You're quoting me saying something that I never said.
>> So much for the truth.
>
> bull ****, my qouting was done using the google
> interface, ... you said it!!!
>
>
>> How many hours a week do you listen to live music with a
>> console at hand? I'm setting no records, but for me it
>> averages about 6-8 hours a week. IOW, your negative
>> comments about my lack of interest in listening are
>> completely bogus inventions of your mind.

> four to six hrs a day here no console, no comb
> filtering, 96/24.

Since its your media space you are wasting and your time you are wasting, no
skin of my noise.

> IOW using a 39.95 sacd player as a
> reference is not serious listening, the spdiff cable I
> use cost more than that cheap piece of crap
> and it is a stock item not a tweakoid cable.

Harry Lavo

April 14th 07, 03:45 PM

"Mike Rivers" > wrote in message
ups.com...
> On Apr 14, 7:37 am, "Natalie Drest" >
> wrote:
>

>snip<

>
> I did learn something interesting about wide bandwidth recording from
> the people who make a monstrous sample library with which they
> simulate large pipe organs with multi-channel installations. They
> record each pipe individually rather than recording the blended sound
> of the organ from a distance, then they play back those pipes and let
> them combine naturally in air. What they claim (and can prove by
> looking at the frequency content of their recordings) is that there's
> ultrasonic energy coming off a pipe, and when two pipes are playing
> together, the beating of those ultrasonic frequencies is down in the
> audible range. What you hear sounds different with and without that
> ultrasonic energy.
>

>snip<

Interesting. Mike, my hunch is that this is what Oohashi's subjects
experienced with the recorded gamelan music they used in his ultrasonic
test. The frequency overtones of the gongs, xylophones, and drums extended
to very high frequencies, and so have the potential to create lots of
"beats". As you recall, in that test of both hearing and physiology
subjects recorded (at a statistically significant level) that the music
sounded more pleasant (and their pleasure centers were activated) when the
ultrasonics were included but not when they were excluded. And yet the
ultrasonics reproduced separately from the audible music were not audible.

Now I'll duck before I get flamed for mentioning Oohashi once again.

Harry Lavo

April 14th 07, 03:53 PM

"Arny Krueger" > wrote in message
...
> "Harry Lavo" > wrote in message
>
>> "Arny Krueger" > wrote in message
>> . ..
>>> "William Sommerwerck" > wrote
>>> in message
>>>

>snip<

..
>
>>> The people who are most desperately convinced that they
>>> already know the right answers and don't need any more
>>> stinkin' tests to convince them of anything, are rather
>>> logically and obviously the people who have done the
>>> fewest tests.
>
>>> The people with the most open minds are the people who
>>> did and continute to put their beliefs (which were not
>>> as one-sided as many would like to believe) to the test
>>> the most often and the most recently.
>
> For a good example of people with closed minds, consider the people who
> don't need tests to know that they are right, like Harry Lavo.
>
>> Arny, you simply don't know what you don't know.
>
> You seem to know so much less than I Harry, that it would be good for you
> to first try to learn what I knew back in 1995.
>
>> You might profitably spend some time monitoring the
>> discussions JJ is having on the Audio Pro
>> newslist.....rather than spouting off here about the
>> "known knowns".
>
> Been there, done that. Frankly, that list often makes me sick. It makes
> some of my friends who publish in the JAES sick. There's so much to know
> about audio that leads to such clear and obvious flaws with the work that
> we do, and here's these guys obsessing over problems that are for most if
> not all practically purposes, solved.

That is how "science" moves forward, Arny. You are by temperament an
engineer, not a scientist. That's fine, so long as you recognize the
difference.

>
>> You know the "known knowns" but you
>> don't know the "known unknowns" much less the "unknown
>> unknowns" and the "unknown knowns"). (Yes, that's that
>> damned Rumsfeld, but it is basically a restating of the
>> Johari Window, so I am allowed :-) ).
>
> Dr. cure thy self.

A concept that any self-aware human being recognizes and acknowledges.

Scott Dorsey

April 14th 07, 04:18 PM

William Sommerwerck > wrote:
>
>I see no reason this should shouldn't (sort of) apply to sound reproduction.
>I'd expect a tweeter that was "naturally" flat to 30kHz would "sound better"
>than one that was flat to 20kHz, then dropped off precipitously.

This makes sense in the analogue world where everything is minimum phase.
It's a good idea to have reasonably flat frequency response out to an octave
beyond the frequency of interest, in order to have flat _phase_ response
out to the frequency of interest. (This turns out to be more important on
the bottom end where phase response is more of a big deal, than on the top
end).

In the digital world we don't have such limitations, and we _can_ make
brickwall filters that have no sonic effects in their passband. It sure
took long enough to figure out how to do it right, too. But now that we
CAN finally do it right, there's no reason not to.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Harry Lavo

April 14th 07, 04:31 PM

"Chel van Gennip" > wrote in message
...
> On Sat, 14 Apr 2007 03:23:49 +0200, Harry Lavo wrote:
>
>> * On top of that, the ratio of hi-rez (DVD-A & SACD) to Classical sales
>> is more than twice as high in Holland as in the US (about 17% vs. about
>> 7%) *
>
> There are errors in your spreadsheet. You missed the drop in (DVD-A &
> SACD) sales value of 33% from 2004 to 2005. The result is that all
> classical sales, SACD/DVD-A sales, and the ratio of hi-rez (DVD-A & SACD)
> to Classical dropped.

I've relooked at the spreadsheet, Chel, and I don't see that I've "missed"
anything. The data is there for all three years, 2003, 2004, and 2005.
Clearly there was a drop from 2003 to 2004. But a volume drop needs a
reference point, so I looked at where 2005 sales were (our latest year of
hard data, and most relevant since we are talking about where hi-rez is in
2006 and 2007). To do this I looked at 2005 sales against the peak year
(2003) and against the prior year (2004). This wasn't meant to be
misleading at all..but any good analyst knows that to truly understand a
sales trend one needs to look at multiple reference points.

Unless you can point to other specifics, it looks to me like all the factual
data is correct...I think what you are objecting to is that I didn't dwell
on the drop from 2003 to 2004. So let me be clear...all segments except US
dollar sales fell 2004 vs. 2003....total market in Holland, Classical in
Holland (US not known), and hi-rez in both Holland and US. But the data
shows that things then stabilized at least in relative volume and share, so
that as of 2005 sales were 40-45% off peak year 2003 for Classical in
Holland and for hi-rez in both markets, but relative share of hi-rez to
classical essentially held up. As I pointed out, when comparing sales
"after the fall" hi-rez was off only modestly in the US (-8.2% same as toal
market) and actually grew in Holland (+3.4%) . Clearer now?

The following summarizes share data, in order to help others as well as
yourself:

Segment % Share/2003/2004/2005

Hi-Rez % Market - Holland/1.63/0.94/1.13
Classical % Market - Holland/8/7/6
Hi-Rez % Classical - Holland/20.4/13.5/18.8

Hi-Rez % Market - US/0.31/0.20/0.20
Classical % Market ) - US(est)/3/3/3
Hi-Rez % Classical - US(est)/10.3/6.7/6.7

Harry Lavo

April 14th 07, 04:35 PM

"Scott Dorsey" > wrote in message
...
> William Sommerwerck > wrote:
>>
>>I see no reason this should shouldn't (sort of) apply to sound
>>reproduction.
>>I'd expect a tweeter that was "naturally" flat to 30kHz would "sound
>>better"
>>than one that was flat to 20kHz, then dropped off precipitously.
>
> This makes sense in the analogue world where everything is minimum phase.
> It's a good idea to have reasonably flat frequency response out to an
> octave
> beyond the frequency of interest, in order to have flat _phase_ response
> out to the frequency of interest. (This turns out to be more important on
> the bottom end where phase response is more of a big deal, than on the top
> end).
>
> In the digital world we don't have such limitations, and we _can_ make
> brickwall filters that have no sonic effects in their passband. It sure
> took long enough to figure out how to do it right, too. But now that we
> CAN finally do it right, there's no reason not to.

Except, according to the discussion on the Pro Audio listserve, even that
conclusion is suspect as it ignores the short "windowing" timing bandwidth
of our hearing which can in turn create transients that extend down to DC
instead of the static phase picture we are used to seeing.

[email protected]

April 14th 07, 04:41 PM

On Apr 14, 10:17 am, "Arny Krueger" > wrote:

> Dr. cure thy self.

damn Arny,
if you only spent as much time and energy listening to music
as you do hyping your dogma...

hank alrich

April 14th 07, 04:46 PM

Arny Krueger wrote:

> > You might profitably spend some time monitoring the
> > discussions JJ is having on the Audio Pro
> > newslist.....rather than spouting off here about the
> > "known knowns".
>
> Been there, done that. Frankly, that list often makes me sick. It makes some
> of my friends who publish in the JAES sick. There's so much to know about
> audio that leads to such clear and obvious flaws with the work that we do,
> and here's these guys obsessing over problems that are for most if not all
> practically purposes, solved.

In fact those guys are among those who have solved many problem by
"obsessing" with details others thought not worthy of detailed
examination. If you can't hang with JJ it's because you can't hang with
JJ. His knowledge runs deeper than that of most self-proclaimed audio
experts. It's possible nobody knows as much about _hearing_ as he does.

They obsess because they care and they are not given to assuming all the
**** has been figured out yet. And those cats have doen a lot of
figuring-out.

--
ha
Iraq is Arabic for Vietnam

hank alrich

April 14th 07, 04:49 PM

Chel van Gennip wrote:

> When I buy lamps, I especialy look for the ones with flat output in
> infrared and utraviolet, to see better greens and reds.

Funny, but you know it's not a relevant reply.

--
ha
Iraq is Arabic for Vietnam

Harry Lavo

April 14th 07, 05:00 PM

"Arny Krueger" > wrote in message
...
> "Harry Lavo" > wrote in message
>
>> "William Sommerwerck" > wrote
>> in message
>> . ..
>>>>> I don't know the source you are quoting, but it is
>>>>> important that it include web sources, since virtually
>>>>> all SACD sales and most DVD-A sales have moved to the
>>>>> web.
>>>
>>> I recently saw a statement that the total number of SACD
>>> sales last year was
>>> about one in a thousand of all optical-media audio
>>> sales. That's not so hot.
>>>
>>> Of course, it doesn't matter how many are sold, as long
>>> as the record companies can at least break even on the
>>> ones they sell.
>
>> Two-tenths of one percent...that's more like
>> one-in-five-hundred. But it was steady from 2004 to
>> 2005...I'm anxious to see the 2006 numbers. Incidentally,
>> if you read my other post in response to Chel, you'll see
>> it is five times higher in Holland, or one-in-a-hundred.
>
> Time to wake up and smell the coffee, Harry. The numbers for the first
> half of 2006 have been out for over 6 months.
>
> Announcement:
>
> http://www.riaa.com/news/newsletter/101206.asp
>
> Actual numbers:
>
> http://www.riaa.com/news/newsletter/pdf/2006midYrStats.pdf
>
> SACD > 44% down from previous period.
> DVD-A > 35% down from previous period
>
> DVD-A+SACD are now less than 0.1% of CD sales.

Your numbers are correct directionally, but not factually for dollar sales,
which is the measure I used as most universal.

Actual combined sales volume fell -38.4% vs a market decline of -16.0%
Hi-Rez Share dropped from 0.16% to 0.12%.

However, there are a few caveats to beware of before jumping to conclusions:

* 2006 saw Best Buy and Circuit City drop SACD and DVD-A, which could
account for much of the decline. If so, it may be a one time drop.
* Please note that only 40% of the year's recorded music volume is done in
first half versus 60% in second half.
* Please note also that hi-rez's first half 2005 share of 0.16% was below
full year share of 0.20%, meaning hi-rez's relative performance during
second half was much better than in the first half in 2005. The same may
well be true in 2007 and it is not inconceivable that the lost share might
be made up, especially if the first half reflected "de-inventorying" of the
big box stores.

Stay tuned.

Harry Lavo

April 14th 07, 05:07 PM

"Arny Krueger" > wrote in message
. ..
> "Harry Lavo" > wrote in message
>
>> "Arny Krueger" > wrote in message
>> ...
>>> "William Sommerwerck" > wrote
>>> in message
>>>>> Has anybody A/B'd a DSD signal with its PCM copy [sic]
>>>>> in a vetted test yet? :)
>>>>
>>>> Actually, a better test would be to compare the DSD
>>>> layer of an SACD with its CD layer. This itself would
>>>> include two comparisons -- when the CD layer is derived
>>>> from the DSD stream (eg, a RCA Living Stereo SACD), and
>>>> when the CD layer is a direct PCM transfer (eg, a
>>>> Mercury Living Presence SACD).
>>>>
>>>> I'm slightly embarrassed to say that, though I've had an
>>>> SACD player for several years, I haven't made the
>>>> comparison. It's clumsy to do well, because the player
>>>> cannot simultaneously output both signals. Perhaps I'll
>>>> find time over the weekend...
>>>
>>> The obvious comparison is to transcribe a SACD to a CD
>>> with a good audio interface, and them compare the two -
>>> level-matched, time-synched, and blind. The hard part
>>> is the time-synched part - you have to keep the two
>>> discs synched within a fraction of second, or a quick
>>> switch will easily identify which is which even if they
>>> are actually identical.
>>
>> If you are doing it blind (especially double blind) why
>> would you know which was which?
>
> That's the point - you don't want to. It breaks the blindness of the test.

Not in an A-B preference test it doesn't (it would in ABX). Since you don't
know which is which, all you know is that the sample has changed...in an AB
preference test you have to know that anyway.

>
>> You don't need to re-record.
>
> You do if you want to include all of the alleged deliterious effects of
> the CD audio format.

No you don't. All you need to know that there was no mastering compression
or eq applied to the CD layer of the hybrid, same as the SACD layer, and
that they were derived from the main master source.

>
> However my proposed experiment is very dangerous to you Harry - its
> actually something that lot of people can do. You could do it, but of
> course you won't.

So is mine. The only difference is you have to have two identical SACDs.
With yours, you still need two identical machines. And if it is ABX, an ABX
switchbox as well.

>
>> You need two identical machines that ears and
>> measurements tell you are identical in both SACD and CD
>> so far as freqency response f rom 30hz-18khz is concerned
>> and which exhibit no other anomalies.
>
> That could work too, if you had recordings that were identical in ever way
> but the format.

Glad you finally see the light.

>
>> You need two identical hybrid sacds...ideally recorded
>> directly in DSD with the same stereo mix fed out of the
>> DSD workstation.
>
> That's not controllable by most people.

But they are available....several.

>
>> Most contemporary classical SACDs are
>> prepared this way. A separate test using really
>> high-quality master tape would be excellent as well as
>> yet another source.
>
> That's not available to most people.

But they are available....several.

>
>> You need to time-synch and double-blind test. Ideally
>> it should be a preference test, rather than an identity
>> test as work at HK suggests this is easier for people to
>> do.
>> Ideally you do it across a cross-section of careful and
>> experienced listeners well attuned to live classical
>> music for a set number of tests each, allowing each to
>> control the switching.
>> You do it in a casual and comfortable home-like setting,
>> with amenities and no pressure and you allow them to stop
>> when fatiqued and resume when they are rested and have
>> taken a break.
>> You sum the responses and apply statistical analysis. When you get over
>> 100-150 separate samples from at least
>> 10 people, you look at the statistics. For scientific
>> "proof" you want 95% reliability (19:1 odds). Great if
>> you get 99% (99:1). Still interesting if you get 80%
>> (4:1).
>> If you want a control (and you should) you provide a
>> blind monadic test individually to 100 similar listeners
>> for each of the two formats, and ask them to rate the
>> music and musical playback quality on a serious of
>> semantic differential scales. When finished you then
>> apply statistics to the scalers between the two groups of
>> 100, and note differences that are statistically
>> signifcant at high rates. The overall results should
>> support the individual pairs conclusions if both tests
>> are valid, and the semantic differential scaling should
>> tell you how and why people perceived one as better than
>> the other.
>> A simple little test....don't know why everybody doesn't
>> do it.......

No comment on the real complexity of doing a truly scientific test, I guess?

Harry Lavo

April 14th 07, 05:08 PM

"Arny Krueger" > wrote in message
. ..
> "Harry Lavo" > wrote in message
>
>> "Arny Krueger" > wrote in message
>> ...
>>> "William Sommerwerck" > wrote
>>> in message
>>>>> Been there done that - the difference is that I dare to
>>>>> listen without having my eyes tell me what the *right*
>>>>> answer should be.
>>>>
>>>> No, you have your "science" telling your ears what they
>>>> ought to hear.
>>>
>>> Doesn't matter. Besides, enough listening is done by
>>> people who are ignorant of the science so that if there
>>> was some hidden truth, it would out.
>>
>> Arny has said elsewhere that it doesn't matter that ABX
>> has no controls against "false negatives" because he
>> "knows" he always gives it his best shot to
>> differentiate.
>
> That would be yet another one of your petty lies, Harry.
>
> What I did say is that otherwise skeptical people tend to get sucked into
> doing their best for listening tests that they might otherwise pooh-pooh.

You said that in addition to saying you always gave your best.

jwvm

April 14th 07, 05:49 PM

On Apr 14, 1:45 am, wrote:
> Harry Lavo > wrote:
> > "Mike Rivers" > wrote in message
> ups.com...
> > > On Apr 13, 8:09 am, "Arny Krueger" > wrote:
>
> > >> I read what Dan said here about them a little differently. And, if he
> > >> thought the issues through from scratch today, he might say things a
> > >> little
> > >> differently than he did the last time he was posting a lot here.
>
> > > I'm not going to fight Lavry's battle for him, but I suspect that it
> > > would be hard to find a 192 kHz converter that sounds as good as his
> > > 96 kHz converter. SOUNDS being the operative word here. Here's a
> > > relevant portion from his 2004 Sampling Rate paper:
> > >snip Lavry's report, irrelevant to my comment below<
> > He and JJ and others have been having a very interesting discussion on the
> > Pro Audio listserve since before March 1 on other aspects of the problem
> > (they agree, btw, that from a practical standpoint, 64khz/20bits is an
> > adequate standard for utter transparency in final reproduction....basically
> > the same conclusion as reached by the hi-rez industry group in Japan back in
> > the mid-90's.) Their current discussion focuses on how the sliding time
> > window of our hearing sensitivity wrecks havoc with the filter
> > characteristics of even the most benign filter at 20khz...offering yet
> > another potential distortion that could be and clearly may be audible when
> > it comes to the conventional CD cutoff. They've also discussed the
> > pre-ringing of PCM filters, and JJ at least, feels that under some
> > circumstances it cleary could be audible.
>
> And those circumstances were clearly and plainly stated to be very narrowly
> defined test signals. Yet you said in RAHE that they were "often" audible
> in a context that seemed to assume general conditions, since you have
> earlier written many times that you think testing with defined test signals is
> irrelevent for music listening. If they were "often" audible with music,
> they would certainly recomend to use it as a test signal. But they don't!
>
> You wanted to know where you misrepresented what they were talking about and
> that is it. Please stop it.

RAHE is famous for spiraling claims and counterclaims regarding magic
LPs, toobs, green markers, $1000 speaker wire/interconnects/power
cords and of course SACDs. Even fairly trivial test suggestions for
proof are routinely ignored by the subjectivists while they make
outlandish claims of phenomenal improvement with their cause de jour.
There is precious little light from these rantings but much smoke and
mirrors, much of it produced by the snake-oil merchants. Scott
Dorsey's comment implying that there is nothing remarkable or
revolutionary about SACD recordings is very reasonable. Given his
stature, experience and expertise, it is not clear why a few posters
(not William!) continue to propagate the RAHE rantings.

William Sommerwerck

April 14th 07, 06:11 PM

>> If price weren't an object, and you could use whatever technology you
>> wanted... Would you design a tweeter that was flat to 20kHz, or flat to
>> 28kHz?

> Maybe price is not an object, physics always is. Any driver is a
> compromise. If you want to get a flat response up to 30kHz, you will have
> to sacrifice below 20kHz somewhere. For me any sacrifice in the audible
> range to get better performance in a not audible range is a bad
> compromise.

You don't know what you're talking about. There is no law of nature that
constrains tweeter design in the way you suggest. You're just spewing words.

William Sommerwerck

April 14th 07, 06:12 PM

> >> When I buy lamps, I especialy look for the ones with flat output in
> >> infrared and utraviolet, to see better greens and reds.
> >
> > Funny, but you know it's not a relevant reply.
>
> Using invisible light to see things is as stupid as using unaudible sound
> to hear things.

Neither of you bothered to read my response. Or if you did, you didn't get
it.

hank alrich

April 14th 07, 06:41 PM

Chel van Gennip wrote:

> If you want to get a flat response up to 30kHz, you will have
> to sacrifice below 20kHz somewhere. For me any sacrifice in the audible
> range to get better performance in a not audible range is a bad
> compromise.

So you prefer single-driver speakers, eh?

--
ha
Iraq is Arabic for Vietnam

hank alrich

April 14th 07, 06:41 PM

William Sommerwerck wrote:

> > >> When I buy lamps, I especialy look for the ones with flat output in
> > >> infrared and utraviolet, to see better greens and reds.
> > >
> > > Funny, but you know it's not a relevant reply.
> >
> > Using invisible light to see things is as stupid as using unaudible sound
> > to hear things.
>
> Neither of you bothered to read my response. Or if you did, you didn't get
> it.

I did and I got it.

--
ha
Iraq is Arabic for Vietnam

Les Cargill

April 14th 07, 06:51 PM

William Sommerwerck wrote:

>>Neither are a lot of other people, particularly since transducers
>>(microphones and loudspeakers) that have useful output above
>>20 kHz aren't very common. But they exist, and recordings of certain
>>instruments have shown that there is considerable energy above the
>>human range of hearing. That doesn't mean that we should all do it all
>>the time, but there's something there that might be useful some day.
>
>
> When designing bridges or skyscrapers, one leaves a substantial margin --
> usually 100% -- for error. That is, if you think a structure will need to
> support only 10,000 pounds, you nevertheless design it for 20,000.
>
> I see no reason this should shouldn't (sort of) apply to sound reproduction.
> I'd expect a tweeter that was "naturally" flat to 30kHz would "sound better"
> than one that was flat to 20kHz, then dropped off precipitously.
>
>

That's fine, so long as the additional bandwidth for the tweeter
doesn't cost too much to make it useable, or negatively
affect the known audible range.

--
Les Cargill

William Sommerwerck

April 14th 07, 07:18 PM

>> I see no reason this should shouldn't (sort of) apply to sound
>> reproduction. I'd expect a tweeter that was "naturally" flat to
>> 30kHz would "sound better" than one that was flat to 20kHz,
>> then dropped off precipitously.

> That's fine, so long as the additional bandwidth for the tweeter
> doesn't cost too much to make it useable, or negatively
> affect the known audible range.

You're missing the engineering point, here. The idea is not to extend the
bandwidth for its own sake, but so that the driver "behaves" better within
the audible band. Bandwidth and "behavior" are not antithetical, but
complementary.

William Sommerwerck

April 14th 07, 07:20 PM

>> You don't know what you're talking about. There is no law of nature that
>> constrains tweeter design in the way you suggest. You're just spewing
>> words.

> One of the laws involved is f=m.a that law limits objects to move as a
> whole at high frequencies.

I'll leave it to a physics major to explain why F=ma has nothing to do with
driver breakup.

Scott Dorsey

April 14th 07, 08:35 PM

hank alrich > wrote:
>Arny Krueger wrote:
>> > You might profitably spend some time monitoring the
>> > discussions JJ is having on the Audio Pro
>> > newslist.....rather than spouting off here about the
>> > "known knowns".
>>
>> Been there, done that. Frankly, that list often makes me sick. It makes some
>> of my friends who publish in the JAES sick. There's so much to know about
>> audio that leads to such clear and obvious flaws with the work that we do,
>> and here's these guys obsessing over problems that are for most if not all
>> practically purposes, solved.
>
>In fact those guys are among those who have solved many problem by
>"obsessing" with details others thought not worthy of detailed
>examination. If you can't hang with JJ it's because you can't hang with
>JJ. His knowledge runs deeper than that of most self-proclaimed audio
>experts. It's possible nobody knows as much about _hearing_ as he does.

Lots of things that everyone thought were solved turn out not to be,
though. On the other hand, some things turn out to be much more solved
than thought. Check out the latest issue of the JAES with the paper on
audibility of group delay.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Harry Lavo

April 14th 07, 08:49 PM

"Chel van Gennip" > wrote in message
...
> On Sat, 14 Apr 2007 17:31:04 +0200, Harry Lavo wrote:
>
>
>>> There are errors in your spreadsheet. You missed the drop in (DVD-A &
>>> SACD) sales value of 33% from 2004 to 2005. The result is that all
>>> classical sales, SACD/DVD-A sales, and the ratio of hi-rez (DVD-A &
>>> SACD) to Classical dropped.
>>
>> I've relooked at the spreadsheet, Chel, and I don't see that I've
>> "missed" anything. The data is there for all three years, 2003, 2004,
>> and 2005.
>
> In the NVPI figures the SACD/DVD-A sales drop from E 2.9M in 2004 to E 2M
> in 2005. In your spreadsheet they rise to E 3M instead.

Okay, thanks for the correction. I thought you were speaking of the RIAA
numbers. You are correct, I misread the 3 for a 2 in the Holland numbers.
Interesting, that gets rid of the anomaly of the +3.4% growth...sales of
hi-rez in Europe fell instead exactly in line with classical sales, and
share ratio stayed pretty much unchanged in Holland 2004-to-2005 just as it
did in the US. The revised share numbers are below:

Segment % Share/2003/2004/2005

Hi-Rez % Market - Holland/1.63/0.94/0.75
Classical % Market - Holland/8/7/6
Hi-Rez % Classical - Holland/20.4/13.5/12.5

Hi-Rez % Market - US/0.31/0.20/0.20
Classical % Market ) - US(est)/3/3/3
Hi-Rez % Classical - US(est)/10.3/6.7/6.7

So based on this, hi-rez in Holland is about 13% of classical sales; hi-rez
in the US is about 6.5-7% of classical sales. While classical sales in
Holland are 6-7% of total market, versus a reputed 3% of total market in the
US. Meaning hi-rez sales in Holland are roughly four times as important as
in the US. Relatively stable during 2004 and 2005 in share ratio but
declining in share of overall market in proportion to classical sales. The
first half 2006 numbers suggest (but cannot predict) that that absolute
decline may be continuing, while saying nothing about the relative stability
of share ratio.

The implications are perhaps different than the assumptions that started
this exchange. Either hi-rez sales are declining because the market for
classical music is declining, or (more likely) both markets are declining
because of common factors. Such factors might be online purchase or
downloading (although unlikely for these segments) or decreasing
brick-and-mortar distribution (very likely) or something else.

Chris Hornbeck

April 15th 07, 02:41 AM

On Sat, 14 Apr 2007 22:46:54 +0200, Chel van Gennip
> wrote:

>> why F=ma has nothing to do with driver breakup.
>
>because you don't understand that from the point where the signal is
>converted into a current in a driver coil, until the moment your nervous
>system detects the sound, Newtons 2nd law is one of the most important
>factors.

Another way of discussing the problem is to generalize to
say that drivers of a particular topo have a particular
fixed bandwidth. For example, a dome tweeter with some particular
construction will have a good practical (define parameters here)
bandwidth of 2000 Hertz to 20000 Hertz.

One decade is a pretty good bandwidth if the definition of "good"
is rigorous and we're talking about conventional dynamic drivers.
It's about three octaves and better than we expect for things
like a car, which even has a transmission (multi-match transformer),
for instance.

And, generally, to push up the upper resonances of a driver
requires reducing their physical size, pushing up their lowest
useful frequency somewhat-more-or-less proportionately.

Size, mass, dimensions, radiating pattern are all tradeoffs,
but within a particular topo, bandwidth is pretty tightly
constrained to a fixed limit. To move the upper limit of a
constained topo from 20000 to 28000 Hertz will require
moving the lower limit from 2000 to 2800 Hertz along with it,
more-or-less.

Personally, crossovers to midrange drivers anywhere in this
half octave work fine for nearfield (1 M) listening if vertical
response is symmetrical, (and the acoustical crossover third order
Butterworth, arf!). So, for me, it's do-able. Others might find
otherwise, or have better drivers available to comment about.

Much thanks to all,

Chris Hornbeck

"Beauty will save the world."
- Feodor Dostoevsky

"Information is not knowledge. Knowledge is not wisdom. Truth is not
beauty. Beauty is not love. Love is not music."
- Frank Zappa

Arny Krueger

April 15th 07, 11:49 AM

"Natalie Drest" > wrote in
message

> "If you're asking, "Why would I want to record
> frequencies no one can hear?" you're not alone. One
> reason is that frequencies we can't hear as pure tones
> may nevertheless audibly influence the color of complex
> tones."

> I'm not sold on this idea...

Hold that thought. It reverses reality.

Frequencies we can hear as pure tones can fail to audibly influence the
color of complex sounds. The reason is simple, well-known and widely
exploited - perceptual masking.

> "In particular, the character of transients may be
> influenced by components whose frequencies are above the
> official range of our hearing."

See my previous statement. The human ear stops interpreting the phase of
components of sounds starting at 1 KHz or below. Below 500 Hz-1KHz the ear
actually follows wave form shapes. Above 500 Hz-1KHz the ear functions like
a spectrum analyzer, and decodes the amplitude of frequency components of
sounds.

> "Another point in favor of high sampling rates is that
> timing differences smaller than 1/44,100 of a second may
> affect our ability to localize sounds.

The idea that digital audio signals can't reproduce timing differences
smaller than one sample period is a myth.

Digital audio signals can easily reproduce timing differences far, far
smaller than one sample period. Digital audio signals can resolve timing
differences at least as small as one sample period divided by the number of
coded amplitude levels. In the case of the CD format this is about
1/(44,100*65,536) of a second.

Arny Krueger

April 15th 07, 11:56 AM

"Chel van Gennip" > wrote in message

> On Sat, 14 Apr 2007 14:18:32 +0200, Scott Dorsey wrote:
>
>> Absolutely, you should definitely understand the Nyquist
>> theorem if you don't already. Nobody debates the
>> accuracy of it, but what they do debate is whether
>> additional frequency response is useful or not. I am
>> inclined these days to say it isn't, but I wouldn't bet
>> my life on it.
>
> As this is as pro(duction) group, it is possible to do
> simple tests: compare input and output. If the output of
> an AD-DA chain is better or worse than the input, the
> device is wrong.

This is the fact that makes fools out of the advocates of so-called hi-rez
formats.

If you take the output of a SACD of DVD-A player playing music, and digitize
it and convert it back to analog with a good ADC/DAC box such as M-Audio's
Flying Cow operating at its 44 KHz sample rate, the interposition of this
device, or bypassing it with a straight wire, cannot be reliably detected
provided you do a good job of matching levels and addressing listener bias.

We first did a similar experiment back in the late 1970s with a comparable
device of the day, and reported the results here:

http://www.provide.net/~djcarlst/abx_digi.htm

From time to time we've repeated this kind of experiement, and invite others
to do the same.

No known exceptions exist.

I challenge *all* of the RAP advocates of hi-rez formats to put their
reputations on the line, and repeat this simple experiment.

Arny Krueger

April 15th 07, 11:59 AM

"William Sommerwerck" > wrote in
message
>> Neither are a lot of other people, particularly since
>> transducers (microphones and loudspeakers) that have
>> useful output above 20 kHz aren't very common. But they
>> exist, and recordings of certain instruments have shown
>> that there is considerable energy above the human range
>> of hearing. That doesn't mean that we should all do it
>> all the time, but there's something there that might be
>> useful some day.

> When designing bridges or skyscrapers, one leaves a
> substantial margin -- usually 100% -- for error. That is,
> if you think a structure will need to support only 10,000
> pounds, you nevertheless design it for 20,000.

The difference is that a record/reproduce format does not irrevocably break
and result in the loss of life and property should it fail to reproduce some
sonic nuance. The good news is that 44/16 is in fact an overkill format, and
does include more than enough safety margins, so that no musical nuances are
lost.

Arny Krueger

April 15th 07, 12:06 PM

> wrote in message
oups.com
> On Apr 14, 10:17 am, "Arny Krueger" >
> wrote:
>
>> Dr. cure thy self.

> damn Arny,

> if you only spent as much time and energy listening to
> music as you do hyping your dogma...

This isn't about dogma, its about observable facts.

Let us know when you do that magical blind listening test showing that
down-coding one of your 24/96 recordings to 44/16 has some audible effect
in a well-designed listening test.

Arny Krueger

April 15th 07, 12:07 PM

"hank alrich" > wrote in message

> Arny Krueger wrote:
>
>>> You might profitably spend some time monitoring the
>>> discussions JJ is having on the Audio Pro
>>> newslist.....rather than spouting off here about the
>>> "known knowns".
>>
>> Been there, done that. Frankly, that list often makes me
>> sick. It makes some of my friends who publish in the
>> JAES sick. There's so much to know about audio that
>> leads to such clear and obvious flaws with the work that
>> we do, and here's these guys obsessing over problems
>> that are for most if not all practically purposes,
>> solved.
>
> In fact those guys are among those who have solved many
> problem by "obsessing" with details others thought not
> worthy of detailed examination.

Name one such thing, and I'll show you the prior art.

William Sommerwerck

April 15th 07, 12:14 PM

>>> why F=ma has nothing to do with driver breakup.

>> because you don't understand that from the point where the
>> signal is converted into a current in a driver coil, until the
>> moment your nervous system detects the sound, Newton's
>> 2nd law is one of the most important factors.

Where do you get this crap?

> Another way of discussing the problem is to generalize to
> say that drivers of a particular topo have a particular fixed
> bandwidth. For example, a dome tweeter with some particular
> construction will have a good practical (define parameters here)
> bandwidth of 2000 hertz to 20000 hertz.

Construction and topology are not the same thing. What about the driver
material?

Your (and I'm talking to several of you) ignorance of driver design is
abysmal -- and I don't consider _myself_ an expert on the subject.

Consider the famous KEF B200 driver. (Duh...? What's that?) It was one of
the first high-quality plastic drivers. The material used had
better-controlled breakup properties than paper, allowing KEF to move the
upper breakup frequency to 3kHz. This permitted the driver to be used in a
two-way system (rather than requiring a separate midrange). Extending the
upper bandwidth had no effect on the low-frequency behavior of the speaker.

What a waste of time to see ignorant people discussing stuff they have
little knowledge of. Being able to quote F = ma doesn't make you a
physicist, any more than saying e = mc^2 means you understand relativity
(and in fact, that equation is the least-important part of Einsteinian
relativity).

If you're seriously interested in speaker design, read Martin Colloms'
"High-Performance Loudspeakers". The title _should_ be "High-Performance
Electrodynamic Loudspeakers", as he gives little space to electrostatic and
planar-magnetic systems (the latter of which _are_ electrodynamic, but give
me a break, huh?).

Arny Krueger

April 15th 07, 12:20 PM

"Harry Lavo" > wrote in message

> "Arny Krueger" > wrote in message
> ...
>> "Harry Lavo" > wrote in message
>>
>>> "William Sommerwerck" > wrote
>>> in message
>>> . ..
>>>>>> I don't know the source you are quoting, but it is
>>>>>> important that it include web sources, since
>>>>>> virtually all SACD sales and most DVD-A sales have
>>>>>> moved to the web.
>>>>
>>>> I recently saw a statement that the total number of
>>>> SACD sales last year was
>>>> about one in a thousand of all optical-media audio
>>>> sales. That's not so hot.
>>>>
>>>> Of course, it doesn't matter how many are sold, as long
>>>> as the record companies can at least break even on the
>>>> ones they sell.
>>
>>> Two-tenths of one percent...that's more like
>>> one-in-five-hundred. But it was steady from 2004 to
>>> 2005...I'm anxious to see the 2006 numbers.
>>> Incidentally, if you read my other post in response to
>>> Chel, you'll see it is five times higher in Holland, or
>>> one-in-a-hundred.
>>
>> Time to wake up and smell the coffee, Harry. The numbers
>> for the first half of 2006 have been out for over 6
>> months. Announcement:
>>
>> http://www.riaa.com/news/newsletter/101206.asp
>>
>> Actual numbers:
>>
>> http://www.riaa.com/news/newsletter/pdf/2006midYrStats.pdf
>>
>> SACD > 44% down from previous period.
>> DVD-A > 35% down from previous period
>>
>> DVD-A+SACD are now less than 0.1% of CD sales.
>
> Your numbers are correct directionally, but not factually
> for dollar sales, which is the measure I used as most
> universal.
> Actual combined sales volume fell -38.4% vs a market
> decline of -16.0% Hi-Rez Share dropped from 0.16% to
> 0.12%.

> However, there are a few caveats to beware of before
> jumping to conclusions:

> * 2006 saw Best Buy and Circuit City drop SACD and DVD-A,
> which could account for much of the decline.

As usual Harry, you are more than a little fact-challenged.

I've seen SACDs amd DVD-As on sale in Best Buy stores in 2007 - in the past
few weeks.

Regrettably, the local Circuit City store closed.

From the best Buy Web site today:

http://www.bestbuy.com/site/olspage.jsp?id=cat12077&categoryRep=cat02000&guideID=1054508946422&type=page

"Best Buy also carries a range of DVD-Audio titles, with many more soon to
hit our shelves. You'll find a variety of genres, including rock, R&B,
classical and jazz. Many of your favorite older albums have been remixed and
remastered to bring you closer than ever to the original recording. There
are also many new releases that have been recorded with DVD-Audio in mind
during the creative process — providing for even more sonic definition!"

"Take a visit to your local Best Buy for the players, surround sound systems
and other gear that'll deliver a peak SACD listening experience. Best Buy
also carries a range of SACD titles, with many new discs on the way. You'll
find a variety of genres, including rock, R&B, classical and jazz. Many of
your favorite older albums have been remixed and remastered to bring you
closer than ever to the original recording (including the recently released
hybrid SACD/CD version of Pink Floyd's Dark Side of the Moon). There are
also many new releases that have been recorded with SACD in mind during the
creative process — promising an astounding surround sound experience.

Searching on "sacd" on the web site finds the following

http://www.bestbuy.com/site/olspage.jsp?_dyncharset=ISO-8859-1&id=pcat17071&type=page&st=sacd&sc=Global&cp=1&nrp=15&sp=&qp=&list=n&iht=y&usc=All+Categories&ks=960

703 items are found including numerous recordings.

Arny Krueger

April 15th 07, 12:23 PM

"Harry Lavo" > wrote in message

>>> You need to time-synch and double-blind test.

Harry you're preaching to the choir and you forgot about leve-matching.

>> > Ideally
>>> it should be a preference test, rather than an identity
>>> test as work at HK suggests this is easier for people to
>>> do.

As usual Harry, you're taking gross liberties with what Harmon's people have
written about the matter.

>>> Ideally you do it across a cross-section of careful and
>>> experienced listeners well attuned to live classical
>>> music for a set number of tests each, allowing each to
>>> control the switching.

Been there, done that.

>>> You do it in a casual and comfortable home-like setting,
>>> with amenities and no pressure and you allow them to
>>> stop

Been there, done that.

>>> when fatiqued and resume when they are rested and have
>>> taken a break.

Doooh!

>>> You sum the responses and apply statistical analysis.
>>> When you get over 100-150 separate samples from at least
>>> 10 people, you look at the statistics. For scientific
>>> "proof" you want 95% reliability (19:1 odds). Great if
>>> you get 99% (99:1). Still interesting if you get 80%
>>> (4:1).

Been there, done that.

>>> If you want a control (and you should) you provide a
>>> blind monadic test individually to 100 similar listeners
>>> for each of the two formats, and ask them to rate the
>>> music and musical playback quality on a serious of
>>> semantic differential scales.

Monadic testing is a hypothesis of Harry Lavo, not an accepted blind testing
protocol.

>>> When finished you then
>>> apply statistics to the scalers between the two groups
>>> of 100, and note differences that are statistically
>>> signifcant at high rates. The overall results should
>>> support the individual pairs conclusions if both tests
>>> are valid, and the semantic differential scaling should
>>> tell you how and why people perceived one as better than
>>> the other.

After you, Harry.

>>> A simple little test....don't know why everybody doesn't
>>> do it.......

> No comment on the real complexity of doing a truly
> scientific test, I guess?

Been there, done that unlike you Harry.

Arny Krueger

April 15th 07, 12:25 PM

"Harry Lavo" > wrote in message

> "Arny Krueger" > wrote in message
> . ..
>> "Harry Lavo" > wrote in message
>>
>>> "Arny Krueger" > wrote in message
>>> ...
>>>> "William Sommerwerck" >
>>>> wrote in message
>>>>
>>>>>> Been there done that - the difference is that I dare
>>>>>> to listen without having my eyes tell me what the
>>>>>> *right* answer should be.
>>>>>
>>>>> No, you have your "science" telling your ears what
>>>>> they ought to hear.
>>>>
>>>> Doesn't matter. Besides, enough listening is done by
>>>> people who are ignorant of the science so that if there
>>>> was some hidden truth, it would out.
>>>
>>> Arny has said elsewhere that it doesn't matter that ABX
>>> has no controls against "false negatives" because he
>>> "knows" he always gives it his best shot to
>>> differentiate.
>>
>> That would be yet another one of your petty lies, Harry.

>> What I did say is that otherwise skeptical people tend
>> to get sucked into doing their best for listening tests
>> that they might otherwise pooh-pooh.

> You said that in addition to saying you always gave your
> best.

As usual, the actual quote should you ever provide it in context, is vastly
different from what you initially presented, Harry.

Arny Krueger

April 15th 07, 12:26 PM

"William Sommerwerck" > wrote in
message
>>> You don't know what you're talking about. There is no
>>> law of nature that constrains tweeter design in the way
>>> you suggest. You're just spewing words.
>
>> One of the laws involved is f=m.a that law limits
>> objects to move as a whole at high frequencies.
>
> I'll leave it to a physics major to explain why F=ma has
> nothing to do with driver breakup.

Sorry William, but every analysis of driver breakup I've ever seen or done
includes F=MA.

No mass, no accelleration, no breakup.

William Sommerwerck

April 15th 07, 12:28 PM

> We first did a similar experiment back in the late 1970s with a
> comparable device of the day, and reported the results here:

> http://www.provide.net/~djcarlst/abx_digi.htm

> From time to time we've repeated this kind of experiement, and invite
> others to do the same.

Why should I perform a listening test that, in my opinion, has no
correlation with the way people normally listen?

How many times do we need to explain to you, Arny, that you're _assuming_
the validity of the experiment a priori? That's not science.

William Sommerwerck

April 15th 07, 01:27 PM

>> I'll leave it to a physics major to explain why F=ma has
>> nothing to do with driver breakup.

> Sorry, William, but every analysis of driver breakup I've ever
> seen or done includes F=ma.

This is like saying that because F=ma is part of the explanation of why eggs
crack when hit on the edge of a bowl, that F=ma explains cake-baking.

Nick's Picks

April 15th 07, 01:42 PM

wow..
now this is a thread hijack!

good reading though. I didn't mean it to be a format debate, but good
reading none the less.
DSD, as I understand it, is not entirely Sony's baby. Also a japonese
professor (who liscened it to Korg), Ed Meitner also had something to
do w/it (?), or just makes it oh so nice.

the comment of "take your SACDs and run it through a good 16/44
DAC"....

impossible, for one. pure SACDs will not play on anything but a
hardware SACD deck. The good ones have a true 1bit DAC and that is
all that is used.
Some SACDs have a redbook layer, and that can be played on normal
decks. But your not hearing the DSD layer at that point, so its a
moot point.

I've got a good DAC. I've had plenty of em. I'm no stranger to PCM
either. my ears tell me that my DSD recording efforts have certainly
taken things to a new level of "realism". I'm trusting my ears at
this point.
Still...., that DSD jukebox eludes me, so for 99% of my listening PCM
is where its at. All my masters are "dumbed up" to redbook as well as
24bit.
Before this, i was mastering at 24/96 or 24/44.1. Mostly the later.
Was very satisfied. But like I said, the Korg is taking things to a
new level.
I'm recording the Charlie Hunter Trio tonight. Jazz trio. Charlie
plays a 10 string guitar (4 bass strings, 6 guitar strings, two
outputs (one for each), and is very talented. This will be an on
stage ambient recording. Minimalist approach. I'm probably running M-
S. Maybe blumlein. I'll post a link to some files (.dff master and
various redbook) if anyone would care to listen and see for
themselves.

[email protected]

April 15th 07, 01:47 PM

On Apr 15, 7:06 am, "Arny Krueger" > wrote:

> Let us know when you do that magical blind listening test showing that
> down-coding one of your 24/96 recordings to 44/16 has some audible effect
> in a well-designed listening test.

who said I need to down sample...

but then if one must, consider these...

< http://www.24-96.net/dither/ >

< http://src.infinitewave.ca/ >

Carey Carlan

April 15th 07, 02:49 PM

"Nick's Picks" > wrote in news:1176640920.374926.166250
@n76g2000hsh.googlegroups.com:

> the comment of "take your SACDs and run it through a good 16/44
> DAC"....
>
> impossible, for one. pure SACDs will not play on anything but a
> hardware SACD deck. The good ones have a true 1bit DAC and that is
> all that is used.

That's probably the greatest advantage of DSD. The DAC is almost trivially
simple. In the very simplest version you can simply apply a low pass
filter to the 3+ mb stream and get a usable output.

hank alrich

April 15th 07, 03:45 PM

Arny Krueger wrote:

> Regrettably, the local Circuit City store closed.

Not at all regrettable, considering these are the kind of pricks that
fire their experienced employees because those folks have worked
steadily for too long and made it too far up the wage scale, to replace
those loyal workers with minimum wage earners so the CEO can keep in the
gravy. **** 'em if they have no loyalty to their own labor force.

--
ha
Iraq is Arabic for Vietnam

Harry Lavo

April 15th 07, 03:53 PM

"Arny Krueger" > wrote in message
. ..
> "Harry Lavo" > wrote in message
>
>> "Arny Krueger" > wrote in message
>> ...
>>> "Harry Lavo" > wrote in message
>>>
>>>> "William Sommerwerck" > wrote
>>>> in message
>>>> . ..
>>>>>>> I don't know the source you are quoting, but it is
>>>>>>> important that it include web sources, since
>>>>>>> virtually all SACD sales and most DVD-A sales have
>>>>>>> moved to the web.
>>>>>
>>>>> I recently saw a statement that the total number of
>>>>> SACD sales last year was
>>>>> about one in a thousand of all optical-media audio
>>>>> sales. That's not so hot.
>>>>>
>>>>> Of course, it doesn't matter how many are sold, as long
>>>>> as the record companies can at least break even on the
>>>>> ones they sell.
>>>
>>>> Two-tenths of one percent...that's more like
>>>> one-in-five-hundred. But it was steady from 2004 to
>>>> 2005...I'm anxious to see the 2006 numbers.
>>>> Incidentally, if you read my other post in response to
>>>> Chel, you'll see it is five times higher in Holland, or
>>>> one-in-a-hundred.
>>>
>>> Time to wake up and smell the coffee, Harry. The numbers
>>> for the first half of 2006 have been out for over 6
>>> months. Announcement:
>>>
>>> http://www.riaa.com/news/newsletter/101206.asp
>>>
>>> Actual numbers:
>>>
>>> http://www.riaa.com/news/newsletter/pdf/2006midYrStats.pdf
>>>
>>> SACD > 44% down from previous period.
>>> DVD-A > 35% down from previous period
>>>
>>> DVD-A+SACD are now less than 0.1% of CD sales.
>>
>> Your numbers are correct directionally, but not factually
>> for dollar sales, which is the measure I used as most
>> universal.
>> Actual combined sales volume fell -38.4% vs a market
>> decline of -16.0% Hi-Rez Share dropped from 0.16% to
>> 0.12%.
>
>> However, there are a few caveats to beware of before
>> jumping to conclusions:
>
>> * 2006 saw Best Buy and Circuit City drop SACD and DVD-A,
>> which could account for much of the decline.
>
> As usual Harry, you are more than a little fact-challenged.
>
> I've seen SACDs amd DVD-As on sale in Best Buy stores in 2007 - in the
> past few weeks.
>
> Regrettably, the local Circuit City store closed.
>
> From the best Buy Web site today:
>
> http://www.bestbuy.com/site/olspage.jsp?id=cat12077&categoryRep=cat02000&guideID=1054508946422&type=page
>
> "Best Buy also carries a range of DVD-Audio titles, with many more soon to
> hit our shelves. You'll find a variety of genres, including rock, R&B,
> classical and jazz. Many of your favorite older albums have been remixed
> and remastered to bring you closer than ever to the original recording.
> There are also many new releases that have been recorded with DVD-Audio in
> mind during the creative process - providing for even more sonic
> definition!"
>
> "Take a visit to your local Best Buy for the players, surround sound
> systems and other gear that'll deliver a peak SACD listening experience.
> Best Buy also carries a range of SACD titles, with many new discs on the
> way. You'll find a variety of genres, including rock, R&B, classical and
> jazz. Many of your favorite older albums have been remixed and remastered
> to bring you closer than ever to the original recording (including the
> recently released hybrid SACD/CD version of Pink Floyd's Dark Side of the
> Moon). There are also many new releases that have been recorded with SACD
> in mind during the creative process - promising an astounding surround
> sound experience.
>
> Searching on "sacd" on the web site finds the following
>
> http://www.bestbuy.com/site/olspage.jsp?_dyncharset=ISO-8859-1&id=pcat17071&type=page&st=sacd&sc=Global&cp=1&nrp=15&sp=&qp=&list=n&iht=y&usc=All+Categories&ks=960
>
> 703 items are found including numerous recordings.

Do you know the difference between brick and mortar and a web site, Arny?

Do you know the difference between having SACD and DVD-A sections, and
placing a few single inventory items in the regular racks/

Because that is the situation at Best Buy, and has been since early 2006.

Harry Lavo

April 15th 07, 04:03 PM

"Arny Krueger" > wrote in message
. ..
> "Harry Lavo" > wrote in message
>
>
>>>> You need to time-synch and double-blind test.
>
> Harry you're preaching to the choir and you forgot about leve-matching.

What do you think two identical players and hybrid SACD's provide, Arny?
You've conveniently eliminated my early listing of requirments.

>
>>> > Ideally
>>>> it should be a preference test, rather than an identity
>>>> test as work at HK suggests this is easier for people to
>>>> do.
>
> As usual Harry, you're taking gross liberties with what Harmon's people
> have written about the matter.

Bull****. Simply not true...I'm not going to get into nitpicking debating
trade **** with you here, Arny......it's another in those set of
inconvenient "facts" that you simply choose to ignore.

>
>>>> Ideally you do it across a cross-section of careful and
>>>> experienced listeners well attuned to live classical
>>>> music for a set number of tests each, allowing each to
>>>> control the switching.
>
> Been there, done that.

I'm talking about a single test that amasses a hundred or more responses
across a dozen people, Arny. Take off your ABX blinders.

>
>>>> You do it in a casual and comfortable home-like setting,
>>>> with amenities and no pressure and you allow them to
>>>> stop
>
> Been there, done that.

See comment above.

>
>>>> when fatiqued and resume when they are rested and have
>>>> taken a break.
>
> Doooh!

This is just meaningless drivel, Arny. Appropriate to RAO perhaps, but not
here.

>
>>>> You sum the responses and apply statistical analysis.
>>>> When you get over 100-150 separate samples from at least
>>>> 10 people, you look at the statistics. For scientific
>>>> "proof" you want 95% reliability (19:1 odds). Great if
>>>> you get 99% (99:1). Still interesting if you get 80%
>>>> (4:1).
>
> Been there, done that.

This is just meaningless drivel, Arny. Appropriate to RAO perhaps, but not
here.

>
>>>> If you want a control (and you should) you provide a
>>>> blind monadic test individually to 100 similar listeners
>>>> for each of the two formats, and ask them to rate the
>>>> music and musical playback quality on a serious of
>>>> semantic differential scales.
>
> Monadic testing is a hypothesis of Harry Lavo, not an accepted blind
> testing protocol.

Monadic testing with semantic differential is one of the classic evaluative
tests in almost any field involving human perception, Arny.

>
>>>> When finished you then
>>>> apply statistics to the scalers between the two groups
>>>> of 100, and note differences that are statistically
>>>> signifcant at high rates. The overall results should
>>>> support the individual pairs conclusions if both tests
>>>> are valid, and the semantic differential scaling should
>>>> tell you how and why people perceived one as better than
>>>> the other.
>
> After you, Harry.

More RAO bull****, Arny.

>
>>>> A simple little test....don't know why everybody doesn't
>>>> do it.......
>
>> No comment on the real complexity of doing a truly
>> scientific test, I guess?
>
> Been there, done that unlike you Harry.

I see, Arny. And you've published the results of your 10 person, 150
judgement test, with control test where Arny? This is not RAO, Arny.

Harry Lavo

April 15th 07, 04:04 PM

"Arny Krueger" > wrote in message
. ..
> "Harry Lavo" > wrote in message
>
>> "Arny Krueger" > wrote in message
>> . ..
>>> "Harry Lavo" > wrote in message
>>>
>>>> "Arny Krueger" > wrote in message
>>>> ...
>>>>> "William Sommerwerck" >
>>>>> wrote in message
>>>>>
>>>>>>> Been there done that - the difference is that I dare
>>>>>>> to listen without having my eyes tell me what the
>>>>>>> *right* answer should be.
>>>>>>
>>>>>> No, you have your "science" telling your ears what
>>>>>> they ought to hear.
>>>>>
>>>>> Doesn't matter. Besides, enough listening is done by
>>>>> people who are ignorant of the science so that if there
>>>>> was some hidden truth, it would out.
>>>>
>>>> Arny has said elsewhere that it doesn't matter that ABX
>>>> has no controls against "false negatives" because he
>>>> "knows" he always gives it his best shot to
>>>> differentiate.
>>>
>>> That would be yet another one of your petty lies, Harry.
>
>>> What I did say is that otherwise skeptical people tend
>>> to get sucked into doing their best for listening tests
>>> that they might otherwise pooh-pooh.
>
>> You said that in addition to saying you always gave your
>> best.
>
> As usual, the actual quote should you ever provide it in context, is
> vastly different from what you initially presented, Harry.

I'm not playing RAO "debating trade" games with you here, Arny. My capture
of what you said stands...I didn't put quotes around the specific words.

Mike Rivers

April 15th 07, 05:24 PM

On Apr 15, 8:42 am, "Nick's Picks" > wrote:

> the comment of "take your SACDs and run it through a good 16/44
> DAC"....
>
> impossible, for one. pure SACDs will not play on anything but a
> hardware SACD deck.

I think Arny's experiment was to connect the analog output (what you
listen to) of the SACD player to a "wire" made up of a good pair of
16/44 A/D and D/A converters back to back. See what's important that's
lost by this not quite straight wire.

> my ears tell me that my DSD recording efforts have certainly
> taken things to a new level of "realism". I'm trusting my ears at
> this point.

That may be true, but it may not be because the DSD process made it
any better. It could be that better care was taken in the recording
process simply because, knowing that there would be an SACD release,
it had darned will better sound better than a CD or the marketing
department would be all over the recording engineer's ass. Same thing
if it was an SACD remake of a project recorded by whatever process and
originally released on CD.

The real test is with your own recordings, which apparently you have
made. If it isn't too much trouble, send the outputs of your mic
preamps to both your DSD recorder and the best PCM A/D converter you
have, and compare the results. But don't bother to put up MP3 files as
"proof." <g>

Scott Dorsey

April 15th 07, 07:16 PM

Mike Rivers > wrote:
>On Apr 15, 8:42 am, "Nick's Picks" > wrote:
>
>> the comment of "take your SACDs and run it through a good 16/44
>> DAC"....
>>
>> impossible, for one. pure SACDs will not play on anything but a
>> hardware SACD deck.
>
>I think Arny's experiment was to connect the analog output (what you
>listen to) of the SACD player to a "wire" made up of a good pair of
>16/44 A/D and D/A converters back to back. See what's important that's
>lost by this not quite straight wire.

It is ALSO possible to take the SACD bitstream and convert it into a
PCM bitstream without leaving the digital domain. Merging Technologies
and EMM have devices to do this. Grimm Audio promises that they will
have one soon.

Does this conversion degrade the sound audibly? I don't know, but that's
the thing we need to find out.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Arny Krueger

April 16th 07, 01:31 AM

> wrote in message
oups.com
> On Apr 15, 7:06 am, "Arny Krueger" >
> wrote:
>
>> Let us know when you do that magical blind listening
>> test showing that down-coding one of your 24/96
>> recordings to 44/16 has some audible effect in a
>> well-designed listening test.
>
> who said I need to down sample...

What's your distribution format?

> but then if one must, consider these...
>
> < http://www.24-96.net/dither/ >

Utter BS. Next time I make a recording with peak levels more than 80 dB
down, I'll put on my dunce cap in sit in the corner until I wise up resume
doing quality work.

> < http://src.infinitewave.ca/ >

Just numbers for the sake of numbers.

Arny Krueger

April 16th 07, 01:33 AM

"Nick's Picks" > wrote in message
oups.com
> wow..
> now this is a thread hijack!
>
>
> good reading though. I didn't mean it to be a format
> debate, but good reading none the less.
> DSD, as I understand it, is not entirely Sony's baby.
> Also a japonese professor (who liscened it to Korg), Ed
> Meitner also had something to do w/it (?), or just makes
> it oh so nice.

> the comment of "take your SACDs and run it through a good
> 16/44 DAC"....

> impossible, for one.

You seem to terribly misinformed.

> pure SACDs will not play on anything but a hardware SACD deck.

So, SACD players have analog outputs, right?

Mine does. What's wrong with yours?

Arny Krueger

April 16th 07, 01:35 AM

"Scott Dorsey" > wrote in message

> Mike Rivers > wrote:
>> On Apr 15, 8:42 am, "Nick's Picks" >
>> wrote:
>>
>>> the comment of "take your SACDs and run it through a
>>> good 16/44 DAC"....
>>>
>>> impossible, for one. pure SACDs will not play on
>>> anything but a hardware SACD deck.
>>
>> I think Arny's experiment was to connect the analog
>> output (what you listen to) of the SACD player to a
>> "wire" made up of a good pair of 16/44 A/D and D/A
>> converters back to back. See what's important that's
>> lost by this not quite straight wire.
>
> It is ALSO possible to take the SACD bitstream and
> convert it into a
> PCM bitstream without leaving the digital domain.
> Merging Technologies and EMM have devices to do this.
> Grimm Audio promises that they will have one soon.
>
> Does this conversion degrade the sound audibly? I don't
> know, but that's the thing we need to find out.

The answer to that question is already known - the answer is no, providing
the listening test is bias-controlled and level-matched.

Note how all the SACD advocates scatter when it is suggested that they
actually do a fair comparison.

Arny Krueger

April 16th 07, 01:39 AM

"Harry Lavo" > wrote in message

> "Arny Krueger" > wrote in message
> . ..
>> "Harry Lavo" > wrote in message
>>
>>> "Arny Krueger" > wrote in message
>>> ...
>>>> "Harry Lavo" > wrote in message
>>>>
>>>>> "William Sommerwerck" >
>>>>> wrote in message
>>>>> . ..
>>>>>>>> I don't know the source you are quoting, but it is
>>>>>>>> important that it include web sources, since
>>>>>>>> virtually all SACD sales and most DVD-A sales have
>>>>>>>> moved to the web.
>>>>>>
>>>>>> I recently saw a statement that the total number of
>>>>>> SACD sales last year was
>>>>>> about one in a thousand of all optical-media audio
>>>>>> sales. That's not so hot.
>>>>>>
>>>>>> Of course, it doesn't matter how many are sold, as
>>>>>> long as the record companies can at least break even
>>>>>> on the ones they sell.
>>>>
>>>>> Two-tenths of one percent...that's more like
>>>>> one-in-five-hundred. But it was steady from 2004 to
>>>>> 2005...I'm anxious to see the 2006 numbers.
>>>>> Incidentally, if you read my other post in response to
>>>>> Chel, you'll see it is five times higher in Holland,
>>>>> or one-in-a-hundred.
>>>>
>>>> Time to wake up and smell the coffee, Harry. The
>>>> numbers for the first half of 2006 have been out for
>>>> over 6 months. Announcement:
>>>>
>>>> http://www.riaa.com/news/newsletter/101206.asp
>>>>
>>>> Actual numbers:
>>>>
>>>> http://www.riaa.com/news/newsletter/pdf/2006midYrStats.pdf
>>>>
>>>> SACD > 44% down from previous period.
>>>> DVD-A > 35% down from previous period
>>>>
>>>> DVD-A+SACD are now less than 0.1% of CD sales.
>>>
>>> Your numbers are correct directionally, but not
>>> factually for dollar sales, which is the measure I used
>>> as most universal.
>>> Actual combined sales volume fell -38.4% vs a market
>>> decline of -16.0% Hi-Rez Share dropped from 0.16% to
>>> 0.12%.
>>
>>> However, there are a few caveats to beware of before
>>> jumping to conclusions:
>>
>>> * 2006 saw Best Buy and Circuit City drop SACD and
>>> DVD-A, which could account for much of the decline.

<Note that Harry made a general statement that included the entire Circuit
City and Best By enterprises, both local stores and web sites.>

>> As usual Harry, you are more than a little
>> fact-challenged.

>> I've seen SACDs and DVD-As on sale in Best Buy stores in
>> 2007 - in the past few weeks.

<snip discussion of Best Buy's Web site. because it seemed to confuse Harry>

> Do you know the difference between brick and mortar and a
> web site, Arny?

Doesn't matter Harry, since your claims were enterprise-wide.

But I did specifically mention a Best Buy retail store.

Can't read, can you Harry?

Or is that you only read what you want to see?

Arny Krueger

April 16th 07, 01:46 AM

"Harry Lavo" > wrote in message

> "Arny Krueger" > wrote in message
> . ..
>> "Harry Lavo" > wrote in message
>>
>>
>>>>> You need to time-synch and double-blind test.

>> Harry you're preaching to the choir and you forgot about
>> leve-matching.

> What do you think two identical players and hybrid SACD's
> provide, Arny? You've conveniently eliminated my early
> listing of requirments.

No Harry, it is you who first eliminated my earlier listing of requirements:

>>> The obvious comparison is to transcribe a SACD to a CD
>>> with a good audio interface, and them compare the two -
>>> level-matched, time-synched, and blind. The hard part
>>> is the time-synched part - you have to keep the two
>>> discs synched within a fraction of second, or a quick
>>> switch will easily identify which is which even if they
>>> are actually identical.

>>>>> Ideally
>>>>> it should be a preference test, rather than an
>>>>> identity test as work at HK suggests this is easier
>>>>> for people to do.
>>
>> As usual Harry, you're taking gross liberties with what
>> Harmon's people have written about the matter.

> Bull****. Simply not true...I'm not going to get into
> nitpicking debating trade **** with you here,

You started it Harry when you deleted my comments

>>> The obvious comparison is to transcribe a SACD to a CD
>>> with a good audio interface, and them compare the two -
>>> level-matched, time-synched, and blind. The hard part
>>> is the time-synched part - you have to keep the two
>>> discs synched within a fraction of second, or a quick
>>> switch will easily identify which is which even if they
>>> are actually identical.

> Arny......it's another in those set of inconvenient
> "facts" that you simply choose to ignore.

No Harry, it is you who first started ignoring mine.

>>>>> Ideally you do it across a cross-section of careful
>>>>> and experienced listeners well attuned to live
>>>>> classical music for a set number of tests each,
>>>>> allowing each to control the switching.
>>
>> Been there, done that.

> I'm talking about a single test that amasses a hundred or
> more responses across a dozen people, Arny.

Been there, done that.

> Take off your ABX blinders.

You're just BSing Harry.

>>>>> You do it in a casual and comfortable home-like
>>>>> setting, with amenities and no pressure and you allow
>>>>> them to stop
>>
>> Been there, done that.
>
> See comment above.

It was BS.

>>>>> when fatiqued and resume when they are rested and have
>>>>> taken a break.
>>
>> Doooh!

> This is just meaningless drivel, Arny. Appropriate to
> RAO perhaps, but not here.

You're talking trash, Harry. You act like you invented the idea that
listeners shouldn't be fatiqued. We beat you to that more than 25 years
ago.

>>>>> You sum the responses and apply statistical analysis.
>>>>> When you get over 100-150 separate samples from at
>>>>> least 10 people, you look at the statistics. For
>>>>> scientific "proof" you want 95% reliability (19:1
>>>>> odds). Great if you get 99% (99:1). Still
>>>>> interesting if you get 80% (4:1).

>> Been there, done that.

> This is just meaningless drivel, Arny. Appropriate to
> RAO perhaps, but not here.

You're talking trash, Harry. You act like you invented the idea that
statistical tests should be used. We beat you to that more than 25 years
ago.

>>>>> If you want a control (and you should) you provide a
>>>>> blind monadic test individually to 100 similar
>>>>> listeners for each of the two formats, and ask them
>>>>> to rate the music and musical playback quality on a
>>>>> serious of semantic differential scales.

>> Monadic testing is a hypothesis of Harry Lavo, not an
>> accepted blind testing protocol.

> Monadic testing with semantic differential is one of the
> classic evaluative tests in almost any field involving
> human perception, Arny.

You're talking trash, Harry. You act like you invented classic listening
tests. We beat you to that more than 25 years ago.

>>>>> When finished you then
>>>>> apply statistics to the scalers between the two groups
>>>>> of 100, and note differences that are statistically
>>>>> signifcant at high rates. The overall results should
>>>>> support the individual pairs conclusions if both tests
>>>>> are valid, and the semantic differential scaling
>>>>> should tell you how and why people perceived one as
>>>>> better than the other.

>> After you, Harry.

> More RAO bull****, Arny.

Yes, that's what you're talking Harry.

>>>>> A simple little test....don't know why everybody
>>>>> doesn't do it.......
>>
>>> No comment on the real complexity of doing a truly
>>> scientific test, I guess?
>
>> Been there, done that unlike you Harry.

> I see, Arny. And you've published the results of your 10
> person, 150 judgement test, with control test where Arny?

I didn't publish it - Stereo Review and S&V have. Nousaine summarized such
tests in an AES paper.

> This is not RAO, Arny.

Arny Krueger

April 16th 07, 01:48 AM

"William Sommerwerck" > wrote in
message

>>> I'll leave it to a physics major to explain why F=ma has
>>> nothing to do with driver breakup.

>> Sorry, William, but every analysis of driver breakup
>> I've ever seen or done includes F=ma.

> This is like saying that because F=ma is part of the
> explanation of why eggs crack when hit on the edge of a
> bowl, that F=ma explains cake-baking.

If a physicist writes the equations of motion relating to the cracking of
eggs on a bowl, he does indeed start with F=MA.

[email protected]

April 16th 07, 02:36 AM

On Apr 15, 8:31 pm, "Arny Krueger" > wrote:

> Just numbers for the sake of numbers.

ah, been there done that....computer! the numbers win!

[email protected]

April 16th 07, 02:40 AM

On Apr 15, 8:31 pm, "Arny Krueger" > wrote:

> What's your distribution format?

dvd-r, niche market, no industry tracking #'s

Steven Sullivan

April 16th 07, 03:01 AM

Mike Rivers > wrote:
> On Apr 13, 7:09 am, "Arny Krueger" > wrote:

> > One practical problem with Sony's DSD strategy is that the so-called
> > hardware problems of 196 KHz PCM converters were solved so thorougly that
> > they can even be found in $39.95 DVD players.

> You may not agree with Dan Lavry, but he doesn't think so. On the
> other hand, he doesn't make $39.95 DVD players, so what people do with
> those doesn't matter a lot fo him.

Last I read Lavry -- which has been some months now --
he was not a proponent of 192, or even 96 kHz.
By his reckoning even the theoretical problems with rebdook standard are
solved by the time the sampling rate reaches ~ 60 kHz.

Of course, he offers the 'industry standard' sampling rates on his gear to
stay competitive...that's just good business. Just like some
loudspeaker manufacturers offered a biwiring option and spiked
stands despite not being 'believers'.

___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

Steven Sullivan

April 16th 07, 03:05 AM

Scott Dorsey > wrote:

> That's true. BUT, the good news about DSD is that it has power
> density issues that make it hard to do abusive processing and get
> away with it. That's a huge plus in my book.

Can abusive processing be performed in the PCM domain, then
the file be transcoded to DSD?

___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

Steven Sullivan

April 16th 07, 03:13 AM

William Sommerwerck > wrote:
> >> That if product
> >> A has measured specs comparable to product B, that they
> >> necessarily sound alike (or very nearly so), EVEN THOUGH
> >> THOSE SPECS HAVE NEVER BEEN CORRELATED WITH
> >> LISTENING QUALITY?

> > Straw man argument.

> Not at all. It's the fundamental issue here. Measurements and specs are
> largely useless correlated with subjective impression.

But there's another fundamental issue -- the potential errancy of
subjective impression.

'Audiophiles' tend to downplay that one, though it's a given in
any careful investigation of audio phenomena.

When 'objectivists' express skepticism over, say, anecdotal reports
(even by recording engineers) of the inherent
'audible superiority' of DSD over Redbook, it is not just because
'specs' suggest such difference is unlikely. It is because
the ever-present chance of self-deception has not been
ruled out.

___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

Steven Sullivan

April 16th 07, 03:17 AM

William Sommerwerck > wrote:
> > Has anybody A/B'd a DSD signal with its PCM copy [sic]
> > in a vetted test yet? :)

> Actually, a better test would be to compare the DSD layer of an SACD with
> its CD layer.
> This itself would include two comparisons -- when the CD layer
> is derived from the DSD stream (eg, a RCA Living Stereo SACD), and when the
> CD layer is a direct PCM transfer (eg, a Mercury Living Presence SACD).

THat would be a better test, IF one can assume that such 'direct' transfer
occurred, with no other differential tweaking. THis has not been a safe
assumption so far (e.g., Dark Side of the Moon).

But even if it could eb assume,d in practice deteriming an *inherent*
difference between a SACD vs CD is a very difficult test, because
differences between playback chains still cannot be ruled out.

___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

Harry Lavo

April 16th 07, 03:43 AM

"Arny Krueger" > wrote in message
. ..
> "Harry Lavo" > wrote in message
>
>> "Arny Krueger" > wrote in message
>> . ..
>>> "Harry Lavo" > wrote in message
>>>
>>>> "Arny Krueger" > wrote in message
>>>> ...
>>>>> "Harry Lavo" > wrote in message
>>>>>
>>>>>> "William Sommerwerck" >
>>>>>> wrote in message
>>>>>> . ..
>>>>>>>>> I don't know the source you are quoting, but it is
>>>>>>>>> important that it include web sources, since
>>>>>>>>> virtually all SACD sales and most DVD-A sales have
>>>>>>>>> moved to the web.
>>>>>>>
>>>>>>> I recently saw a statement that the total number of
>>>>>>> SACD sales last year was
>>>>>>> about one in a thousand of all optical-media audio
>>>>>>> sales. That's not so hot.
>>>>>>>
>>>>>>> Of course, it doesn't matter how many are sold, as
>>>>>>> long as the record companies can at least break even
>>>>>>> on the ones they sell.
>>>>>
>>>>>> Two-tenths of one percent...that's more like
>>>>>> one-in-five-hundred. But it was steady from 2004 to
>>>>>> 2005...I'm anxious to see the 2006 numbers.
>>>>>> Incidentally, if you read my other post in response to
>>>>>> Chel, you'll see it is five times higher in Holland,
>>>>>> or one-in-a-hundred.
>>>>>
>>>>> Time to wake up and smell the coffee, Harry. The
>>>>> numbers for the first half of 2006 have been out for
>>>>> over 6 months. Announcement:
>>>>>
>>>>> http://www.riaa.com/news/newsletter/101206.asp
>>>>>
>>>>> Actual numbers:
>>>>>
>>>>> http://www.riaa.com/news/newsletter/pdf/2006midYrStats.pdf
>>>>>
>>>>> SACD > 44% down from previous period.
>>>>> DVD-A > 35% down from previous period
>>>>>
>>>>> DVD-A+SACD are now less than 0.1% of CD sales.
>>>>
>>>> Your numbers are correct directionally, but not
>>>> factually for dollar sales, which is the measure I used
>>>> as most universal.
>>>> Actual combined sales volume fell -38.4% vs a market
>>>> decline of -16.0% Hi-Rez Share dropped from 0.16% to
>>>> 0.12%.
>>>
>>>> However, there are a few caveats to beware of before
>>>> jumping to conclusions:
>>>
>>>> * 2006 saw Best Buy and Circuit City drop SACD and
>>>> DVD-A, which could account for much of the decline.
>
> <Note that Harry made a general statement that included the entire Circuit
> City and Best By enterprises, both local stores and web sites.>
>
>>> As usual Harry, you are more than a little
>>> fact-challenged.
>
>>> I've seen SACDs and DVD-As on sale in Best Buy stores in
>>> 2007 - in the past few weeks.
>
> <snip discussion of Best Buy's Web site. because it seemed to confuse
> Harry>
>
>> Do you know the difference between brick and mortar and a
>> web site, Arny?
>
> Doesn't matter Harry, since your claims were enterprise-wide.
>
> But I did specifically mention a Best Buy retail store.
>
> Can't read, can you Harry?
>
> Or is that you only read what you want to see?

Fact of the matter is, Arny, that I spend a lot of time on the hi-rez forum
in Audio Asylum....you don't...so we *know* when Best Buy moved to take down
their SACD and DVD-A pretty much across the country. If you found a store
selling SACD or DVD-A then it is out of the mainstream Best Buy chain. And
since Best Buy was hardly even a presence on the web a year ago, when I (or
somebody else) refers to Best Buy they are usually talking about the largest
chain of retail bricks and mortar stores in the country, not a come-lately
web site.

No amount of "debating trade" is going to obscure the reality, Arny....Best
Buy as a chain is pretty much out of the DVD-A and SACD business.

Harry Lavo

April 16th 07, 03:46 AM

"Steven Sullivan" > wrote in message
...
> Mike Rivers > wrote:
>> On Apr 13, 7:09 am, "Arny Krueger" > wrote:
>
>> > One practical problem with Sony's DSD strategy is that the so-called
>> > hardware problems of 196 KHz PCM converters were solved so thorougly
>> > that
>> > they can even be found in $39.95 DVD players.
>
>> You may not agree with Dan Lavry, but he doesn't think so. On the
>> other hand, he doesn't make $39.95 DVD players, so what people do with
>> those doesn't matter a lot fo him.
>
> Last I read Lavry -- which has been some months now --
> he was not a proponent of 192, or even 96 kHz.
> By his reckoning even the theoretical problems with rebdook standard are
> solved by the time the sampling rate reaches ~ 60 kHz.
>
> Of course, he offers the 'industry standard' sampling rates on his gear to
> stay competitive...that's just good business. Just like some
> loudspeaker manufacturers offered a biwiring option and spiked
> stands despite not being 'believers'.

First, his units do not go above 96khz...for a reason that he argues
cogently.

Second, I specifically referred to the discussion during the month of March
on the Audio Pro listserve.

Harry Lavo

April 16th 07, 03:54 AM

"Steven Sullivan" > wrote in message
...
> William Sommerwerck > wrote:
>> > Has anybody A/B'd a DSD signal with its PCM copy [sic]
>> > in a vetted test yet? :)
>
>> Actually, a better test would be to compare the DSD layer of an SACD with
>> its CD layer.
>> This itself would include two comparisons -- when the CD layer
>> is derived from the DSD stream (eg, a RCA Living Stereo SACD), and when
>> the
>> CD layer is a direct PCM transfer (eg, a Mercury Living Presence SACD).
>
> THat would be a better test, IF one can assume that such 'direct' transfer
> occurred, with no other differential tweaking. THis has not been a safe
> assumption so far (e.g., Dark Side of the Moon).
>
> But even if it could eb assume,d in practice deteriming an *inherent*
> difference between a SACD vs CD is a very difficult test, because
> differences between playback chains still cannot be ruled out.
>

I have pointed out to you specifically on other forums that Chesky's "Swing
Low" is was done this way in is available on SACD and CD hybrid, as well as
DVD-A.

I also pointed out that Stereophile has a Mozart recording done exactly for
this purpose and can be compared SACD to CD, and either to LP.

I also pointed out years ago that many of Columbia's early non-hybrid
releases were SACD versions of classical works from their earlier periods
that had no differences in dynamic range or other sonic anomalies and could
be used for valid comparisons of formats....I specifically mentions the
Ormandy/Philadelphia/Verdi Requiem, and the Walter/Columbia/Beethoven 5th in
this regard.

All you need are two identical universal players to compare CD/SACD/and
DVD-A, especially ones that are acknowledge to be excellent in each format.
Either that, or some subset...such as my Sony XA2000ES's which can do SACD
and CD.

William Sommerwerck

April 16th 07, 11:57 AM

> But there's another fundamental issue -- the potential errancy
> of subjective impression.

That's precisely the reason I've argued for long-term listening. And not
just for one auditioner, but many.

It's ludicrous to claim that "forced-decision" listening tests, under
conditions that have nothing to do with the way we actually listen to music,
are the last word. Useful, probably. But not definitive.

William Sommerwerck

April 16th 07, 12:00 PM

> But even if it could be assumed in practice, deteriming an *inherent*
> difference between a SACD and CD is a very difficult test, because
> differences between playback chains still cannot be ruled out.

What would be the difference in playback chains?

When playing a hybrid disk, you're using the same D/A converters for both CD
and SACD. At least, in the case of the SACD player I own, you are.

Mike Rivers

April 16th 07, 12:57 PM

On Apr 16, 6:57 am, "William Sommerwerck" >
wrote:

> That's precisely the reason I've argued for long-term listening. And not
> just for one auditioner, but many.

One good test for problems (without analyzing just what the problems
are) is how long you can listen without getting fatigued. Since
someone who can listen to opera all day can get fatigued after a few
minutes of hip-hop (all other things being equal) means that there can
be no standardized listening test. It's possible to collect anecdotal
data and this is where you get the reports that SACD or DSD sounds
SOOOOO much better than (whatever).

Arny Krueger

April 16th 07, 01:23 PM

"Harry Lavo" > wrote in message

> Fact of the matter is, Arny, that I spend a lot of time
> on the hi-rez forum in Audio Asylum....you don't..

Yeah Harry, I'm too busy doing real- world audio and computer work to spend
as much time as you do yapping on various audio forums.

For example I did almost 100 live recordings for hiren the past month or
so - how many did you do, Harry?

I created 4 video DVDs ranging from 5 minutes to 45 minutes, some original
work, some compendiums of other works that were major components of 4 live
performances, each attended by up to 360 people.

I mixed 10 rehearsals and 5 live performances with a total attendance of
over 1,500 people.

I produced 5 CDs that were sold to the general public, tiny niche market
products to be sure, but neverthess stuff that sold.

That doesn't make me any kind of exceptional audio guy - there's people
around here who did more in just the past few days, but what did you do,
Harry? I'm quite sure you can't even bring yourself to admit what you did
that would comparable, since it was likely zero.

> .so we
> *know* when Best Buy moved to take down their SACD and
> DVD-A pretty much across the country.

I provided considerable real-world evidence that says otherwise. Just
another case of Harry trying to turn rumor into facts.

> If you found a
> store selling SACD or DVD-A then it is out of the
> mainstream Best Buy chain.

Yeah, like their web site is completely out of the mainstream Best Buy
chain. Tell us another story, Harry.

> And since Best Buy was hardly
> even a presence on the web a year ago, when I (or
> somebody else) refers to Best Buy they are usually
> talking about the largest chain of retail bricks and
> mortar stores in the country, not a come-lately web site.

Yap, yap, yap, yap.

> No amount of "debating trade" is going to obscure the
> reality, Arny....Best Buy as a chain is pretty much out
> of the DVD-A and SACD business.

In fact Harry, every alleged fact that you presented was adequately falsifed
by real-world evidence. And that was just a footnote on the story of the
collapse of hi-rez media in the mainstream music market, as documented by no
less than the RIAA.

Denial isn't a river in Egypt - its a major component of Harry Lavo's life.
Too bad about those people who spent the big bucks on DVD-A and SACD
players. I hope they enjoy them, because the sun is setting on them. Slowly
for sure, but the sun is setting and HD DVD and Blu-Ray have their sights
set on their market segment.

Arny Krueger

April 16th 07, 01:23 PM

"Steven Sullivan" > wrote in message

> Scott Dorsey > wrote:
>
>> That's true. BUT, the good news about DSD is that it
>> has power density issues that make it hard to do abusive
>> processing and get away with it. That's a huge plus in
>> my book.
>
> Can abusive processing be performed in the PCM domain,
> then
> the file be transcoded to DSD?

Absolutely. It's already been done for commercial releases. The form of the
abusive processing would be traditional audio processing as conducted in the
pre-digital era.

Arny Krueger

April 16th 07, 01:26 PM

"Steven Sullivan" > wrote in message

> William Sommerwerck > wrote:
>>> Has anybody A/B'd a DSD signal with its PCM copy [sic]
>>> in a vetted test yet? :)
>
>> Actually, a better test would be to compare the DSD
>> layer of an SACD with its CD layer.
>> This itself would include two comparisons -- when the CD
>> layer
>> is derived from the DSD stream (eg, a RCA Living Stereo
>> SACD), and when the CD layer is a direct PCM transfer
>> (eg, a Mercury Living Presence SACD).
>
> THat would be a better test, IF one can assume that such
> 'direct' transfer occurred, with no other differential
> tweaking. THis has not been a safe assumption so far
> (e.g., Dark Side of the Moon).
>
> But even if it could eb assume,d in practice deteriming
> an *inherent* difference between a SACD vs CD is a very
> difficult test, because differences between playback
> chains still cannot be ruled out.

I'm willing to turn a blind eye to considerable differences in the playback
chain, because those differences can easily still so small that positive
results in a proper subjective evaluation will highly unlikely. And if the
results are otherwise, well then it is a learning experience. We all win
either way.

Thats why I specifically mentioned interposing a device as mediocre as the
old 48 Khz Flying Cow in a signal chain. It can be reasonably expected to
demonstrate the lack of practical advantages to the so-called hi-rex
formats.

Arny Krueger

April 16th 07, 01:57 PM

"William Sommerwerck" > wrote in
message
>> We first did a similar experiment back in the late 1970s
>> with a comparable device of the day, and reported the
>> results here:
>
>> http://www.provide.net/~djcarlst/abx_digi.htm
>
>> From time to time we've repeated this kind of
>> experiement, and invite others to do the same.
>
> Why should I perform a listening test that, in my
> opinion, has no correlation with the way people normally
> listen?

Don't bother William - you already know the *right* answer. You spent $2k on
a fancy SACD player with hopes that were technically impossible, and can't
bring yourself to admit its limitations.

> How many times do we need to explain to you, Arny, that
> you're _assuming_ the validity of the experiment a
> priori?

Shows what little you know about the practice of science, William. No sane
person would knowingly spend any effort at all on an experiment that had
abosolutely no validity, except for just the fun of it. And, if its only for
fun, then its valid is only fun, and nothing that relates to anybody with a
more serious intent.

Before one starts an experiment with any serious intent, they must believe
that its outcome has some relevance and validity. It isn't blindly assumed,
it has to have some logical reason for being.

So no, I didn't assume the validity of any of the ABX experiements a priori.
The belief in their validity was based on the best possible investigation of
what it took to make a valid audio experiment, by a group of people with
surprisingly good credentials - one person with a BS in Math, several with
BS+advanced course work in electrical engineering, one with a BS in
experimental psychology, a medical practitiioner, teacher and researcher;
and several with years of professional experience in audio production and
product development.

You gonna say that's all just an a priori assumption, William? Pshawww....
;-)

> That's not science.

Scott Dorsey

April 16th 07, 02:14 PM

Steven Sullivan > wrote:
>Scott Dorsey > wrote:
>
>> That's true. BUT, the good news about DSD is that it has power
>> density issues that make it hard to do abusive processing and get
>> away with it. That's a huge plus in my book.
>
>Can abusive processing be performed in the PCM domain, then
>the file be transcoded to DSD?

Yes, but it doesn't make the end result any louder, so there is no point.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

William Sommerwerck

April 16th 07, 02:56 PM

>> Why should I perform a listening test that, in my
>> opinion, has no correlation with the way people normally
>> listen?

> Don't bother William - you already know the *right* answer.
> You spent $2k on a fancy SACD player with hopes that were
> technically impossible, and can't bring yourself to admit its
> limitations.

Wait a second... In a previous post, you said that the limitations of SACD
(or more precisely, DSD) were at the threshold of audibility, if at all
audible. Which is it?

>> How many times do we need to explain to you, Arny,
>> that you're _assuming_ the validity of the experiment
>> a priori?

> Shows what little you know about the practice of science,
> William. No sane person would knowingly spend any effort
> at all on an experiment that had abosolutely no validity,
> except for just the fun of it. And, if its only for fun, then its
> valid is only fun, and nothing that relates to anybody with
> a more serious intent.

It's done every day, Arny. "Serious" researchers set up poorly designed
experiments and/or misinterpret them. A serious intent is no guarantee of
accuracy or utility.

> Before one starts an experiment with any serious intent,
> they must believe that its outcome has some relevance
> and validity. It isn't blindly assumed, it has to have some
> logical reason for being.

And that's the point -- ABX testing has no reason for being. Because you
haven't demonstrated (and I use that word deliberately) its connection with
what people here when they sit down and listen to music in a hi-fi system.
You cannot assume there is a valid connection or correlation -- because the
conditions of the test are not the same as common listening conditions, you
have to demonstrate, to prove the connection.

> So no, I didn't assume the validity of any of the ABX experiements a
priori.
> The belief in their validity was based on the best possible investigation
of
> what it took to make a valid audio experiment, by a group of people with
> surprisingly good credentials -- one person with a BS in Math, several
with
> BS+advanced course work in electrical engineering, one with a BS in
> experimental psychology, a medical practitiioner, teacher and researcher;
> and several with years of professional experience in audio production and
> product development.

> You gonna say that's all just an a priori assumption, William? Pshawww....
> ;-)

Yes, it is. Because we're using "validity" in two altogether different
ways -- and your use is the wrong one.

Here's how Arny's thinking worked 30 years ago -- and still works:
"Double-blind testing is a widely accepted protocol in scientific studies.
Therefore, any form of testing using it must be scientific and must
necessarily produce valid results."

Let's ignore the fact that "A implies B" does not necessarily mean that "B
implies A".

The results of an experiment using double-blind protocols might be
mathematically valid. And they might be valid in the sense that the test was
properly conducted according to those (and other commonly accepted)
protocols. But are they valid in any practical sense?

Not always, because it's not always possible to set up an experiment that
doesn't interfere with what you're trying to measure (such as when the
subject is aware of the experiment), or that properly isolates the single
factor we're trying to learn about.

The latter is a major problem in sociological and psychological studies --
into which ABX testing of audio components falls. In order to show that such
testing is valid, you need to show that it correctly predicts "subjective"
long-term listening that reflects the way we commonly listen to music. As
you haven't done this, you can't say that ABX testing is valid. It might
very well be valid -- but you have no experimental or practical proof of it.

To put it more bluntly -- you're saying that because an experiment meets
certain mathematical and procedural requirements, that it _necessarily_
produces useful, valid, correct results. Since when?

All that ABX testing shows is that some things are audibly different under
the conditions of the test, others aren't. And the connection of this test
with what I hear when I drop an SACD in my overpriced player and sit down to
listen is...? Fill in the blank, Arny. Having trouble? I thought so.

Again, I'll put this more bluntly. How do I know that the results of ABX
testing correlate with ordinary listening? Where is your proof? Or just
strong evidence?

Arny, you're a perfect example of Pope's observation that "A little learning
is a dangerous thing." * Your mind is so full of "facts" that understanding
seems to have pushed out (if it ever was there). It's pathetic arguing with
someone who has no understanding of what Einstein meant when he said that
"Imagination is more important than knowledge."

You're a fine example of the "science nazi" who blathers and blathers and
blathers -- and attracts a coterie of True Believers -- but who fails to
convince.

* A little learning is a dangerous thing;
Drink deep, or taste not the Pierian spring:
There shallow draughts intoxicate the brain,
And drinking largely sobers us again.

Harry Lavo

April 16th 07, 02:58 PM

"Arny Krueger" > wrote in message
...
> "Harry Lavo" > wrote in message
>
>
>> Fact of the matter is, Arny, that I spend a lot of time
>> on the hi-rez forum in Audio Asylum....you don't..
>
> Yeah Harry, I'm too busy doing real- world audio and computer work to
> spend as much time as you do yapping on various audio forums.
>
> For example I did almost 100 live recordings for hiren the past month or
> so - how many did you do, Harry?
>
> I created 4 video DVDs ranging from 5 minutes to 45 minutes, some original
> work, some compendiums of other works that were major components of 4
> live performances, each attended by up to 360 people.
>
> I mixed 10 rehearsals and 5 live performances with a total attendance of
> over 1,500 people.
>
> I produced 5 CDs that were sold to the general public, tiny niche market
> products to be sure, but neverthess stuff that sold.
>
> That doesn't make me any kind of exceptional audio guy - there's people
> around here who did more in just the past few days, but what did you do,
> Harry? I'm quite sure you can't even bring yourself to admit what you did
> that would comparable, since it was likely zero.
>
>> .so we
>> *know* when Best Buy moved to take down their SACD and
>> DVD-A pretty much across the country.
>
> I provided considerable real-world evidence that says otherwise. Just
> another case of Harry trying to turn rumor into facts.
>
>> If you found a
>> store selling SACD or DVD-A then it is out of the
>> mainstream Best Buy chain.
>
> Yeah, like their web site is completely out of the mainstream Best Buy
> chain. Tell us another story, Harry.
>
>> And since Best Buy was hardly
>> even a presence on the web a year ago, when I (or
>> somebody else) refers to Best Buy they are usually
>> talking about the largest chain of retail bricks and
>> mortar stores in the country, not a come-lately web site.
>
> Yap, yap, yap, yap.
>
>> No amount of "debating trade" is going to obscure the
>> reality, Arny....Best Buy as a chain is pretty much out
>> of the DVD-A and SACD business.
>
> In fact Harry, every alleged fact that you presented was adequately
> falsifed by real-world evidence. And that was just a footnote on the story
> of the collapse of hi-rez media in the mainstream music market, as
> documented by no less than the RIAA.
>
> Denial isn't a river in Egypt - its a major component of Harry Lavo's
> life. Too bad about those people who spent the big bucks on DVD-A and SACD
> players. I hope they enjoy them, because the sun is setting on them.
> Slowly for sure, but the sun is setting and HD DVD and Blu-Ray have their
> sights set on their market segment.

All this posturing and and anti-SACD and anti-Harry ranting doesn't change
the original point that I made....that 2006 was the year that Best Buy (as
well as Circuit City) removed their SACD and DVD-A sections and eliminated
their inventory for the most part....thus possibily contributing to the
slump in hi-rez sales during the first half of 2005.

You can nitpick all you want, but this isn't RAO and I'm not going to play
"debating trade" games with you.

I think the discussion of hi-rez sales has run its course.

Steven Sullivan

April 16th 07, 03:19 PM

Harry Lavo > wrote:

> "Steven Sullivan" > wrote in message
> ...
> > Mike Rivers > wrote:
> >> On Apr 13, 7:09 am, "Arny Krueger" > wrote:
> >
> >> > One practical problem with Sony's DSD strategy is that the so-called
> >> > hardware problems of 196 KHz PCM converters were solved so thorougly
> >> > that
> >> > they can even be found in $39.95 DVD players.
> >
> >> You may not agree with Dan Lavry, but he doesn't think so. On the
> >> other hand, he doesn't make $39.95 DVD players, so what people do with
> >> those doesn't matter a lot fo him.
> >
> > Last I read Lavry -- which has been some months now --
> > he was not a proponent of 192, or even 96 kHz.
> > By his reckoning even the theoretical problems with rebdook standard are
> > solved by the time the sampling rate reaches ~ 60 kHz.
> >
> > Of course, he offers the 'industry standard' sampling rates on his gear to
> > stay competitive...that's just good business. Just like some
> > loudspeaker manufacturers offered a biwiring option and spiked
> > stands despite not being 'believers'.

> First, his units do not go above 96khz...for a reason that he argues
> cogently.

96 kHz being an industry standard...and he agrees it's overkill.

___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

Steven Sullivan

April 16th 07, 03:23 PM

William Sommerwerck > wrote:
> > But there's another fundamental issue -- the potential errancy
> > of subjective impression.

> That's precisely the reason I've argued for long-term listening. And not
> just for one auditioner, but many.

And again, what part of 'long term listening' isn't also 'subjective'?

Or do you mean long-term *double blind* listeneing? The main obstacle there
is doing enough trials to perform useful statistics.

> It's ludicrous to claim that "forced-decision" listening tests, under
> conditions that have nothing to do with the way we actually listen to music,
> are the last word. Useful, probably. But not definitive.

Sighted long term listening is arguably much less definitive.

___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

Steven Sullivan

April 16th 07, 03:26 PM

William Sommerwerck > wrote:
> > But even if it could be assumed in practice, deteriming an *inherent*
> > difference between a SACD and CD is a very difficult test, because
> > differences between playback chains still cannot be ruled out.

> What would be the difference in playback chains?

> When playing a hybrid disk, you're using the same D/A converters for both CD
> and SACD. At least, in the case of the SACD player I own, you are.

Most SACD players I've seen have separate user-adjustable settings for the
SACD versus CD outputs.

___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

Steven Sullivan

April 16th 07, 03:29 PM

Arny Krueger > wrote:
> "Steven Sullivan" > wrote in message
>
> > Scott Dorsey > wrote:
> >
> >> That's true. BUT, the good news about DSD is that it
> >> has power density issues that make it hard to do abusive
> >> processing and get away with it. That's a huge plus in
> >> my book.
> >
> > Can abusive processing be performed in the PCM domain,
> > then
> > the file be transcoded to DSD?

> Absolutely. It's already been done for commercial releases. The form of the
> abusive processing would be traditional audio processing as conducted in the
> pre-digital era.

I was thinking more or radical digital compression and limiting, as done
in the last 15 years or so.

___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

Arny Krueger

April 16th 07, 03:30 PM

"Harry Lavo" > wrote in message

> "Arny Krueger" > wrote in message
> ...
>> "Harry Lavo" > wrote in message
>>
>>
>>> Fact of the matter is, Arny, that I spend a lot of time
>>> on the hi-rez forum in Audio Asylum....you don't..
>>
>> Yeah Harry, I'm too busy doing real- world audio and
>> computer work to spend as much time as you do yapping on
>> various audio forums. For example I did almost 100 live recordings for
>> hiren
>> the past month or so - how many did you do, Harry?
>>
>> I created 4 video DVDs ranging from 5 minutes to 45
>> minutes, some original work, some compendiums of other
>> works that were major components of 4 live
>> performances, each attended by up to 360 people. I mixed 10 rehearsals
>> and 5 live performances with a
>> total attendance of over 1,500 people.
>>
>> I produced 5 CDs that were sold to the general public,
>> tiny niche market products to be sure, but neverthess
>> stuff that sold. That doesn't make me any kind of exceptional audio guy -
>> there's people around here who did more in just the past
>> few days, but what did you do, Harry? I'm quite sure you
>> can't even bring yourself to admit what you did that
>> would comparable, since it was likely zero.
>>> .so we
>>> *know* when Best Buy moved to take down their SACD and
>>> DVD-A pretty much across the country.
>>
>> I provided considerable real-world evidence that says
>> otherwise. Just another case of Harry trying to turn
>> rumor into facts.
>>> If you found a
>>> store selling SACD or DVD-A then it is out of the
>>> mainstream Best Buy chain.
>>
>> Yeah, like their web site is completely out of the
>> mainstream Best Buy chain. Tell us another story, Harry.
>>
>>> And since Best Buy was hardly
>>> even a presence on the web a year ago, when I (or
>>> somebody else) refers to Best Buy they are usually
>>> talking about the largest chain of retail bricks and
>>> mortar stores in the country, not a come-lately web
>>> site.
>>
>> Yap, yap, yap, yap.
>>
>>> No amount of "debating trade" is going to obscure the
>>> reality, Arny....Best Buy as a chain is pretty much out
>>> of the DVD-A and SACD business.
>>
>> In fact Harry, every alleged fact that you presented was
>> adequately falsifed by real-world evidence. And that was
>> just a footnote on the story of the collapse of hi-rez
>> media in the mainstream music market, as documented by
>> no less than the RIAA. Denial isn't a river in Egypt - its a major
>> component of
>> Harry Lavo's life. Too bad about those people who spent
>> the big bucks on DVD-A and SACD players. I hope they
>> enjoy them, because the sun is setting on them. Slowly
>> for sure, but the sun is setting and HD DVD and Blu-Ray
>> have their sights set on their market segment.

> All this posturing and and anti-SACD and anti-Harry
> ranting doesn't change the original point that I
> made....that 2006 was the year that Best Buy (as well as
> Circuit City) removed their SACD and DVD-A sections

Trouble is, said removal is invisible in BB's brick-and-mortar stores and
web site.

> and eliminated their inventory for the most part.

Whatever that means.

>...thus possibily contributing to the slump in hi-rez sales
> during the first half of 2005.

Now Harry we see the transformation of your original claim:

"2006 saw Best Buy and Circuit City drop SACD and DVD-A"

to

"eliminated their inventory for the most part."

Keep changing your story Harry - pretty soon you'll come up with a
believable tale.

> You can nitpick all you want, but this isn't RAO and I'm
> not going to play "debating trade" games with you.

Harry, you've been plying the debating trade for all you can squeeze out of
it, for who knows how long?

The example I quoted from you taken about 2 days apart, being just one
example of many.

> I think the discussion of hi-rez sales has run its course.

The RIAA seems to have the final word - hi-rez sales are diving towards the
bottom of the tank, and the only people who care the haplesss consumers who
invested in the formats.

Now that Blu Ray and DVD-HD are here, the hi-rez audio format schtick will
probably again be milked for all it is worth. Trouble is that the industry
knows that isn't much.

Here's what to expect:

Blu Ray:

"
Linear PCM (LPCM) - offers up to 8 channels of uncompressed audio.
Dolby Digital (DD) - format used for DVDs also known as AC3, offers
5.1-channel surround sound.
Dolby Digital Plus (DD+) - extension of DD, offers increased bitrates and
7.1-channel surround sound.
Dolby TrueHD - extension of MLP Lossless, offers lossless encoding of up to
8 channels of audio.
DTS Digital Surround - format used for DVDs, offers 5.1-channel surround
sound.
DTS-HD - extension of DTS, offers increased bitrates and up to 8 channels of
audio.
"

HD DVD:

"
HD DVD discs support encoding in up to 24-bit/192 kHz for two channels, or
up to eight channels of up to 24-bit/96 kHz encoding.[9] For reference, even
new big-budget Hollywood films are mastered in only 24-bit/48 kHz, with
16-bit/48 kHz being common for ordinary films.
All HD DVD players are required to decode linear (uncompressed) PCM, Dolby
Digital AC-3, Dolby Digital EX, DTS, Dolby Digital Plus and Dolby
TrueHD.[10] A secondary soundtrack, if present, can be stored in any of the
aforementioned formats, or in one of the HD DVD optional codecs: DTS-HD High
Resolution Audio and DTS-HD Master Audio.

For the highest-fidelity audio experience, HD DVD offers content-producers
the choice of linear PCM, Dolby TrueHD and DTS-HD Master Audio. Due to the
high-bandwidth requirements of linear-PCM, lossless audio on HD DVD movies
has thus far been delivered in the lossless format Dolby True-HD.

Please note that this simply means that all Blu-ray players and recorders
will have to support playback of these audio codecs, it will still be up to
the movie studios to decide which audio codec(s) they use for their
releases.
"

I'll leave it to the readers to figure out what well-known so-called hi-rez
format is missing.... ;-)

BTW, just so nobody thinks it is just me that thinks that the old hi-rez
audo formats failed in the marketplace:

http://www.audioholics.com/news/editorials/10-reasons-why-high-definition-dvd-formats-have-already-failed/

Steven Sullivan

April 16th 07, 03:31 PM

Scott Dorsey > wrote:
> Steven Sullivan > wrote:
> >Scott Dorsey > wrote:
> >
> >> That's true. BUT, the good news about DSD is that it has power
> >> density issues that make it hard to do abusive processing and get
> >> away with it. That's a huge plus in my book.
> >
> >Can abusive processing be performed in the PCM domain, then
> >the file be transcoded to DSD?

> Yes, but it doesn't make the end result any louder, so there is no point.

I guess I'm not sure what you're defining as 'abusive processing' then.

___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

Arny Krueger

April 16th 07, 03:41 PM

"William Sommerwerck" > wrote in
message
>>> Why should I perform a listening test that, in my
>>> opinion, has no correlation with the way people normally
>>> listen?
>
>> Don't bother William - you already know the *right*
>> answer.
>> You spent $2k on a fancy SACD player with hopes that were
>> technically impossible, and can't bring yourself to
>> admit its limitations.

> Wait a second... In a previous post, you said that the
> limitations of SACD (or more precisely, DSD) were at the
> threshold of audibility, if at all audible. Which is it?

The technically impossible hope was that the DSD format would have better
audible SQ than CD-A.

>>> How many times do we need to explain to you, Arny,
>>> that you're _assuming_ the validity of the experiment
>>> a priori?

>> Shows what little you know about the practice of science,
>> William. No sane person would knowingly spend any effort
>> at all on an experiment that had abosolutely no validity,
>> except for just the fun of it. And, if its only for fun,
>> then its
>> valid is only fun, and nothing that relates to anybody
>> with
>> a more serious intent.

> It's done every day, Arny. "Serious" researchers set up
> poorly designed experiments and/or misinterpret them. A
> serious intent is no guarantee of accuracy or utility.

A truism, but so what? A valiant attempt that fails is not the same as
assuming the validity of the experiment a priori.

>> Before one starts an experiment with any serious intent,
>> they must believe that its outcome has some relevance
>> and validity. It isn't blindly assumed, it has to have
>> some
>> logical reason for being.

> And that's the point -- ABX testing has no reason for
> being.

Try to sell that to the world.

> Because you haven't demonstrated (and I use that
> word deliberately) its connection with what people here
> when they sit down and listen to music in a hi-fi system.

Other than in the minds of a few obsessive audiophiles, the connection
between ABX and sitting down and listening are obvious to all who have
investigated it.

> You cannot assume there is a valid connection or
> correlation -- because the conditions of the test are not
> the same as common listening conditions, you have to
> demonstrate, to prove the connection.

The differences that exist are of little or no importance to all but a tiny
minority of true believers in various kinds of well-known snake oil,
including the so-called hi-rez audio distribution format.

>> So no, I didn't assume the validity of any of the ABX
>> experiements a priori. The belief in their validity was
>> based on the best possible investigation of what it took
>> to make a valid audio experiment, by a group of people
>> with surprisingly good credentials -- one person with a
>> BS in Math, several with BS+advanced course work in
>> electrical engineering, one with a BS in experimental
>> psychology, a medical practitiioner, teacher and
>> researcher; and several with years of professional
>> experience in audio production and product development.
>
>> You gonna say that's all just an a priori assumption,
>> William? Pshawww.... ;-)

> Yes, it is. Because we're using "validity" in two
> altogether different ways -- and your use is the wrong one.

Oh yes, I'm always wrong and you're always right, William. Pardon me for
deluding myself into thinking that my friends and I have a brain.

> Here's how Arny's thinking worked 30 years ago -- and
> still works: "Double-blind testing is a widely accepted
> protocol in scientific studies. Therefore, any form of
> testing using it must be scientific and must necessarily
> produce valid results."

The analysis and practice went a lot deeper than that.

The fact of the matter is that *everything* that the most recent detailed
studies of human hearing say is audible can be, and in many cases has been
confirmed by ABX tests.

The only things that ABX can't confirm is the garbage that a few frantic and
obsessive audiophiles say they can hear, that the study of psychoacoustics
and human hearing should be totally masked or otherwise inaudible. You
know, golden wires, magic amplifiers, blue CD pens, ultrasonic sample rates,
and dynamic range wildly in excess of what we observe in concert halls. ABX
doesn't confirm them, so let's sit around and think of wild theories about
why ABX isn't valid. Yes William, you can do that sort of thing on your
time, but not mine. I've got too much real world audio work to get done.

Scott Dorsey

April 16th 07, 03:49 PM

Harry Lavo > wrote:
>
>All this posturing and and anti-SACD and anti-Harry ranting doesn't change
>the original point that I made....that 2006 was the year that Best Buy (as
>well as Circuit City) removed their SACD and DVD-A sections and eliminated
>their inventory for the most part....thus possibily contributing to the
>slump in hi-rez sales during the first half of 2005.

Yes, and 2007 was the year that Tower Records shut down. First quarter
numbers for CDs look really depressing too.

>I think the discussion of hi-rez sales has run its course.

I don't think we can have a discussion of high resolution format sales
except in the context of ALL record sales. Because all record sales are
tanking dramatically, it's hard to say anything for sure about any niche
formats.

I hesitate to say anything except possibly to shake my arms like a character
in a Greek tragedy and shout "We're all DOOMED."
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Scott Dorsey

April 16th 07, 03:53 PM

Steven Sullivan > wrote:
>Scott Dorsey > wrote:
>> Steven Sullivan > wrote:
>> >Scott Dorsey > wrote:
>> >
>> >> That's true. BUT, the good news about DSD is that it has power
>> >> density issues that make it hard to do abusive processing and get
>> >> away with it. That's a huge plus in my book.
>> >
>> >Can abusive processing be performed in the PCM domain, then
>> >the file be transcoded to DSD?
>
>> Yes, but it doesn't make the end result any louder, so there is no point.
>
>I guess I'm not sure what you're defining as 'abusive processing' then.

Brickwall limiting that causes flattopping for the most part. You just
can't get away with it on SACD.

You can still compress the hell of of things to bring the average levels
way up, but the aggressive limiting buys you nothing. Much like it was
in the vinyl era except without the low end level issues.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Scott Dorsey

April 16th 07, 03:56 PM

Arny Krueger > wrote:
>"William Sommerwerck" > wrote in
>message
>>>> Why should I perform a listening test that, in my
>>>> opinion, has no correlation with the way people normally
>>>> listen?
>>
>>> Don't bother William - you already know the *right*
>>> answer.
>>> You spent $2k on a fancy SACD player with hopes that were
>>> technically impossible, and can't bring yourself to
>>> admit its limitations.
>
>> Wait a second... In a previous post, you said that the
>> limitations of SACD (or more precisely, DSD) were at the
>> threshold of audibility, if at all audible. Which is it?
>
>The technically impossible hope was that the DSD format would have better
>audible SQ than CD-A.

The problem is that formats don't HAVE sound quality. Only implementations
have sound quality.

The argument with DSD is that it's _easier_ to build higher sound quality
implementations with it. Whether this is true or not has yet to really be
determined. I suspect, sadly, that it will be a moot point in the age of
low-fi MP3 distribution.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[email protected]

April 16th 07, 04:31 PM

On Apr 16, 10:41 am, "Arny Krueger" > wrote:

> Oh yes, I'm always wrong .... Pardon me for
> deluding myself into thinking that my friends and I have a brain.

I will not suggest such,
(how many friends for a brain ... arny +3= a brain)

(I'll go to my room now and not make any noise....)

maybe a tin ear with friends

> why ABX isn't valid. Yes ..., you can do that sort of thing on your
> time, but not mine. I've got too much real world audio work to get done.

okay abx is valid, now will you settle down and go do you "listening"
etudes???

[email protected]

April 16th 07, 04:34 PM

William Sommerwerck

April 16th 07, 04:44 PM

>> Wait a second... In a previous post, you said that the
>> limitations of SACD (or more precisely, DSD) were at the
>> threshold of audibility, if at all audible. Which is it?

> The technically impossible hope was that the DSD format
> would have better audible SQ than CD-A.

But you define sound quality in a very narrow way, according to easily
measurable parameters that supposedly are the only ones that have any
meaningful subjective correlation with sound quality.

> Other than in the minds of a few obsessive audiophiles, the connection
> between ABX and sitting down and listening are obvious to all who have
> investigated it.

There is NO connection Arny. They're not the same thing.

>> You cannot assume there is a valid connection or
>> correlation -- because the conditions of the test are not
>> the same as common listening conditions, you have to
>> demonstrate, to prove the connection.

> The differences that exist are of little or no importance to all but a
tiny
> minority of true believers in various kinds of well-known snake oil,
> including the so-called hi-rez audio distribution format.

The differences are the sine qua non of this whole issue.

People... Arny Krueger (whose initials, by the way, are short for a very
rude desriptive noun in Yiddish) is not worth discussing anything with. He
assumes that if he believes something, it is necessarily true, and then he
twists everything around to make it come out true in this own mind.

If Arny _was_ ever able to argue intelligently about anything, he's lost it.
He's out in his own little world, incapable of understanding any other point
of view but his own.

> Oh yes, I'm always wrong and you're always right, William.
> Pardon me for deluding myself into thinking that my friends
> and I have a brain.

Uh... Neither you nor your friends express what I would consider
"intelligence". I apologize for understanding things better than you. Sorry.
Just a filthy little habit I've acquired over the years.

>> Here's how Arny's thinking worked 30 years ago -- and
>> still works: "Double-blind testing is a widely accepted
>> protocol in scientific studies. Therefore, any form of
>> testing using it must be scientific and must necessarily
>> produce valid results."

> The analysis and practice went a lot deeper than that.

You've produced nothing to demonstrate this.

When I listen to music, I put on a favorite disk and plop down in the chair.
I don't run ABX comparisons of one amplifier or CD player with another.
Please tell me the correlation between these. You never have, and never
will, because has ever researched the correlation.

> The fact of the matter is that *everything* that the most recent detailed
> studies of human hearing say is audible can be, and in many cases has
> been confirmed by ABX tests.

> The only things that ABX can't confirm is the garbage that a few frantic
and
> obsessive audiophiles say they can hear, that the study of psychoacoustics
> and human hearing should be totally masked or otherwise inaudible.

Light needs a medium to propagate. Right? We know that, because it's just
common sense, isn't it?

> You know, golden wires, magic amplifiers, blue [sic] CD pens, ultrasonic
> sample rates, and dynamic range wildly in excess of what we observe
> in concert halls. ABX doesn't confirm them, so let's sit around and think
> of wild theories about why ABX isn't valid.

One doesn't need wild theories, merely the common sense to recognize that
the lack of an audible difference in ABX testing only proves that it doesn't
show up under that form of testing. The difference might very well NOT be
audible under ANY conditions. But its absence in one form of testing is not
proof of its non-existance.

What any honest researcher tries to do is to determine the truth -- not to
prove (or disprove) preconceived notions or theories. It's obvious that
there's nothing intellectually honest about Arny's research. Never has been.
It's always been directed at confirming his view of "scientific fact".

What Arny is actually saying here (and you can look at his other postings to
confirm this) is... If a device has distortion and noise characteristics
below a certain measurable level, then its behavior is indistinguishable
from any other device measuring similarly. And ABX testing "proves" this.

Arny is, in a sense, psychotic. Instead of trying to understand "what's
going on", he wants the world to behave in the way he _thinks_ it ought to
behave, according to his "scientific" view of things. Had he been running
the Michelson-Morley experiment, he would have screwed with the test setup
to "bugger" the experiment in such a way as to show that the Luminiferous
Ether actually existed -- and then defended his work on the basis that it
_has_ to exist because our "scientific" knowledge of wave propagation
requires it to.

Arny, there is so much useful work you could have done in audio -- and you
haven't. You've almost nothing of value, because you simply don't understand
the limitations of your work. They're only a useful beginning -- they're
hardly not the last word.

William Sommerwerck

April 16th 07, 04:46 PM

>> You're a fine example of the "science nazi" who blathers and blathers
>> and blathers -- and attracts a coterie of True Believers -- but who
>> fails to convince.

> Typically words from someone who states that the statement "Newtons 2nd
> law describes the behaviour of moving objects" is "crap" I dont think it
> is possible to "convince" that kind of "True Believers", who detests
> experiments and only hear what they want to hear and have payed for.

If you're back to the topic of cone breakup (not just the behavior of moving
objects) -- Newton's 2nd Law, by itself, tells us nothing about cone
breakup. Other factors are required.

Arny Krueger

April 16th 07, 05:38 PM

"William Sommerwerck" > wrote in
message
>>> Wait a second... In a previous post, you said that the
>>> limitations of SACD (or more precisely, DSD) were at the
>>> threshold of audibility, if at all audible. Which is it?

>> The technically impossible hope was that the DSD format
>> would have better audible SQ than CD-A.

> But you define sound quality in a very narrow way,

No, I use two very broad measures. Bandwidth and dynamic range are sort of
the like the executive summary of audio performance.

> according to easily measurable parameters that supposedly
> are the only ones that have any meaningful subjective
> correlation with sound quality.

Name something that is generally agreed-upon to adversly affect sonic
performance that doesn't adversly affect dynamic range and bandwidth.

>> Other than in the minds of a few obsessive audiophiles,
>> the connection between ABX and sitting down and
>> listening are obvious to all who have investigated it.

> There is NO connection Arny. They're not the same thing.

Hyperbole is fun, but it has got to be used in moderation, William. There
are a zillion things that are not the same thing, like $0.99 and $1.00, but
that are strongly connected. Ya know, they are both money, right? ;-)

ABX is first and foremost composed of sitting down and listening. I know of
no way to do an ABX test without sitting down and listening, do you William?

>>> You cannot assume there is a valid connection or
>>> correlation -- because the conditions of the test are
>>> not the same as common listening conditions, you have to
>>> demonstrate, to prove the connection.

>> The differences that exist are of little or no
>> importance to all but a tiny minority of true believers
>> in various kinds of well-known snake oil, including the
>> so-called hi-rez audio distribution format.

> The differences are the sine qua non of this whole issue.

Answer me how a person can do an ABX test without sitting down and
listening, and you've got a deal, William. If you know how to do this, I'm
sure that at minimum, all the original ABX partners and the 30-odd surviving
(two have passed AFAIK) SMWTMS members that participated in the ABX tests
through the ones that were published in Stereo Review in the late 80s will
salute you. Listening can be a lot of work, and if there's some way to do
ABX without listening, it would be a great time and trouble saver.

> People... Arny Krueger (whose initials, by the way, are
> short for a very rude desriptive noun in Yiddish) is not
> worth discussing anything with. He assumes that if he
> believes something, it is necessarily true, and then he
> twists everything around to make it come out true in this
> own mind.

Just answer me how one can do an ABX test without listening, William.

> If Arny _was_ ever able to argue intelligently about
> anything, he's lost it. He's out in his own little world,
> incapable of understanding any other point of view but
> his own.

Just answer me how one can do an ABX test without listening, William.

>> Oh yes, I'm always wrong and you're always right,
>> William.
>> Pardon me for deluding myself into thinking that my
>> friends
>> and I have a brain.

> Uh... Neither you nor your friends express what I would
> consider "intelligence".

We did a good job of fooling the audio industry. ;-)

> I apologize for understanding
> things better than you. Sorry. Just a filthy little habit
> I've acquired over the years.

Yeah William. Sure. I'm sure there's something, even about audio that you
understand better than I. It would be interesting to find out what that is
were time available and you could coherently express it...

But when it comes to listening tests...

When it comes to characterizing equipment performance...

William, I haven't seen anything from you that I didn't already know that
was worthy of a salute. Here's your chance - help me with a difficult
problem that I've never resoloved. Just answer me how one can do an ABX test
without listening, William.

>>> Here's how Arny's thinking worked 30 years ago -- and
>>> still works: "Double-blind testing is a widely accepted
>>> protocol in scientific studies. Therefore, any form of
>>> testing using it must be scientific and must necessarily
>>> produce valid results."
>
>> The analysis and practice went a lot deeper than that.
>
> You've produced nothing to demonstrate this.

Sure I did.

> When I listen to music, I put on a favorite disk and plop
> down in the chair.

Just answer me how one can do an ABX test without doing something exactly
that, William. Well one way is to use non-disc media which I admit I do a
lot. But I think you know what I'm trying to say. Don't you?

> I don't run ABX comparisons of one
> amplifier or CD player with another.

I wasn't able to get through life so easily. Something about thirst for
knowlege about what matters about sound quality with some troubling issues
resolved. The big troubling issue was bias.

> Please tell me the
> correlation between these. You never have, and never
> will, because has ever researched the correlation.

What about all the times I did an ABX test by putting on a favorite disk and
plopping down in my favorite listening chair?

>> The fact of the matter is that *everything* that the
>> most recent detailed studies of human hearing say is
>> audible can be, and in many cases has been confirmed by
>> ABX tests.

>> The only things that ABX can't confirm is the garbage
>> that a few frantic and obsessive audiophiles say they
>> can hear, that the study of psychoacoustics and human
>> hearing should be totally masked or otherwise inaudible.

> Light needs a medium to propagate. Right? We know that,
> because it's just common sense, isn't it?

OK, so???

>> You know, golden wires, magic amplifiers, blue [sic] CD
>> pens, ultrasonic sample rates, and dynamic range wildly
>> in excess of what we observe
>> in concert halls. ABX doesn't confirm them, so let's sit
>> around and think of wild theories about why ABX isn't
>> valid.

> One doesn't need wild theories, merely the common sense
> to recognize that the lack of an audible difference in
> ABX testing only proves that it doesn't show up under
> that form of testing.

But, ABX correlates with what shows up or not with ABC/hr tests,
same/different tests, whaterver it was Dolby did in their jitter tests, even
many sighted tests, etc., etc. It correlates with the tests that
Fletcher/Munson did and the redos over the years, however they did them. ABX
tests often provides more sensitive results than one finds in standard tests
for JNDs.

ABX just doesn't do well with CD pens and the like. It does do a good job of
detecting problems with dynamic range and bandwidth, in the cases where we
know that dyanmic range and bandwidth adversely affect SQ.

> The difference might very well NOT
> be audible under ANY conditions. But its absence in one
> form of testing is not proof of its non-existance.

Never said it was. But after you spend say 30 years looking for a pot of
gold at the end of many, many rainbows and find none, what does that say
about rainbows?

> What any honest researcher tries to do is to determine
> the truth -- not to prove (or disprove) preconceived
> notions or theories.

Been there and done that.

> It's obvious that there's nothing
> intellectually honest about Arny's research. Never has
> been. It's always been directed at confirming his view of
> "scientific fact".

Nonsense. I built the first ABX comparator and did the first ABX test
sincerily hoping that I would prove the *truth* that was being published in
the early issues of The Absolute Sound to be right, and thus *educate*
certain nay-sayers. I'd still like to make Tom Nousaine, Dave Clark, and
Stan Lip****z eat crow. I know them all personally, you know. I stopped
holding my breath about that a few years back! ;-)

> What Arny is actually saying here (and you can look at
> his other postings to confirm this) is... If a device has
> distortion and noise characteristics below a certain
> measurable level, then its behavior is indistinguishable
> from any other device measuring similarly. And ABX
> testing "proves" this.

So does every other reliable listening test.

> Arny is, in a sense, psychotic. Instead of trying to
> understand "what's going on", he wants the world to
> behave in the way he _thinks_ it ought to behave,
> according to his "scientific" view of things. Had he been
> running the Michelson-Morley experiment, he would have
> screwed with the test setup to "bugger" the experiment in
> such a way as to show that the Luminiferous Ether
> actually existed -- and then defended his work on the
> basis that it _has_ to exist because our "scientific"
> knowledge of wave propagation requires it to.

Nonsense.

> Arny, there is so much useful work you could have done in
> audio -- and you haven't. You've almost nothing of value,
> because you simply don't understand the limitations of
> your work. They're only a useful beginning -- they're
> hardly not the last word.

I'll stand on what I've done. Stuff like the 46,000 Google mentions of
"PCABX audio".

Mentions like these:

http://www.edn.com/article/CA74935.html

http://www.bostonaudiosociety.org/bas_speaker/wishful_thinking.htm

http://www.hydrogenaudio.org/forums/lofiversion/index.php/t4088.html

http://www.soundonsound.com/sos/nov00/articles/pcnotes.htm

http://www.stereophile.com/news/050905debate/

http://www.mytekdigital.com/compare/comparison1.htm

http://www.music.miami.edu/programs/mue/Research/sbrowne/thesis.pdf

http://www.24bitfaq.org/

http://forumz.tomshardware.com/ce/Difference-Betweeen-96khz-192khz-ftopict49110.html

http://www.ocf.berkeley.edu/~ashon/audio/phase/phaseaud2.htm

http://club.cdfreaks.com/showthread.php?t=54734

http://forums.audioholics.com/forums/archive/index.php/t-3556.html

http://www.silcom.com/~aludwig/Sound_demos.html

http://ask.slashdot.org/comments.pl?sid=9868&threshold=1&commentsort=0&mode=thread&cid=516328

http://www.badscience.net/?p=209

Arny Krueger

April 16th 07, 05:44 PM

"Chel van Gennip" > wrote in message

> On Mon, 16 Apr 2007 17:46:39 +0200, William Sommerwerck
> wrote:
>
>>>> You're a fine example of the "science nazi" who
>>>> blathers and blathers and blathers -- and attracts a
>>>> coterie of True Believers -- but who fails to convince.
>>
>>> Typically words from someone who states that the
>>> statement "Newtons 2nd law describes the behaviour of
>>> moving objects" is "crap" I dont think it is possible
>>> to "convince" that kind of "True Believers", who
>>> detests experiments and only hear what they want to
>>> hear and have payed for.
>>
>> If you're back to the topic of cone breakup (not just
>> the behavior of moving objects) -- Newton's 2nd Law, by
>> itself, tells us nothing about cone breakup. Other
>> factors are required.
>
> As I wrote, it is impossible to convince a "True
> Believer" like you that the forces needed to accelerate
> tha mass of the cone, as described in Newton's 2nd Law,
> are the reason of cone breakup.

Agreed. First we hear that F=MA has nothing to do with cone breakup. Then
without admitting the rather grotesque first error, we are told that
"Newton's 2nd Law, by itself, tells us nothing about cone breakup" which is
obviously true but obfuscates the first error.

Dooh, our model also needs a spring and a damper not just a mass. Nobody
ever said otherwise except for the one who said that F=MA had nothing to do
with cone breakup.

hank alrich

April 16th 07, 05:57 PM

Chel van Gennip > wrote:

> It seems obvious that 10 extra years of development in digital and analog
> electronics made that advantage vanish.

Why would those very factors not apply to the approach that was easier
in the first place?

> Secondly it is hard to speak about "higher sound quality" when widly
> accepted tests don't show an audible improvement.

Does every CD you've heard offer the exact same sound quality, or is
there a variety of results, based not on a medium, but upon the
practices that led to the content on that disc?

--
ha
Iraq is Arabic for Vietnam

Arny Krueger

April 16th 07, 06:03 PM

"hank alrich" > wrote in message
...
> Chel van Gennip > wrote:
>
>> It seems obvious that 10 extra years of development in digital and analog
>> electronics made that advantage vanish.
>
> Why would those very factors not apply to the approach that was easier
> in the first place?

Because the ease of implementation for the two alternatives has progressed
to the point where relative ease isn't an issue.

For example Crystal Semiconductor has several chips for decoding SACD, but
they all also do PCM.

DSD advantage: lost

>> Secondly it is hard to speak about "higher sound quality" when widly
>> accepted tests don't show an audible improvement.

> Does every CD you've heard offer the exact same sound quality, or is
> there a variety of results, based not on a medium, but upon the
> practices that led to the content on that disc?

That's the point - CD's vary greatly in terms of SQ, but the limits set by
the CD medium itself are far from intrusive. They are moot for a
distribution format.

Harry Lavo

April 16th 07, 09:57 PM

"Scott Dorsey" > wrote in message
...
> Arny Krueger > wrote:
>>"William Sommerwerck" > wrote in
>>message
>>>>> Why should I perform a listening test that, in my
>>>>> opinion, has no correlation with the way people normally
>>>>> listen?
>>>
>>>> Don't bother William - you already know the *right*
>>>> answer.
>>>> You spent $2k on a fancy SACD player with hopes that were
>>>> technically impossible, and can't bring yourself to
>>>> admit its limitations.
>>
>>> Wait a second... In a previous post, you said that the
>>> limitations of SACD (or more precisely, DSD) were at the
>>> threshold of audibility, if at all audible. Which is it?
>>
>>The technically impossible hope was that the DSD format would have better
>>audible SQ than CD-A.
>
> The problem is that formats don't HAVE sound quality. Only
> implementations
> have sound quality.
>
> The argument with DSD is that it's _easier_ to build higher sound quality
> implementations with it. Whether this is true or not has yet to really be
> determined. I suspect, sadly, that it will be a moot point in the age of
> low-fi MP3 distribution.

Scott, just copied this out of the Acoustic Sounds ebulletin. Nice to see
somebody is still trying:
Music SACD Hybrid Hybrid Multichannel SACD

Blue Coast Collection
The E.S.E. Sessions

Various Artists
Label:Blue Coast Records
Genre:Various
Product No.:CBCR 1012 SA
Availability:In Stock
Format:Hybrid Multichannel SACD
NEW ARRIVAL

$29.99

Acoustic Sounds Exclusive
Blue Coast Records Sets a New Standard in Audio Excellence

"One of the most natural sounding recordings I've ever heard. Very
enjoyable."
David Schwartz - FOUNDER OF MIX MAGAZINE

"The sound production is so clear you can drink it from the tap!"
Allen Butler- PRESIDENT, SONY MUSIC NASHVILLE, 1993-2003

Label founder Cookie Marenco, a five-time Grammy nominated
producer/engineer, has developed a patented-pending surround sound recording
technique called Extended Sound Environment (ESE) that creates a holographic
sonic image. ESE makes use of room and direct microphone placement to
optimize full tonal response from the instruments and character of the room.
Marenco assembled a cast of world-class engineers and technicians noted for
their commitment to audio excellence. The disc features the work of 10
acoustic artists of various styles, and each recording used only the highest
quality gear, including silver cabling throughout the process. No
headphones, digital effects or overdubs were used. And while the musicians
are relatively unknown, they are of top caliber and were hand picked for
their performance abilities.

TRANSLATION: This is an audiophile release of the highest possible
standards, acoustic music in its most lifelike form.

Also, I just noticed in this bulletin a new RCA classical SACD (not Living
Presence) recorded within the last ten years. Could it be?

William Sommerwerck

April 16th 07, 10:36 PM

> ABX is first and foremost composed of sitting down and listening.
> I know of no way to do an ABX test without sitting down and
> listening, do you William?

I know lots of ways of sitting down and listening that are not ABX testing.

>> Uh... Neither you nor your friends express what I would
>> consider "intelligence".

> We did a good job of fooling the audio industry. ;-)

Which also consists of a lot of idiots -- on _both_ sides of this issue.

> William, I haven't seen anything from you that I didn't already know
> that was worthy of a salute. Here's your chance - help me with a
> difficult problem that I've never resoloved. Just answer me how one
> can do an ABX test without listening, William.

Non-sequitur. You're trying to claim that any form of listening is
perceptually equivalent to any other form of listening. It's not.

> Just answer me how one can do an ABX test without doing
> something exactly that, William.

But ABX testing isn't done that way. It's formal, not casual.

> I wasn't able to get through life so easily. Something about thirst
> for knowlege about what matters about sound quality with some
> troubling issues resolved. The big troubling issue was bias.

Bias is only one of the issues.

> ABX just doesn't do well with CD pens and the like. It does do a
> good job of detecting problems with dynamic range and bandwidth,
> in the cases where we know that dyanmic range and bandwidth
> adversely affect SQ.

Arny, because you sound "reasonable" at this (very) brief moment in time
(but don't worry, it'll pass), I'm going to give you one more chance (though
I shouldn't, because you'll twist things around in your self-serving,
weasely way).

I'm an ex-subjective reviewer, who quit, in part, because he felt he wasn't
providing reliable information to his readers. But I haven't forgotten a
number of troubling listening experiences over the years.

One of them was the proximate cause of my quitting. I was auditioning
cables -- just for fun. I knew which were connected, etc. I was in a good
mood -- I'd learned that the best time to listen was early on a weekend
morning, when it was quiet, your ears were rested, etc. I heard very
noticeable differences among the cables, and found that I preferred (note
the word) the $50 cable from one company over a $350 cable from another
company. The best of all was the $400 cable from the first company, which
sounded qualitatively like their $50 cable, but "eversomuchmoreso". (Yes, I
read Beverly Cleary when I was a kid.)

The "better" (to me) cables had more depth; the others were "flatter" and
more "forward"-sounding. Out of curiosity (and, in my book, like a good
scientist), I used one cable from each pair, and got an ear-twisting result
that sounded as if I were hearing the recording from two different
perspectives at the same time. This was strong evidence (not proof) that I
was hearing _real_ differences, as it seemed unlikely my mind could cook up
such a sonic mess and present it convincingly.

I should add the all the perceived differences among the cables were gross,
in-your-face effects, and not the least-bit subtle.

So interconnects really do sound different, right? They must, as I audtioned
them under conditions I knew would maximize my ability to hear differences.

Not so fast.

The next morning I packed up the cables to ship back, but decided to first
give them another listen. This time, I was prejudiced to hear what I'd heard
the day before. But there were _no_ differences among the cables. They all
sounded the same. I was in a similar mood, and played the same recordings,
but got grossly different (subjective) results.

So, which audition revealed "the truth"? Don't say "Neither, because there
was no control." One of them _had_ to be much closer to the truth than the
other, because both of them couldn't have been correct. I couldn't tell
which, because I was caught up (as we all are) in my own mental space, and
had no way to "look outside" to determine the state of my mind or
perceptions.

So are we to believe that ABX testing does not impose its own "mental space"
on the listeners? That it provides a neutral test bed in which the listeners
will hear what is actually there? I want proof that it does.

I'll give you an example that suggests this might not be the case with ABX
testing, that it has its own set of problems.

Back in the early '80s, I bought an Acoustat system -- amps, preamps,
speakers, all designed by Jim Strickland. After six months or so of
listening, I decided I didn't like the tonal balance -- it sounded
consistently "sway-backed", as if the top and the bottom were "up" a skosh
from the midrange. The system had _not_ sounded like this when I first
bought it -- that is, I had not _perceived_ it as sounding this way.

Most listeners have experienced something like this -- what sounds neutral
at first gradually begins revealing its colorations. I contend that it's
possible for a piece of electronics to have subtle colorations that don't
show up in initial listening -- and are not audible in even a direct
comparison with a neutral bypass -- until the listener has learned to hear
them. In other words, ABX testing won't work with small errors that are not
immediately noticeable, but need time to gradually reveal themselves.

This is speculation, of course, but no worse than assuming ABX testing is
correct and complete.

I really want to know "the truth" -- and I don't feel either side in this
debate is delivering it. Or knows what it's talking about, for that matter.

>> It's obvious that there's nothing intellectually honest about
>> Arny's research. Never has been. It's always been directed
>> at confirming his view of "scientific fact".

> Nonsense. I built the first ABX comparator and did the first ABX
> test sincerely hoping that I would prove the *truth* that was being
> published in the early issues of The Absolute Sound to be right,
> and thus *educate* certain nay-sayers.

And _everything_ that TAS, The Stereophile, et al, claimed was wrong? That
there wasn't a single product evaluation that they "got right"? This isn't a
rhetorical question. ABX testing disproved every claim of sonic differences
among electronics?

Forgive me, Arny, but an "intelligent" person wouldn't assume that the
failure of ABX testing to confirm subjective reactions necessarily _proves_
the latter to be incorrect -- particularly if there were some agreement
among the subjective reviewers. Rather he might react -- "What the hell is
going on here?", as I did. I want to know what's going on. And I'm not
stupid enough to think that an uncorrelated "scientifc" test is going to
supply all the answers any more than uncontrolled anecdotal listening.

I want long-term double-blind "casual" listening tests -- running two or
three years -- with a least a dozen listeners and as many different
amplifiers. I want to see if, at the end of that time, there's any consensus
as to what the amplifiers "sound like", and whether such a consensus carries
over to ABX testing. Once such a test is done, we'll begin to _understand_
the strengths and weaknesses of both types of testing and how they correlate
(or not) with each other.

If anyone would like to discuss setting up such a test, I'd be happy to
discuss it with them. But I'm not holding my breath.

> > What Arny is actually saying here (and you can look at
> > his other postings to confirm this) is... If a device has
> > distortion and noise characteristics below a certain
> > measurable level, then its behavior is indistinguishable
> > from any other device measuring similarly. And ABX
> > testing "proves" this.

> So does every other reliable listening test.

In your book, a reliable listening test is one whose conclusions fit your
preconceptions.

> I'll stand on what I've done. Stuff like the 46,000 Google mentions
> of "PCABX audio".

I get 35,600,000 for Hitler.

[email protected]

April 17th 07, 01:29 AM

Hi all! Acoustic Sounds has the first copies of the SACD in the
United States. We brought it to the Salon HiFI Show in Paris last
month and received an extremely warm reception. We have been selling
the cd and now the SACD from a hidden page on our website to our club
members. If you want to preview the music before buying, please do!
you can find it at www.bluecoastrecords.com

We're hoping for more releases this year, reviving and remastering
some incredible performances for SA-CD... hopefully no more plants
will close. We love the format!
Check out the previews and let us know what you think. It's
multigenre, live in the studio, no headphones, no overdubs, no digital
efx.

Best, and glad to know there are music and sound lovers still around!
Cookie Marenco
founder, producer
Blue Coast Records

Chris Hornbeck

April 17th 07, 03:10 AM

On Mon, 16 Apr 2007 10:41:00 -0400, "Arny Krueger" >
wrote:

>Other than in the minds of a few obsessive audiophiles, the connection
>between ABX and sitting down and listening are obvious to all who have
>investigated it.

The validity or non- or this argument is the crux of the bisquit.

Much thanks, as always,

Chris Hornbeck

"Beauty will save the world."
- Feodor Dostoevsky

"Information is not knowledge. Knowledge is not wisdom. Truth is not
beauty. Beauty is not love. Love is not music."
- Frank Zappa

Chris Hornbeck

April 17th 07, 03:13 AM

On Mon, 16 Apr 2007 12:38:07 -0400, "Arny Krueger" >
wrote:

>Name something that is generally agreed-upon to adversly affect sonic
>performance that doesn't adversly affect dynamic range and bandwidth.

Monotonicity. Crossover distortion. Small signal threshold
effects. Quantization. Off the top of my head.

All good fortune,

Chris Hornbeck

"Beauty will save the world."
- Feodor Dostoevsky

"Information is not knowledge. Knowledge is not wisdom. Truth is not
beauty. Beauty is not love. Love is not music."
- Frank Zappa

William Sommerwerck

April 17th 07, 03:34 AM

>> Name something that is generally agreed-upon to adversly affect sonic
>> performance that doesn't adversly affect dynamic range and bandwidth.

> Monotonicity. Crossover distortion. Small signal threshold
> effects. Quantization. Off the top of my head.

Arny will tell you that all those can subsumed under "dynamic range".

Chris Hornbeck

April 17th 07, 03:50 AM

On Mon, 16 Apr 2007 19:34:32 -0700, "William Sommerwerck"
> wrote:

>>> Name something that is generally agreed-upon to adversly affect sonic
>>> performance that doesn't adversly affect dynamic range and bandwidth.
>
>> Monotonicity. Crossover distortion. Small signal threshold
>> effects. Quantization. Off the top of my head.
>
>Arny will tell you that all those can subsumed under "dynamic range".

All deviations from perfection fall into the categories of
either amplitude or time. Whether we appreciate them properly
or ignore them, ie. whether they're included in our numbers
or not, is relevant.

Much thanks, as always,

Chris Hornbeck

"Beauty will save the world."
- Feodor Dostoevsky

"Information is not knowledge. Knowledge is not wisdom. Truth is not
beauty. Beauty is not love. Love is not music."
- Frank Zappa

William Sommerwerck

April 17th 07, 12:32 PM

>>>> Name something that is generally agreed-upon to adversly affect sonic
>>>> performance that doesn't adversly affect dynamic range and bandwidth.

>>> Monotonicity. Crossover distortion. Small signal threshold
>>> effects. Quantization. Off the top of my head.

>> Arny will tell you that all those can subsumed under "dynamic range".

> All deviations from perfection fall into the categories of
> either amplitude or time. Whether we appreciate them properly
> or ignore them, ie. whether they're included in our numbers
> or not, is relevant.

Harmonic and IM distortion can be considered as amplitude errors. But we
don't _hear_ them as alterations to dynamics.

Arny Krueger

April 17th 07, 12:36 PM

"William Sommerwerck" > wrote in message
...

>>> Other than in the minds of a few obsessive audiophiles, the connection
>>> between ABX and sitting down and listening are obvious to all who have
>>> investigated it.

>>>There is NO connection Arny. They're not the same thing.

>> ABX is first and foremost composed of sitting down and listening.

>> I know of no way to do an ABX test without sitting down and
>> listening, do you William?

> I know lots of ways of sitting down and listening that are not ABX
> testing.

How does this answer the question: Do you know of any way to do an ABX test
without sitting down and listening?

>>> Uh... Neither you nor your friends express what I would
>>> consider "intelligence".

>> We did a good job of fooling the audio industry. ;-)

> Which also consists of a lot of idiots -- on _both_ sides of this issue.

>> William, I haven't seen anything from you that I didn't already know
>> that was worthy of a salute. Here's your chance - help me with a
>> difficult problem that I've never resoloved. Just answer me how one
>> can do an ABX test without listening, William.

> Non-sequitur. You're trying to claim that any form of listening is
> perceptually equivalent to any other form of listening. It's not.

William, you said there is no connection at all. How is the above relevant
to that claim?

>> Just answer me how one can do an ABX test without doing
>> something exactly that, William.

> But ABX testing isn't done that way. It's formal, not casual.

It's as formal as you make it. It's as informal as you make it.

Several of us have experimented with long-term ABX listening. We set an ABX
comparator to a certain position and leave it that way for hours or even
days without changing it or even looking at it. We just listened and thought
informally about what we thought we heard.

>> I wasn't able to get through life so easily. Something about thirst
>> for knowlege about what matters about sound quality with some
>> troubling issues resolved. The big troubling issue was bias.

> Bias is only one of the issues.

A truism. Sheds no specifical light on the question at hand.

>> ABX just doesn't do well with CD pens and the like. It does do a
>> good job of detecting problems with dynamic range and bandwidth,
>> in the cases where we know that dyanmic range and bandwidth
>> adversely affect SQ.

> Arny, because you sound "reasonable" at this (very) brief moment in time
> (but don't worry, it'll pass), I'm going to give you one more chance
> (though
> I shouldn't, because you'll twist things around in your self-serving,
> weasely way).

Gosh William could you point your nose at a steeper angle and look down it
further?

> I'm an ex-subjective reviewer, who quit, in part, because he felt he
> wasn't
> providing reliable information to his readers. But I haven't forgotten a
> number of troubling listening experiences over the years.

> One of them was the proximate cause of my quitting. I was auditioning
> cables -- just for fun. I knew which were connected, etc. I was in a good
> mood -- I'd learned that the best time to listen was early on a weekend
> morning, when it was quiet, your ears were rested, etc. I heard very
> noticeable differences among the cables, and found that I preferred (note
> the word) the $50 cable from one company over a $350 cable from another
> company. The best of all was the $400 cable from the first company, which
> sounded qualitatively like their $50 cable, but "eversomuchmoreso". (Yes,
> I
> read Beverly Cleary when I was a kid.)

> The "better" (to me) cables had more depth; the others were "flatter" and
> more "forward"-sounding. Out of curiosity (and, in my book, like a good
> scientist), I used one cable from each pair, and got an ear-twisting
> result
> that sounded as if I were hearing the recording from two different
> perspectives at the same time. This was strong evidence (not proof) that I
> was hearing _real_ differences, as it seemed unlikely my mind could cook
> up
> such a sonic mess and present it convincingly.

> I should add the all the perceived differences among the cables were
> gross,
> in-your-face effects, and not the least-bit subtle.

Oh yes, and the kind of thing that completely disappears when done under
bias-controlled conditions.

> So interconnects really do sound different, right? They must, as I
> audtioned
> them under conditions I knew would maximize my ability to hear
> differences.

> Not so fast.

> The next morning I packed up the cables to ship back, but decided to first
> give them another listen. This time, I was prejudiced to hear what I'd
> heard
> the day before. But there were _no_ differences among the cables. They all
> sounded the same. I was in a similar mood, and played the same recordings,
> but got grossly different (subjective) results.
>
> So, which audition revealed "the truth"? Don't say "Neither, because there
> was no control." One of them _had_ to be much closer to the truth than the
> other, because both of them couldn't have been correct. I couldn't tell
> which, because I was caught up (as we all are) in my own mental space, and
> had no way to "look outside" to determine the state of my mind or
> perceptions.

A test with inadequate controls is like a room full of a million monkeys and
keyboards. You know the rest.

The idea that one of the monkeys is a better speller because the garbage he
typed came closer to being the Gettyberg Address for a few seconds is of
course, completely missing the point.

> So are we to believe that ABX testing does not impose its own "mental
> space"
> on the listeners?

It's whatever you want it to be.

> That it provides a neutral test bed in which the listeners
> will hear what is actually there? I want proof that it does.

ABX is no different than any other kind of critical listening.

> I'll give you an example that suggests this might not be the case with ABX
> testing, that it has its own set of problems.

> Back in the early '80s, I bought an Acoustat system -- amps, preamps,
> speakers, all designed by Jim Strickland. After six months or so of
> listening, I decided I didn't like the tonal balance -- it sounded
> consistently "sway-backed", as if the top and the bottom were "up" a skosh
> from the midrange. The system had _not_ sounded like this when I first
> bought it -- that is, I had not _perceived_ it as sounding this way.

> Most listeners have experienced something like this -- what sounds neutral
> at first gradually begins revealing its colorations. I contend that it's
> possible for a piece of electronics to have subtle colorations that don't
> show up in initial listening -- and are not audible in even a direct
> comparison with a neutral bypass -- until the listener has learned to hear
> them.

William, I can tell you think that:

(1) This is something that myself and the ABXers don't know.

(2) It's a deep insight.

Horsefeathers on both counts. We've known this for decades and include
allowances for it in our test procedures.

William, you would be wise to stop talking down to people like me who
learned more about listening than you can posssibly ever know, because
unlike you, we've been working with reliable measures that help us separate
sensitivity from illusion, and have been doing so for decades.

Arny Krueger

April 17th 07, 12:48 PM

"Chris Hornbeck" > wrote in message
...
> On Mon, 16 Apr 2007 12:38:07 -0400, "Arny Krueger" >
> wrote:
>
>>Name something that is generally agreed-upon to adversly affect sonic
>>performance that doesn't adversly affect dynamic range and bandwidth.

> Monotonicity. Crossover distortion. Small signal threshold
> effects. Quantization. Off the top of my head.

They all adversely affect dynamic range by creating spurious responses. This
effectively raises the noise floor.

Dynamic Range is *not* the same thing as SNR.

Arny Krueger

April 17th 07, 12:48 PM

"William Sommerwerck" > wrote in message
...
>>> Name something that is generally agreed-upon to adversly affect sonic
>>> performance that doesn't adversly affect dynamic range and bandwidth.
>
>> Monotonicity. Crossover distortion. Small signal threshold
>> effects. Quantization. Off the top of my head.
>
> Arny will tell you that all those can subsumed under "dynamic range".

Right, because they all raise the noise floor with a signal present.

Arny Krueger

April 17th 07, 12:49 PM

"Chris Hornbeck" > wrote in message
...
> On Mon, 16 Apr 2007 19:34:32 -0700, "William Sommerwerck"
> > wrote:
>
>>>> Name something that is generally agreed-upon to adversly affect sonic
>>>> performance that doesn't adversly affect dynamic range and bandwidth.
>>
>>> Monotonicity. Crossover distortion. Small signal threshold
>>> effects. Quantization. Off the top of my head.
>>
>>Arny will tell you that all those can subsumed under "dynamic range".
>
> All deviations from perfection fall into the categories of
> either amplitude or time. Whether we appreciate them properly
> or ignore them, ie. whether they're included in our numbers
> or not, is relevant.

They're not ignored by proper evaluations of dynamic range.

Again, dynamic range is *not* SNR.

Arny Krueger

April 17th 07, 12:56 PM

"William Sommerwerck" > wrote in message
. ..
>>>>> Name something that is generally agreed-upon to adversly affect sonic
>>>>> performance that doesn't adversly affect dynamic range and bandwidth.
>
>>>> Monotonicity. Crossover distortion. Small signal threshold
>>>> effects. Quantization. Off the top of my head.
>
>>> Arny will tell you that all those can subsumed under "dynamic range".
>
>> All deviations from perfection fall into the categories of
>> either amplitude or time. Whether we appreciate them properly
>> or ignore them, ie. whether they're included in our numbers
>> or not, is relevant.

> Harmonic and IM distortion can be considered as amplitude errors. But we
> don't _hear_ them as alterations to dynamics.

Actually, nonlinear distortion does change dynamics, and also often has the
effect of additionally changing the perception of dynamics. IOW, nonlinear
distortion can make a signal sound louder than it actually is. People
exploit this to make their songs appear to be louder.

An ideal compressor is a nonlinear amplifier where the local gain is based
on a time-weighted measure of the amplitude of the signal, not its
instantaneous amplitude. If it wasn't for the time-weighted part, it would
just be another ordinary nonlinear amplifier.

Arny Krueger

April 17th 07, 01:00 PM

"Chris Hornbeck" > wrote in message
...
> On Mon, 16 Apr 2007 10:41:00 -0400, "Arny Krueger" >
> wrote:
>
>>Other than in the minds of a few obsessive audiophiles, the connection
>>between ABX and sitting down and listening are obvious to all who have
>>investigated it.
>
> The validity or non- or this argument is the crux of the bisquit.

Agreed. And the source of the problem is the fear and ignorance of the
audiophiles. For about 30 years many audiophiles have been sitting around
making up self-pitying stories about how horrible and stupid reliable
subjective testing is.

There are very few people with significant experience actually doing
reliable subjective tests that don't respect the process.

Having a reliable measure and applying it, is hugely effective in fostering
progress within a technology.

William Sommerwerck

April 17th 07, 01:37 PM

> How does this answer the question: Do you know of any way
> to do an ABX test without sitting down and listening?

The question is meaningless, and I can't believe you asked it -- let alone
take it seriously.

>> Non-sequitur. You're trying to claim that any form of listening is
>> perceptually equivalent to any other form of listening. It's not.

> William, you said there is no connection at all. How is the above
> relevant to that claim?

My terse response might not have been clear. What I meant was that an ABX
test is inherently formal, and is not equivalent to casual listening. The
fact that the listeners are expect to make a decision "formalizes" it.

> Several of us have experimented with long-term ABX listening.
> We set an ABX comparator to a certain position and leave it
> that way for hours or even days without changing it or even
> looking at it. We just listened and thought informally about
> what we thought we heard.

That's good, but it's still not casual listening.

>> Arny, because you sound "reasonable" at this (very) brief
>> moment in time (but don't worry, it'll pass), I'm going to give
>> you one more chance (though I shouldn't, because you'll twist
>> things around in your self-serving, weasely way).

> Gosh, William, could you point your nose at a steeper angle
> and look down it further?

I'm not responsible for your inability to understand what is incomplete
about your thinking and testing.

>> I should add the all the perceived differences among the cables
>> were gross, in-your-face effects, and not the least-bit subtle.

> Oh yes, and the kind of thing that completely disappears
> when done under bias-controlled conditions.

And that "proves" they don't exist?

This, more than anything, shows the difference in how Arny and I think. Arny
want to prove or disprove something. I want to understand what's going on.
There is a wide gulf between such points of view.

> A test with inadequate controls is like a room full of a million
> monkeys and keyboards. You know the rest.

Badly mixed metaphor, Arny.

In other words, the use of controls guarantees correct results? You've got
to be kidding!

It doesn't matter how good the controls are, if the experiment doesn't "ask
the right question". Why can't you understand this? What is so difficult
about this not-at-all-subtle philosophical point to understand?

> The idea that one of the monkeys is a better speller because
> the garbage he typed came closer to being the Gettyberg Address
> for a few seconds is of course, completely missing the point.

Bad, bad analogy. We KNOW that monkeys don't have language and can't spell.
We DO NOT know whether or not cables "sound different".

I agree that cables _shouldn't_ "sound different", assuming they're driven
by a load-insensitive source. But my assumption about such things proves
nothing, and the failure of a double-blind test to replicate these results
is merely evidence, not proof.

>> So are we to believe that ABX testing does not impose
>> its own "mental space" on the listeners?

> It's whatever you want it to be.

What is that supposed to mean? How can a listener arbitrarily choose their
mental state, or the person running the test impose a desirable mental state
on the participants?

> William, you would be wise to stop talking down to people like me who
> learned more about listening than you can posssibly ever know, because
> unlike you, we've been working with reliable measures that help us
separate
> sensitivity from illusion, and have been doing so for decades.

You're exactly the sort of person who needs talking-down-to -- and an
intellectual dressing-down. You've been performing experiments that confirm
what you would like to believe -- that your parochial view of what "science"
comprises is always correct. You've only barely begun to perform the
experiments needed to _demonstrate_ that what you claim is true.

I would dearly love for some Famous Scientist (such as the late Richard
Feynman) to carefully examine our arguments and points of view, and then say
"Yes, William, your view of what a good scientific experiment is closer to
the truth" and "Arny's conclusions might be correct, but they're not
experimentally valid". But it's not possible, and even if it were, I
wouldn't do it, because I don't like "appealing to authority" to resolve an
issue. What can I say to you, Arny, to convince you that your overall
approach to science is just plain wrong, something that no real scientist
would take seriously?

William Sommerwerck

April 17th 07, 01:42 PM

> An ideal compressor is a nonlinear amplifier where the local gain
> is based on a time-weighted measure of the amplitude of the signal,
> not its instantaneous amplitude. If it wasn't for the time-weighted part,
> it would just be another ordinary nonlinear amplifier.

Ignoring our running argument for a moment...

What you say is broadly correct, but the gain changes nevertheless introduce
IM. (These effects also occur in surround decoders that use dynamic
crosstalk cancellation.) I'm not sure that you can say one is "ordinary" and
the other not, except for the fact that dynamic-range controllers aren't
common components.

I recently sold my dbx 5BX. I'd used it very little, because I didn't like
the way it sounded. The expansion was virtually inaudible, but it did
"something" to the sound that I can't put into words.

William Sommerwerck

April 17th 07, 01:44 PM

>>> Other than in the minds of a few obsessive audiophiles, the connection
>>> between ABX and sitting down and listening are obvious to all who have
>>> investigated it.

>> The validity or non- or this argument is the crux of the bisquit.

> Agreed. And the source of the problem is the fear and ignorance
> of the audiophiles. For about 30 years many audiophiles have been
> sitting around making up self-pitying stories about how horrible and
> stupid reliable subjective testing is.

> There are very few people with significant experience actually doing
> reliable subjective tests that don't respect the process.

> Having a reliable measure and applying it, is hugely effective in
fostering
> progress within a technology.

But you've never demonstrated the "reliability" of ABX testing -- outside of
its own little narrow universe.

The issue remains, and won't go away until someone actually performs valid
experiments.

Arny Krueger

April 17th 07, 02:19 PM

"William Sommerwerck" > wrote in message
. ..

> What I meant was that an ABX
> test is inherently formal, and is not equivalent to casual listening. The
> fact that the listeners are expect to make a decision "formalizes" it.

And that differs from listening to support a purchase decision, how?

>> Several of us have experimented with long-term ABX listening.
>> We set an ABX comparator to a certain position and leave it
>> that way for hours or even days without changing it or even
>> looking at it. We just listened and thought informally about
>> what we thought we heard.

> That's good, but it's still not casual listening.

Neither is listening with the idea of writing a review, or just talking to a
friend.

>>> Arny, because you sound "reasonable" at this (very) brief
>>> moment in time (but don't worry, it'll pass), I'm going to give
>>> you one more chance (though I shouldn't, because you'll twist
>>> things around in your self-serving, weasely way).
>
>> Gosh, William, could you point your nose at a steeper angle
>> and look down it further?

> I'm not responsible for your inability to understand what is incomplete
> about your thinking and testing.

See what I said? ;-)

>>> I should add the all the perceived differences among the cables
>>> were gross, in-your-face effects, and not the least-bit subtle.

>> Oh yes, and the kind of thing that completely disappears
>> when done under bias-controlled conditions.

> And that "proves" they don't exist?

What is proof? Let's all do a chorus of "All findings of science are
provisional, and only relevant until we have better evidence and.or
analysis" Now that the sing-along is over lets get serious:

Provisional proof for audio production people is typified by say, a
purchase decision, a decision to repair or replace, a recommendation to
obtain, a release of a design to manufactring, a mic or gobo choice or
position, a decision to make this the final take, mix or master, etc.

> This, more than anything, shows the difference in how Arny and I think.

For openers, I think early and often. I'm still active as a designer and
producer.

> Arny want to prove or disprove something.

No, I just want to get a first-rate or at least an appropriate satisfactory
job done.

>I want to understand what's going on.

In the past 20 years, I've no doubt designed and built more electronic gear
than 99% of the people who post here.

> There is a wide gulf between such points of view.

Yup, its the difference between producers and consumers.

>> a million monkeys and keyboards. You know the rest.

> Badly mixed metaphor, Arny.

Downward pointed nose noted.

> In other words, the use of controls guarantees correct results? You've got
> to be kidding!

Since you said it William, it can be as stupid as you like.

What I'm saying is that the absence of appropriate controls guarantees
chaos. Controls are a necessary but not necessarily sufficient condition.
Good brakes won't make a car with a broken engine run. But, a car with
broken brakes and a working engine isn't a very good car.

> It doesn't matter how good the controls are, if the experiment doesn't
> "ask
> the right question". Why can't you understand this?

Because I understand that in ways you've never been able to comprehend, but
your ego won't let you see it, William.

> What is so difficult
> about this not-at-all-subtle philosophical point to understand?

You tell me, William because its the absence of understanding of this point
that has sent your life off into the woods. If you and Atkinson both
understood this, you'd both be very different men, and probably a lot closer
and happier with each other today than you are.

>> The idea that one of the monkeys is a better speller because
>> the garbage he typed came closer to being the Gettyberg Address
>> for a few seconds is of course, completely missing the point.

> Bad, bad analogy. We KNOW that monkeys don't have language and can't
> spell.
> We DO NOT know whether or not cables "sound different".

Actually, we know that some do, and some don't. Being effective in audio
production is partially contingent on managing the sonic properties of
cables on a scale that transcends mere audiophilia by several orders of
magnitude.

William, you're no doubt heard the old saw about the futility of arguing
with people who buy ink by the barrel. You need to recognize the futility of
arguing cables with people who go through 500 and 1,000 foot rolls of
high-quality shileded cable like rolls of toilet paper. Been there, done
that.

> I agree that cables _shouldn't_ "sound different", assuming they're driven
> by a load-insensitive source. But my assumption about such things proves
> nothing, and the failure of a double-blind test to replicate these results
> is merely evidence, not proof.

I guess we need another chorus of: "All findings of science are provisional,
and only relevant until we have better evidence and.or analysis" Now that
the sing-along is over, lets get back to our regularly scheduled
programming:

>>> So are we to believe that ABX testing does not impose
>>> its own "mental space" on the listeners?

>> It's whatever you want it to be.

> What is that supposed to mean? How can a listener arbitrarily choose their
> mental state, or the person running the test impose a desirable mental
> state
> on the participants?

The same way you do it for a sighted evaluation.

>> William, you would be wise to stop talking down to people like me who
>> learned more about listening than you can posssibly ever know, because
>> unlike you, we've been working with reliable measures that help us
> separate sensitivity from illusion, and have been doing so for decades.

> You're exactly the sort of person who needs talking-down-to -- and an
> intellectual dressing-down.

Given all the ordinary people, PhDs and industry luminaries that have
praised the approach I'm espousing, you're going to have do more than you
can possibly do, William.

> You've been performing experiments that confirm
> what you would like to believe

Insufffable delusions of mind-reading noted.

> -- that your parochial view of what "science" comprises is always
> correct.

Me always correct? LOL! However, I found out long ago that there's a
considerable advantage in simply being more correct, but obviously still
somewhat erroneous, than your competition.

> You've only barely begun to perform the
> experiments needed to _demonstrate_ that what you claim is true.

What did I claim - that Information Theory is relevant to audio? Why do I
need to defend *that* among well-educated people?

> I would dearly love for some Famous Scientist (such as the late Richard
> Feynman) to carefully examine our arguments and points of view, and then
> say
> "Yes, William, your view of what a good scientific experiment is closer to
> the truth" and "Arny's conclusions might be correct, but they're not
> experimentally valid".

I'll settle for the JAES review board and AES board of governors. ;-)

> But it's not possible, and even if it were, I
> wouldn't do it, because I don't like "appealing to authority" to resolve
> an
> issue. What can I say to you, Arny, to convince you that your overall
> approach to science is just plain wrong, something that no real scientist
> would take seriously?

Just get the AES to rescind their explicit approval of the stuff I wrote or
helped write, that they've published. ;-)

Harry Lavo

April 17th 07, 02:56 PM

"Arny Krueger" > wrote in message
...
>
> "William Sommerwerck" > wrote in message
> ...
>

>snip<

>> But ABX testing isn't done that way. It's formal, not casual.

It's as formal as you make it. It's as informal as you make it.

Several of us have experimented with long-term ABX listening. We set an ABX
comparator to a certain position and leave it that way for hours or even
days without changing it or even looking at it. We just listened and thought
informally about what we thought we heard

>snip<

That is hardly the "scientific" validation that a new test technique
requires if it is to be accepted as valid.

William Sommerwerck

April 17th 07, 02:59 PM

>> What I meant was that an ABX test is inherently formal, and
>> is not equivalent to casual listening. The fact that the listeners
>> are expected to make a decision "formalizes" it.

> And that differs from listening to support a purchase decision, how?

It doesn't, and that's a valid point. (I was going to bring it up, but
decided not to.) Comparing two or more products, deciding which to buy,
introduces the same sort of "make a decision" psychological effects that
occur in ABX testing. The kind of casual-but-controlled testing I'd like to
perform minimizes these.

>> That's good, but it's still not casual listening.

> Neither is listening with the idea of writing a review,
> or just talking to a friend.

Agreed, but that's not the kind of "casual" testing I want to perform.

>> I'm not responsible for your inability to understand what
>> is incomplete about your thinking and testing.

> See what I said? ;-)

If Arny were a good scientist, HE would be questioning the validity and
completeness of his work. I wouldn't have to do it.

>>>> I should add the all the perceived differences among the cables
>>>> were gross, in-your-face effects, and not the least-bit subtle.

>>> Oh yes, and the kind of thing that completely disappears
>>> when done under bias-controlled conditions.

>> And that "proves" they don't exist?

> What is proof? Let's all do a chorus of "All findings of science are
> provisional, and only relevant until we have better evidence and.or
> analysis".

See what I mean when I gripe about Arny's weasling? What he just said is
correct, but he also wants us to believe that the results of his testing are
"definitive". (See his preceding comments about the rest of the audio world
having moved on.)

>> It doesn't matter how good the controls are, if the experiment doesn't
>> "ask the right question". Why can't you understand this?

> Because I understand that in ways you've never been able to comprehend,
> but but your ego won't let you see it, William.

What ego? I have no ego in these matters, Arny. That's not a joke.

>> What is so difficult
>> about this not-at-all-subtle philosophical point to understand?

> You tell me, William because its the absence of understanding of this
> point that has sent your life off into the woods. If you and Atkinson both
> understood this, you'd both be very different men, and probably a lot
closer
> and happier with each other today than you are.

John Atkinson has as an intellectually clogged mind as you, Arny. I wanted
to be friends with him, but gradually discovered he is incapable of having
an intelligent conversation with anyone, about anything. He knows what he
knows, and that's it.

I'm not interested in "being happy" with each other. He's married and not
really "my type". (I like skinny li'l bearded geezers, and I've got my
sights on one at the moment.)

>>> The idea that one of the monkeys is a better speller because
>>> the garbage he typed came closer to being the Gettyberg Address
>>> for a few seconds is of course, completely missing the point.

>> Bad, bad analogy. We KNOW that monkeys don't have language
>> and can't spell. We DO NOT know whether or not cables
>> "sound different".

> Actually, we know that some do, and some don't.

I've never fully believed that signing chimps truly have language in the
sense that humans do.

> Being effective in audio production is partially contingent on managing
> the sonic properties of cables on a scale that transcends mere
> audiophilia by several orders of magnitude.

> William, you're no doubt heard the old saw about the futility of arguing
> with people who buy ink by the barrel. You need to recognize the futility
of
> arguing cables with people who go through 500 and 1,000 foot rolls of
> high-quality shileded cable like rolls of toilet paper. Been there, done
> that.

I think Arny's intellectual confusion is plain. I see no point in spilling
more electronic ink over it.

>> What is that supposed to mean? How can a listener arbitrarily choose
>> their mental state, or the person running the test impose a desirable
>> mental state on the participants?

> The same way you do it for a sighted evaluation.

If you can arbitrarily set your mental state, Arny, you're a better man than
I.

>>> William, you would be wise to stop talking down to people like me who
>>> learned more about listening than you can posssibly ever know, because
>>> unlike you, we've been working with reliable measures that help us
>>> separate sensitivity from illusion, and have been doing so for decades.

>> You're exactly the sort of person who needs talking-down-to -- and an
>> intellectual dressing-down.

> Given all the ordinary people, PhDs and industry luminaries that have
> praised the approach I'm espousing, you're going to have do more than
> you can possibly do, William.

Ever heard the expression "a majority of one"?

I've decided to stop responding at this point. Arny is not mentally ill, in
any sense, but he's wandered off into a corridor of intellectual
self-delusion that should be plain. His jumping back and forth between
contradictory points of view is "proof" <ar, ar>.

I will end by saying that I really don't care what the members or leadership
of the JAES thinks about anything. I'm a member, by the way, and both Jon
Dahlquist and Saul Marantz approved me for membership. So there. I can drop
names, too.

Arny Krueger

April 17th 07, 03:39 PM

"William Sommerwerck" > wrote in message
. ..

>> An ideal compressor is a nonlinear amplifier where the local gain
>> is based on a time-weighted measure of the amplitude of the signal,
>> not its instantaneous amplitude. If it wasn't for the time-weighted part,
>> it would just be another ordinary nonlinear amplifier.

> Ignoring our running argument for a moment...

> What you say is broadly correct, but the gain changes nevertheless
> introduce
> IM.

Of a sort. Did I say otherwise?

> (These effects also occur in surround decoders that use dynamic
> crosstalk cancellation.) I'm not sure that you can say one is "ordinary"
> and
> the other not, except for the fact that dynamic-range controllers aren't
> common components.

I can say what I want, if at least a few people get my drift. ;-)

> I recently sold my dbx 5BX. I'd used it very little, because I didn't like
> the way it sounded. The expansion was virtually inaudible, but it did
> "something" to the sound that I can't put into words.

Whatever. I still have a legacy UREI limiter (LA4?) in the amplifier stack
at church... I could probably sell it and buy a new amplifier stack!

Arny Krueger

April 17th 07, 03:55 PM

"William Sommerwerck" > wrote in message
...
>>> What I meant was that an ABX test is inherently formal, and
>>> is not equivalent to casual listening. The fact that the listeners
>>> are expected to make a decision "formalizes" it.
>
>> And that differs from listening to support a purchase decision, how?

> It doesn't, and that's a valid point. (I was going to bring it up, but
> decided not to.) Comparing two or more products, deciding which to buy,
> introduces the same sort of "make a decision" psychological effects that
> occur in ABX testing. The kind of casual-but-controlled testing I'd like
> to
> perform minimizes these.

>>> That's good, but it's still not casual listening.

>> Neither is listening with the idea of writing a review,
>> or just talking to a friend.

> Agreed, but that's not the kind of "casual" testing I want to perform.

Well William, if you haven't figured it out yet, let me be clearer. I'm not
here to be your servant, and I'm not here to create the perfect listening
test. I'm here to get some jobs done. I don't *care* about some undefined
ideal test, I want a test that will get me through my next technical
decision based on sound quality.

>>> I'm not responsible for your inability to understand what
>>> is incomplete about your thinking and testing.
>
>> See what I said? ;-)

> If Arny were a good scientist, HE would be questioning the validity and
> completeness of his work. I wouldn't have to do it.

William, I don't need you to question me. I can do far better myself because
of my greater relevant experience and education. When that doesn't make it,
I've got friends and associates with far more experience and far better
credentials to talk to.

>>>>> I should add the all the perceived differences among the cables
>>>>> were gross, in-your-face effects, and not the least-bit subtle.
>
>>>> Oh yes, and the kind of thing that completely disappears
>>>> when done under bias-controlled conditions.
>
>>> And that "proves" they don't exist?
>
>> What is proof? Let's all do a chorus of "All findings of science are
>> provisional, and only relevant until we have better evidence and.or
>> analysis".

> See what I mean when I gripe about Arny's weasling?

It's not a weasel, its something that apparently you don't want to live
with.

> What he just said is
> correct, but he also wants us to believe that the results of his testing
> are
> "definitive". (See his preceding comments about the rest of the audio
> world
> having moved on.)

Non sequitor. First off, I just said that the results of my tests are
provisional, not definitive. In common speech, they are what they are -
take them or leave them. That the mainstream audio world has moved on from
primitive bias-ridden listening tests is a well known fact.

>>> It doesn't matter how good the controls are, if the experiment doesn't
>>> "ask the right question". Why can't you understand this?

>> Because I understand that in ways you've never been able to comprehend,
>> but but your ego won't let you see it, William.

> What ego? I have no ego in these matters, Arny. That's not a joke.

LOL!

>>> What is so difficult
>>> about this not-at-all-subtle philosophical point to understand?

>> You tell me, William because its the absence of understanding of this
>> point that has sent your life off into the woods. If you and Atkinson
>> both
>> understood this, you'd both be very different men, and probably a lot
>> closer
>> and happier with each other today than you are.

> John Atkinson has as an intellectually clogged mind as you, Arny.

No, far worse. And that's pretty bad.

> I wanted
> to be friends with him, but gradually discovered he is incapable of having
> an intelligent conversation with anyone, about anything. He knows what he
> knows, and that's it.

Which makes him different from you how, William? ;-)

You see, I don't know exactly what I know, and I certainly don't know
exactly what I don't know. But I know that I don't know it all, and I know
that just about everything I've ever done or will do will or has been
improved on. For example, I'm painfully aware of how ABC/hr is for many
applications a great improvement on ABX.

> I'm not interested in "being happy" with each other.

I'm in favor in intrapersonal peace.

> He's married and not
> really "my type". (I like skinny li'l bearded geezers, and I've got my
> sights on one at the moment.)

Well, I think you know about me. ;-)

>>>> The idea that one of the monkeys is a better speller because
>>>> the garbage he typed came closer to being the Gettyberg Address
>>>> for a few seconds is of course, completely missing the point.

>>> Bad, bad analogy. We KNOW that monkeys don't have language
>>> and can't spell. We DO NOT know whether or not cables
>>> "sound different".

>> Actually, we know that some do, and some don't.

> I've never fully believed that signing chimps truly have language in the
> sense that humans do.

Whatever.

>> Being effective in audio production is partially contingent on managing
>> the sonic properties of cables on a scale that transcends mere
>> audiophilia by several orders of magnitude.

>> William, you're no doubt heard the old saw about the futility of arguing
>> with people who buy ink by the barrel. You need to recognize the futility
>> of
>> arguing cables with people who go through 500 and 1,000 foot rolls of
>> high-quality shileded cable like rolls of toilet paper. Been there, done
>> that.

> I think Arny's intellectual confusion is plain. I see no point in spilling
> more electronic ink over it.

>>> What is that supposed to mean? How can a listener arbitrarily choose
>>> their mental state, or the person running the test impose a desirable
>>> mental state on the participants?
>
>> The same way you do it for a sighted evaluation.

> If you can arbitrarily set your mental state, Arny, you're a better man
> than
> I.

Who said anything about arbitrarily? But, I probably know more than the
average bear about attracting, engaging, and transforming people. BTW, that
implies directing their state of mind.

>>>> William, you would be wise to stop talking down to people like me who
>>>> learned more about listening than you can posssibly ever know, because
>>>> unlike you, we've been working with reliable measures that help us
>>>> separate sensitivity from illusion, and have been doing so for decades.
>
>>> You're exactly the sort of person who needs talking-down-to -- and an
>>> intellectual dressing-down.

>> Given all the ordinary people, PhDs and industry luminaries that have
>> praised the approach I'm espousing, you're going to have do more than
>> you can possibly do, William.

> Ever heard the expression "a majority of one"?

Yeah, as applied to this William guy. I'm not buying it.

> I've decided to stop responding at this point. Arny is not mentally ill,
> in
> any sense, but he's wandered off into a corridor of intellectual
> self-delusion that should be plain. His jumping back and forth between
> contradictory points of view is "proof" <ar, ar>.

> I will end by saying that I really don't care what the members or
> leadership
> of the JAES thinks about anything. I'm a member, by the way, and both Jon
> Dahlquist and Saul Marantz approved me for membership. So there. I can
> drop
> names, too.

Did I drop a name?

They guys who signed my AES app are relatively unknown compared to Dahlquist
and Marantz. But, they have designed audio systems that have arguably
provided far more listening pleasure to far more people, and they are still
at it.

William Sommerwerck

April 17th 07, 04:35 PM

> Well William, if you haven't figured it out yet, let me be clearer.
> I'm not here to be your servant, and I'm not here to create the
> perfect listening test.

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

What do you think your claims for ABX testing are?

> I'm here to get some jobs done. I don't *care* about some undefined
> ideal test, I want a test that will get me through my next technical
> decision based on sound quality.

But you claim that your form of testing provides a definitive evaluation of
sound quality -- THEN you say "nothing is ever defninitive".

There is a very fine -- and I claim meaningless -- line between "definitive"
and "perfect".

What an intellectual weasel you are.

> William, I don't need you to question me.

I'm exaxtly the person you need to question. I'm not your friend, and I
don't agree with you. That's the sort of person you need to keep you honest.

You are obviously very bothered that someone of about your age, IQ, and
education refuses to ingest your codswallop. Who refuses to blindly accept
what all the sycophants who hang on your ever word believe.

> Non sequitor. First off, I just said that the results of my tests are
> provisional, not definitive. In common speech, they are what they are -
> take them or leave them. That the mainstream audio world has moved
> on from primitive bias-ridden listening tests is a well-known fact.

You want people to believe that your "provisional" results are definitive.
Then when someone points out how incomplete they are, you won't listen. You
want credit for research that you cannot demonstrate the validity of.

>> I wanted to be friends with him, but gradually discovered he is
>> incapable of having an intelligent conversation with anyone,
>> about anything. He knows what he knows, and that's it.

> Which makes him different from you how, William? ;-)

What's with these continuing "wink signs"?

Well, for one thing, I have lots of friends I have intelligent conversations
with, that I don't always agree with. If I were to tell (names omitted) I
thought they were wrong about something, they would take me seriously and
thoughfully consider what I have to say.

> You see, I don't know exactly what I know, and I certainly don't
> know exactly what I don't know. But I know that I don't know it all,
> and I know that just about everything I've ever done or will do will
> or has been improved on.

So why is it that when someone points out your intellectual errors, you
don't seriously consider them? That's wise, that's smart, that's mature?

My number is 425-235-9579. You may call me at any time (including the middle
of the night), and we'll discuss whatever you care to discuss. But you
won't, because you're afraid that, as we're talking, the light bulb will
suddenly go on, and the intellectual sandpile you've built your research on
will slide away.

> Who said anything about arbitrarily? But, I probably know more than
> the average bear about attracting, engaging, and transforming people.
> BTW, that implies directing their state of mind.

To what ends? Please be specific.

> Did I drop a name?

You did, more or less. Rather less than more, but you appealed to the belief
that you have the respect of people the rest of us are supposed to look up
to.

Steven Sullivan

April 17th 07, 05:44 PM

Scott Dorsey > wrote:
> Harry Lavo > wrote:
> >
> >All this posturing and and anti-SACD and anti-Harry ranting doesn't change
> >the original point that I made....that 2006 was the year that Best Buy (as
> >well as Circuit City) removed their SACD and DVD-A sections and eliminated
> >their inventory for the most part....thus possibily contributing to the
> >slump in hi-rez sales during the first half of 2005.

> Yes, and 2007 was the year that Tower Records shut down. First quarter
> numbers for CDs look really depressing too.

> >I think the discussion of hi-rez sales has run its course.

> I don't think we can have a discussion of high resolution format sales
> except in the context of ALL record sales. Because all record sales are
> tanking dramatically, it's hard to say anything for sure about any niche
> formats.

> I hesitate to say anything except possibly to shake my arms like a character
> in a Greek tragedy and shout "We're all DOOMED."

People are still listening to recorded music, and lots of it. They're just obtaining it in
shall we say nontraditional ways. The industry will figure it out eventually. Recording
professionals aren't doom, the demand for the 'product' is still there.

What may be doomed, though, for the time being at least, is the 'album' as we knew and
loved it...but's lets remember that prior to the mid 60's, most pop albums
were not conceived as 'works' in themselves, just convenient repositories
for hit singles + filler.

___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

Steven Sullivan

April 17th 07, 06:19 PM

William Sommerwerck > wrote:
> > ABX is first and foremost composed of sitting down and listening.
> > I know of no way to do an ABX test without sitting down and
> > listening, do you William?

> I know lots of ways of sitting down and listening that are not ABX testing.

Yes, and which of those circumvent the problem of bias?

You listen to something on CD then on SACD, by whatever sighted method, long term, short term,
whatever. You decide the SACD, which has been marketed as a sonic improvement on
CD, sounds better. Then you conclude: it sounds better *because it's an SACD*.

If my years of reading audio forums is any indication, this is a typical 'audiophile' result
of 'listening that is not ABX testing'. Do you still not see how essentially dubious it is?

> One of them was the proximate cause of my quitting. I was auditioning
> cables -- just for fun. I knew which were connected, etc. I was in a good
> mood -- I'd learned that the best time to listen was early on a weekend
> morning, when it was quiet, your ears were rested, etc. I heard very
> noticeable differences among the cables, and found that I preferred (note
> the word) the $50 cable from one company over a $350 cable from another
> company. The best of all was the $400 cable from the first company, which
> sounded qualitatively like their $50 cable, but "eversomuchmoreso". (Yes, I
> read Beverly Cleary when I was a kid.)

> The "better" (to me) cables had more depth; the others were "flatter" and
> more "forward"-sounding. Out of curiosity (and, in my book, like a good
> scientist), I used one cable from each pair, and got an ear-twisting result
> that sounded as if I were hearing the recording from two different
> perspectives at the same time. This was strong evidence (not proof) that I
> was hearing _real_ differences, as it seemed unlikely my mind could cook up
> such a sonic mess and present it convincingly.

> I should add the all the perceived differences among the cables were gross,
> in-your-face effects, and not the least-bit subtle.

> So interconnects really do sound different, right? They must, as I audtioned
> them under conditions I knew would maximize my ability to hear differences.

> Not so fast.

> The next morning I packed up the cables to ship back, but decided to first
> give them another listen. This time, I was prejudiced to hear what I'd heard
> the day before. But there were _no_ differences among the cables. They all
> sounded the same. I was in a similar mood, and played the same recordings,
> but got grossly different (subjective) results.

> So, which audition revealed "the truth"? Don't say "Neither, because there
> was no control." One of them _had_ to be much closer to the truth than the
> other, because both of them couldn't have been correct. I couldn't tell
> which, because I was caught up (as we all are) in my own mental space, and
> had no way to "look outside" to determine the state of my mind or
> perceptions.

Study of human psychology is rife with instances of 'inevitable illusions' cast up by the
mind, that lead people to INSIST that they saw/heard something that simply wasn't real. To
them it was 'gross, in your face' and they are astonished and chagrined to find that it simply
wasn't real. (Classic cases involve experiments on 'eyewitness testimony') People very much
tend to underestimate how bad their judgement and perceptual accuracy can be!

Given this body of evidence, it's quite wrong to proceed as if the two alternatives
are *equally* likely. It's long past time for audiophiles to realize just how easily
they can be fooled -- and to keep that always in mind as a cautionary filter.

> So are we to believe that ABX testing does not impose its own "mental space"
> on the listeners? That it provides a neutral test bed in which the listeners
> will hear what is actually there? I want proof that it does.

ABX is simply a variant on double-blind methodology. Your argument boils down
to nothing more than 'participating in a test affects the results of the test'.

> Most listeners have experienced something like this -- what sounds neutral
> at first gradually begins revealing its colorations.

Many experience the opposite -- a sense of 'difference' at first that disappears as
the user psychologically accomodates to the new factor in the environment.

> I contend that it's
> possible for a piece of electronics to have subtle colorations that don't
> show up in initial listening -- and are not audible in even a direct
> comparison with a neutral bypass -- until the listener has learned to hear
> them. In other words, ABX testing won't work with small errors that are not
> immediately noticeable, but need time to gradually reveal themselves.

Then do the ABX test *after* you're convinced that you hear them. That is all anyone has ever
asked of the many, many claims of audiophiles (it's in line with using ABX for research
purposes, when one would first train subjects to hear differences). . It's what I asked John
Atkinson to do when he trotted out his 'Saul to Paul' conversion story (from objectivist to
subjectivist) at an HE meeting a few years back -- he did an ABX, found he couldn't tell a
beloved tube amp from a SS amp, so went with the SS amp, regretted it after months of
llistening, went back to the tube amp. This, to him, was apparently evidence that ABX is
useless. You see the elementary logical error here, I trust.

> This is speculation, of course, but no worse than assuming ABX testing is
> correct and complete.

The idea that someone assumes that, in such a case, is a straw man.

> I really want to know "the truth" -- and I don't feel either side in this
> debate is delivering it. Or knows what it's talking about, for that matter.

The side that is unaware or 'in denial' of utterly uncontroversial facts of human
psychology -- facts which underly virtually all scientific protocols today --
is definitely at a rhetorical disadvantage.

> failure of ABX testing to confirm subjective reactions necessarily _proves_
> the latter to be incorrect -- particularly if there were some agreement
> among the subjective reviewers. Rather he might react -- "What the hell is
> going on here?", as I did. I want to know what's going on. And I'm not
> stupid enough to think that an uncorrelated "scientifc" test is going to
> supply all the answers any more than uncontrolled anecdotal listening.

The thing is, *there is an answer already*. Science has fundamentally acknowledged the
existence of psychological 'noise making', by its use of controls and blind
methods. It uses those to help rule out 'mental noise' as the answer. And no,
it doesn't really matter much that 'subjective reviewers agree', any more
than ten top scientists performing the same uncontrolled experiment will arrive at
a 'more believable' answer.

> I want long-term double-blind "casual" listening tests -- running two or
> three years -- with a least a dozen listeners and as many different
> amplifiers. I want to see if, at the end of that time, there's any consensus
> as to what the amplifiers "sound like", and whether such a consensus carries
> over to ABX testing. Once such a test is done, we'll begin to _understand_
> the strengths and weaknesses of both types of testing and how they correlate
> (or not) with each other.

All I want is for audiophiles -- led by their press -- to acknowledge the surprising strength
of psychological bias -- the way science does. And to tailor its claims accordingly. The
consumer can only win from such practices.

But in the meantime, do you hold 'difference' and 'no difference' to be
*equally* likely? If so, why?

> If anyone would like to discuss setting up such a test, I'd be happy to
> discuss it with them. But I'm not holding my breath.

Who should really be setting up these tests? The people who acknowledge the fallibility of
human perception, and make their claims accordingly, or the ones who are sure they hear
differences based on sighted listening?

> > > What Arny is actually saying here (and you can look at
> > > his other postings to confirm this) is... If a device has
> > > distortion and noise characteristics below a certain
> > > measurable level, then its behavior is indistinguishable
> > > from any other device measuring similarly. And ABX
> > > testing "proves" this.

> > So does every other reliable listening test.

> In your book, a reliable listening test is one whose conclusions fit your
> preconceptions.

Whetehr that's bad or not entirely depends on the nature of those
'preconceptions'.

___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

Steven Sullivan

April 17th 07, 06:21 PM

William Sommerwerck > wrote:
> >>> Other than in the minds of a few obsessive audiophiles, the connection
> >>> between ABX and sitting down and listening are obvious to all who have
> >>> investigated it.

> >> The validity or non- or this argument is the crux of the bisquit.

> > Agreed. And the source of the problem is the fear and ignorance
> > of the audiophiles. For about 30 years many audiophiles have been
> > sitting around making up self-pitying stories about how horrible and
> > stupid reliable subjective testing is.

> > There are very few people with significant experience actually doing
> > reliable subjective tests that don't respect the process.

> > Having a reliable measure and applying it, is hugely effective in
> fostering
> > progress within a technology.

> But you've never demonstrated the "reliability" of ABX testing -- outside of
> its own little narrow universe.

> The issue remains, and won't go away until someone actually performs valid
> experiments.

One might wonder if a 'valid experiment' is merely one that fits your preconceptions.

;>

___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

Steven Sullivan

April 17th 07, 06:27 PM

William Sommerwerck > wrote:
> >> What I meant was that an ABX test is inherently formal, and
> >> is not equivalent to casual listening. The fact that the listeners
> >> are expected to make a decision "formalizes" it.

> > And that differs from listening to support a purchase decision, how?

> It doesn't, and that's a valid point. (I was going to bring it up, but
> decided not to.) Comparing two or more products, deciding which to buy,
> introduces the same sort of "make a decision" psychological effects that
> occur in ABX testing. The kind of casual-but-controlled testing I'd like to
> perform minimizes these.

Isn't this getting rather rarefied?

Your 'valid' experiment would seem to involve mental jujitsu whereby the listener isn't aware
that he's (it will almost certainly be 'he', let's face it) judging the sound he's currently
hearing in comparison to another sound, yet spontaneously comes to believe (one might say
'decides') that one of them sounds better than another. How does one 'judge' sound quality
without reference to other sound (even if it's 'the absolute sound of acoustic instruments in
real space")?

I'm unclear on your method for avoiding 'deciision'. How, exactly, would you long term, casual
double-blind test be carried out?

___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

Harry Lavo

April 17th 07, 06:43 PM

"Steven Sullivan" > wrote in message
...
> William Sommerwerck > wrote:
>> >> What I meant was that an ABX test is inherently formal, and
>> >> is not equivalent to casual listening. The fact that the listeners
>> >> are expected to make a decision "formalizes" it.
>
>> > And that differs from listening to support a purchase decision, how?
>
>> It doesn't, and that's a valid point. (I was going to bring it up, but
>> decided not to.) Comparing two or more products, deciding which to buy,
>> introduces the same sort of "make a decision" psychological effects that
>> occur in ABX testing. The kind of casual-but-controlled testing I'd like
>> to
>> perform minimizes these.
>
>
> Isn't this getting rather rarefied?
>
> Your 'valid' experiment would seem to involve mental jujitsu whereby the
> listener isn't aware
> that he's (it will almost certainly be 'he', let's face it) judging the
> sound he's currently
> hearing in comparison to another sound, yet spontaneously comes to believe
> (one might say
> 'decides') that one of them sounds better than another. How does one
> 'judge' sound quality
> without reference to other sound (even if it's 'the absolute sound of
> acoustic instruments in
> real space")?
>
> I'm unclear on your method for avoiding 'deciision'. How, exactly, would
> you long term, casual
> double-blind test be carried out?
..
That's the whole point of considering monadic testing as a better
alternative, Steven. All the listener has to do is listen to one variable
and after having enough exposure to feel confident in his ability to "rate",
then "rating" his listening experience on a series of semantic scales.
That's it. No comparison to anything other than his own subjective reaction
to the music, and perhaps to live music. When you get 100-150 people
(minimum) doing this for each of the variables (minimum two) then you can
apply statistical difference measurements to determine what if anything is
different, at what level of statistical significance.

IMO this type of test represents the "gold standard" of audio testing. It
has been done..it is not anything new. But it most closely approximates a
normal listening state (especially if done at home) and requires not
artificial comparison, or even any particular noting "during" listening.
All it requires is the willingness to apply ratings judgement after the
experience.

Harry Lavo

April 17th 07, 06:49 PM

"Steven Sullivan" > wrote in message
...
> William Sommerwerck > wrote:
>> > ABX is first and foremost composed of sitting down and listening.
>> > I know of no way to do an ABX test without sitting down and
>> > listening, do you William?
>
>> I know lots of ways of sitting down and listening that are not ABX
>> testing.
>
> Yes, and which of those circumvent the problem of bias?
>
> You listen to something on CD then on SACD, by whatever sighted method,
> long term, short term,
> whatever. You decide the SACD, which has been marketed as a sonic
> improvement on
> CD, sounds better. Then you conclude: it sounds better *because it's an
> SACD*.
>
> If my years of reading audio forums is any indication, this is a typical
> 'audiophile' result
> of 'listening that is not ABX testing'. Do you still not see how
> essentially dubious it is?
>
>
>> One of them was the proximate cause of my quitting. I was auditioning
>> cables -- just for fun. I knew which were connected, etc. I was in a good
>> mood -- I'd learned that the best time to listen was early on a weekend
>> morning, when it was quiet, your ears were rested, etc. I heard very
>> noticeable differences among the cables, and found that I preferred (note
>> the word) the $50 cable from one company over a $350 cable from another
>> company. The best of all was the $400 cable from the first company, which
>> sounded qualitatively like their $50 cable, but "eversomuchmoreso". (Yes,
>> I
>> read Beverly Cleary when I was a kid.)
>
>> The "better" (to me) cables had more depth; the others were "flatter" and
>> more "forward"-sounding. Out of curiosity (and, in my book, like a good
>> scientist), I used one cable from each pair, and got an ear-twisting
>> result
>> that sounded as if I were hearing the recording from two different
>> perspectives at the same time. This was strong evidence (not proof) that
>> I
>> was hearing _real_ differences, as it seemed unlikely my mind could cook
>> up
>> such a sonic mess and present it convincingly.
>
>> I should add the all the perceived differences among the cables were
>> gross,
>> in-your-face effects, and not the least-bit subtle.
>
>> So interconnects really do sound different, right? They must, as I
>> audtioned
>> them under conditions I knew would maximize my ability to hear
>> differences.
>
>> Not so fast.
>
>> The next morning I packed up the cables to ship back, but decided to
>> first
>> give them another listen. This time, I was prejudiced to hear what I'd
>> heard
>> the day before. But there were _no_ differences among the cables. They
>> all
>> sounded the same. I was in a similar mood, and played the same
>> recordings,
>> but got grossly different (subjective) results.
>
>> So, which audition revealed "the truth"? Don't say "Neither, because
>> there
>> was no control." One of them _had_ to be much closer to the truth than
>> the
>> other, because both of them couldn't have been correct. I couldn't tell
>> which, because I was caught up (as we all are) in my own mental space,
>> and
>> had no way to "look outside" to determine the state of my mind or
>> perceptions.
>
> Study of human psychology is rife with instances of 'inevitable illusions'
> cast up by the
> mind, that lead people to INSIST that they saw/heard something that simply
> wasn't real. To
> them it was 'gross, in your face' and they are astonished and chagrined to
> find that it simply
> wasn't real. (Classic cases involve experiments on 'eyewitness
> testimony') People very much
> tend to underestimate how bad their judgement and perceptual accuracy can
> be!
>
> Given this body of evidence, it's quite wrong to proceed as if the two
> alternatives
> are *equally* likely. It's long past time for audiophiles to realize just
> how easily
> they can be fooled -- and to keep that always in mind as a cautionary
> filter.
>
>> So are we to believe that ABX testing does not impose its own "mental
>> space"
>> on the listeners? That it provides a neutral test bed in which the
>> listeners
>> will hear what is actually there? I want proof that it does.
>
> ABX is simply a variant on double-blind methodology. Your argument boils
> down
> to nothing more than 'participating in a test affects the results of the
> test'.
>
>> Most listeners have experienced something like this -- what sounds
>> neutral
>> at first gradually begins revealing its colorations.
>
> Many experience the opposite -- a sense of 'difference' at first that
> disappears as
> the user psychologically accomodates to the new factor in the environment.
>
>
>> I contend that it's
>> possible for a piece of electronics to have subtle colorations that don't
>> show up in initial listening -- and are not audible in even a direct
>> comparison with a neutral bypass -- until the listener has learned to
>> hear
>> them. In other words, ABX testing won't work with small errors that are
>> not
>> immediately noticeable, but need time to gradually reveal themselves.
>
> Then do the ABX test *after* you're convinced that you hear them. That is
> all anyone has ever
> asked of the many, many claims of audiophiles (it's in line with using ABX
> for research
> purposes, when one would first train subjects to hear differences). .
> It's what I asked John
> Atkinson to do when he trotted out his 'Saul to Paul' conversion story
> (from objectivist to
> subjectivist) at an HE meeting a few years back -- he did an ABX, found he
> couldn't tell a
> beloved tube amp from a SS amp, so went with the SS amp, regretted it
> after months of
> llistening, went back to the tube amp. This, to him, was apparently
> evidence that ABX is
> useless. You see the elementary logical error here, I trust.
>
>
>> This is speculation, of course, but no worse than assuming ABX testing is
>> correct and complete.
>
> The idea that someone assumes that, in such a case, is a straw man.
>
>> I really want to know "the truth" -- and I don't feel either side in this
>> debate is delivering it. Or knows what it's talking about, for that
>> matter.
>
> The side that is unaware or 'in denial' of utterly uncontroversial facts
> of human
> psychology -- facts which underly virtually all scientific protocols
> today --
> is definitely at a rhetorical disadvantage.
>
>
>> failure of ABX testing to confirm subjective reactions necessarily
>> _proves_
>> the latter to be incorrect -- particularly if there were some agreement
>> among the subjective reviewers. Rather he might react -- "What the hell
>> is
>> going on here?", as I did. I want to know what's going on. And I'm not
>> stupid enough to think that an uncorrelated "scientifc" test is going to
>> supply all the answers any more than uncontrolled anecdotal listening.
>
> The thing is, *there is an answer already*. Science has fundamentally
> acknowledged the
> existence of psychological 'noise making', by its use of controls and
> blind
> methods. It uses those to help rule out 'mental noise' as the answer. And
> no,
> it doesn't really matter much that 'subjective reviewers agree', any more
> than ten top scientists performing the same uncontrolled experiment will
> arrive at
> a 'more believable' answer.
>
>> I want long-term double-blind "casual" listening tests -- running two or
>> three years -- with a least a dozen listeners and as many different
>> amplifiers. I want to see if, at the end of that time, there's any
>> consensus
>> as to what the amplifiers "sound like", and whether such a consensus
>> carries
>> over to ABX testing. Once such a test is done, we'll begin to
>> _understand_
>> the strengths and weaknesses of both types of testing and how they
>> correlate
>> (or not) with each other.
>
> All I want is for audiophiles -- led by their press -- to acknowledge the
> surprising strength
> of psychological bias -- the way science does. And to tailor its claims
> accordingly. The
> consumer can only win from such practices.
>
> But in the meantime, do you hold 'difference' and 'no difference' to be
> *equally* likely? If so, why?
>
>
>> If anyone would like to discuss setting up such a test, I'd be happy to
>> discuss it with them. But I'm not holding my breath.
>
> Who should really be setting up these tests? The people who acknowledge
> the fallibility of
> human perception, and make their claims accordingly, or the ones who are
> sure they hear
> differences based on sighted listening?
>
>
>> > > What Arny is actually saying here (and you can look at
>> > > his other postings to confirm this) is... If a device has
>> > > distortion and noise characteristics below a certain
>> > > measurable level, then its behavior is indistinguishable
>> > > from any other device measuring similarly. And ABX
>> > > testing "proves" this.
>
>> > So does every other reliable listening test.
>
>> In your book, a reliable listening test is one whose conclusions fit your
>> preconceptions.
>
> Whetehr that's bad or not entirely depends on the nature of those
> 'preconceptions'.
..

A good scientist says that if all the common type "tests" are prone to
preconceptions (positive or negative) then it is incumbent to come up with a
test approach that eliminates the preconceptions. Monadic, scaler testing
without the listeners being aware of *what* they are testing (and thus
avoiding preconceptions) is the best approach I can think of for audio.
That's why several years ago on RAHE I proposed it as one of two
"validation" tests to see if ABX results are consisten. ABX has the
potential to eliminate "positive" bias but it has no controls against
"negative" bias, and it is a highly intrusive test (in the normal
psychological sense) that is highly likely to alter the ear/brain processing
of music.

Harry Lavo

April 17th 07, 06:52 PM

"Steven Sullivan" > wrote in message
...
> William Sommerwerck > wrote:
>> >>> Other than in the minds of a few obsessive audiophiles, the
>> >>> connection
>> >>> between ABX and sitting down and listening are obvious to all who
>> >>> have
>> >>> investigated it.
>
>> >> The validity or non- or this argument is the crux of the bisquit.
>
>> > Agreed. And the source of the problem is the fear and ignorance
>> > of the audiophiles. For about 30 years many audiophiles have been
>> > sitting around making up self-pitying stories about how horrible and
>> > stupid reliable subjective testing is.
>
>> > There are very few people with significant experience actually doing
>> > reliable subjective tests that don't respect the process.
>
>> > Having a reliable measure and applying it, is hugely effective in
>> fostering
>> > progress within a technology.
>
>> But you've never demonstrated the "reliability" of ABX testing -- outside
>> of
>> its own little narrow universe.
>
>> The issue remains, and won't go away until someone actually performs
>> valid
>> experiments.
>
>
> One might wonder if a 'valid experiment' is merely one that fits your
> preconceptions.
..
No, he's raising a valid issue. See my other posts commenting on how
monadic testing could serve to get around the issues he raises.
That is, for really scientific testing. It is not a "quick fix", single
person test. But neither should we be using such a test if it has problems
for its intended purpose, which is what many of us feel...especially those
of us schooled in the social sciences.

Arny Krueger

April 17th 07, 07:15 PM

"William Sommerwerck" > wrote in message
. ..
>> Well William, if you haven't figured it out yet, let me be clearer.
>> I'm not here to be your servant, and I'm not here to create the
>> perfect listening test.
>
> !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
>
> What do you think your claims for ABX testing are?

Imperfect but often useful for a certain range of tests.

ABC/hr is often a far superior testing methodology, as are other
less-well-known methods.

>> I'm here to get some jobs done. I don't *care* about some undefined
>> ideal test, I want a test that will get me through my next technical
>> decision based on sound quality.

> But you claim that your form of testing provides a definitive evaluation
> of
> sound quality -- THEN you say "nothing is ever defninitive".

Where did I say that ABX is definitive? Quote me or admit that I didn't say
it.

> There is a very fine -- and I claim meaningless -- line between
> "definitive"
> and "perfect".

Doesn't matter to me, since I believe that reality reveals itself to us in a
progressive fashion. ABX's provisional results can still be useful enough
of the time so that it can be useful and worthwhile.

> What an intellectual weasel you are.

Spare us all. ;-)

>> William, I don't need you to question me.

> I'm exaxtly the person you need to question. I'm not your friend, and I
> don't agree with you. That's the sort of person you need to keep you
> honest.

Honesty isn't really an issue here for my side of the discusison. What we is
generally recognized as being useful in many cases.

> You are obviously very bothered that someone of about your age, IQ, and
> education refuses to ingest your codswallop.

Not at all. I do enjoy disagreement to a degree. Agreement has its moments
as well.

> Who refuses to blindly accept
> what all the sycophants who hang on your ever word believe.

I've I've got any sycophants hanging on, please help me scare them away.

>> Non sequitor. First off, I just said that the results of my tests are
>> provisional, not definitive. In common speech, they are what they are -
>> take them or leave them. That the mainstream audio world has moved
>> on from primitive bias-ridden listening tests is a well-known fact.

> You want people to believe that your "provisional" results are definitive.

No.

> Then when someone points out how incomplete they are, you won't listen.

I've heard it before, many times. The failings of those kinds of argument
are well-known.

> You want credit for research that you cannot demonstrate the validity of.

I could care less about credit of that kind. Cash, checks and Paypal would
be appreciated, but are not expected. ;-) ABX has pretty well been a
financial loss for most of us. Clark took it to the bank better than I did,
but he put tons more work into it. Added reward for added value - I get
that.

>>> I wanted to be friends with him, but gradually discovered he is
>>> incapable of having an intelligent conversation with anyone,
>>> about anything. He knows what he knows, and that's it.
>
>> Which makes him different from you how, William? ;-)

> What's with these continuing "wink signs"?

I'm sitting here smiling. One has to take this sort of thing with humor if
one is to take it at all.

> Well, for one thing, I have lots of friends I have intelligent
> conversations
> with, that I don't always agree with. If I were to tell (names omitted) I
> thought they were wrong about something, they would take me seriously and
> thoughfully consider what I have to say.

This is an old, old show and I see no added value this time through.

>> You see, I don't know exactly what I know, and I certainly don't
>> know exactly what I don't know. But I know that I don't know it all,
>> and I know that just about everything I've ever done or will do will
>> or has been improved on.

> So why is it that when someone points out your intellectual errors, you
> don't seriously consider them? That's wise, that's smart, that's mature?

What errors?

> My number is 425-235-9579. You may call me at any time (including the
> middle
> of the night), and we'll discuss whatever you care to discuss. But you
> won't, because you're afraid that, as we're talking, the light bulb will
> suddenly go on, and the intellectual sandpile you've built your research
> on
> will slide away.

Ain't gonna happen. Atkinson made me a far more attractive offer, and I just
barely took it. I pretty much turned it down at least once before I took it.

>> Who said anything about arbitrarily? But, I probably know more than
>> the average bear about attracting, engaging, and transforming people.
>> BTW, that implies directing their state of mind.

> To what ends? Please be specific.

Well, you know I work with Christian churches.

>> Did I drop a name?

> You did, more or less. Rather less than more, but you appealed to the
> belief
> that you have the respect of people the rest of us are supposed to look up
> to.

Namely who?

Arny Krueger

April 17th 07, 07:17 PM

Jim Gilliland

April 17th 07, 08:17 PM

Harry Lavo wrote:
> .
> That's the whole point of considering monadic testing as a better
> alternative, Steven. All the listener has to do is listen to one variable
> and after having enough exposure to feel confident in his ability to "rate",
> then "rating" his listening experience on a series of semantic scales.
> That's it. No comparison to anything other than his own subjective reaction
> to the music, and perhaps to live music. When you get 100-150 people
> (minimum) doing this for each of the variables (minimum two) then you can
> apply statistical difference measurements to determine what if anything is
> different, at what level of statistical significance.
>
> IMO this type of test represents the "gold standard" of audio testing.

Since each of those hundreds of people would apply their own definition and
standard for said "variable", as well as the thresholds for values on your
"semantic scales", the results of such a test would be absolutely meaningless.

[email protected]

April 17th 07, 09:23 PM

Jim Gilliland > wrote:
> Harry Lavo wrote:
> > .
> > That's the whole point of considering monadic testing as a better
> > alternative, Steven. All the listener has to do is listen to one variable
> > and after having enough exposure to feel confident in his ability to "rate",
> > then "rating" his listening experience on a series of semantic scales.
> > That's it. No comparison to anything other than his own subjective reaction
> > to the music, and perhaps to live music. When you get 100-150 people
> > (minimum) doing this for each of the variables (minimum two) then you can
> > apply statistical difference measurements to determine what if anything is
> > different, at what level of statistical significance.
> >
> > IMO this type of test represents the "gold standard" of audio testing.

> Since each of those hundreds of people would apply their own definition and
> standard for said "variable", as well as the thresholds for values on your
> "semantic scales", the results of such a test would be absolutely meaningless.

Agreed. It's marketing bat**** "science."

Harry Lavo

April 17th 07, 10:26 PM

"Jim Gilliland" > wrote in message
...
> Harry Lavo wrote:
>> .
>> That's the whole point of considering monadic testing as a better
>> alternative, Steven. All the listener has to do is listen to one
>> variable and after having enough exposure to feel confident in his
>> ability to "rate", then "rating" his listening experience on a series of
>> semantic scales. That's it. No comparison to anything other than his own
>> subjective reaction to the music, and perhaps to live music. When you
>> get 100-150 people (minimum) doing this for each of the variables
>> (minimum two) then you can apply statistical difference measurements to
>> determine what if anything is different, at what level of statistical
>> significance.
>>
>> IMO this type of test represents the "gold standard" of audio testing.
>
> Since each of those hundreds of people would apply their own definition
> and standard for said "variable", as well as the thresholds for values on
> your "semantic scales", the results of such a test would be absolutely
> meaningless.

Sorry...the groups are randomly matched and the law of averages which is
inherent in statistical evaluation comes into play. It is a tried and true
appoach in the social sciences for many purposes.

And as for "each having his own reference" that is exactly what you want if
you are trying to evaluation audio equipment for musical reproduction in the
home. You don't care if you have a "composite"...you just want to be sure
you have two panels of evaluators who "statistically match".

Harry Lavo

April 17th 07, 10:28 PM

> wrote in message
...
> Jim Gilliland > wrote:
>> Harry Lavo wrote:
>> > .
>> > That's the whole point of considering monadic testing as a better
>> > alternative, Steven. All the listener has to do is listen to one
>> > variable
>> > and after having enough exposure to feel confident in his ability to
>> > "rate",
>> > then "rating" his listening experience on a series of semantic scales.
>> > That's it. No comparison to anything other than his own subjective
>> > reaction
>> > to the music, and perhaps to live music. When you get 100-150 people
>> > (minimum) doing this for each of the variables (minimum two) then you
>> > can
>> > apply statistical difference measurements to determine what if anything
>> > is
>> > different, at what level of statistical significance.
>> >
>> > IMO this type of test represents the "gold standard" of audio testing.
>
>> Since each of those hundreds of people would apply their own definition
>> and
>> standard for said "variable", as well as the thresholds for values on
>> your
>> "semantic scales", the results of such a test would be absolutely
>> meaningless.
>
> Agreed. It's marketing bat**** "science."

Sorry, John, it isn't. It is a pretty standard "social science" type of
testing. Do you consider all the social sciences "marketing"? Do you
consider them all "bat****"?

Jim Gilliland

April 17th 07, 11:34 PM

Harry Lavo wrote:

> Sorry...the groups are randomly matched and the law of averages which is
> inherent in statistical evaluation comes into play. It is a tried and true
> appoach in the social sciences for many purposes.

I understand perfectly well the concept of population sampling in the social
sciences. If you feel that a longer term and more relaxed test gives a more
meaningful answer, that's fine. As long as it is conducted under double blind
conditions, the results should have some meaning.

It's a whole lot more time and work than an ABX test, though, and I see no
additional value in it. In fact, in general I'd expect it to be less accurate
since the test subjects aren't necessarily doing any careful critical listening,
and since their experiences are being captured after the fact when they are no
longer "fresh". But despite those flaws, I still think you could achieve some
useful results. So if you prefer that approach, go for it. Just don't expect
everyone to follow.

There are lots of ways to put together a meaningful test. The common ground,
though, must be that the testers and test subjects are kept completely unaware
of the specifics of the systems under test. You can't "average away" the bias
that comes from a sighted test.

[email protected]

April 17th 07, 11:40 PM

Harry Lavo > wrote:

> > wrote in message
> ...
> > Jim Gilliland > wrote:
> >> Harry Lavo wrote:
> >> > .
> >> > That's the whole point of considering monadic testing as a better
> >> > alternative, Steven. All the listener has to do is listen to one
> >> > variable
> >> > and after having enough exposure to feel confident in his ability to
> >> > "rate",
> >> > then "rating" his listening experience on a series of semantic scales.
> >> > That's it. No comparison to anything other than his own subjective
> >> > reaction
> >> > to the music, and perhaps to live music. When you get 100-150 people
> >> > (minimum) doing this for each of the variables (minimum two) then you
> >> > can
> >> > apply statistical difference measurements to determine what if anything
> >> > is
> >> > different, at what level of statistical significance.
> >> >
> >> > IMO this type of test represents the "gold standard" of audio testing.
> >
> >> Since each of those hundreds of people would apply their own definition
> >> and
> >> standard for said "variable", as well as the thresholds for values on
> >> your
> >> "semantic scales", the results of such a test would be absolutely
> >> meaningless.
> >
> > Agreed. It's marketing bat**** "science."

> Sorry, John, it isn't. It is a pretty standard "social science" type of
> testing. Do you consider all the social sciences "marketing"? Do you
> consider them all "bat****"?

I certainly consider marketing as bat****.

Harry Lavo

April 18th 07, 02:29 AM

> wrote in message
...
> Harry Lavo > wrote:
>
>> > wrote in message
>> ...
>> > Jim Gilliland > wrote:
>> >> Harry Lavo wrote:
>> >> > .
>> >> > That's the whole point of considering monadic testing as a better
>> >> > alternative, Steven. All the listener has to do is listen to one
>> >> > variable
>> >> > and after having enough exposure to feel confident in his ability to
>> >> > "rate",
>> >> > then "rating" his listening experience on a series of semantic
>> >> > scales.
>> >> > That's it. No comparison to anything other than his own subjective
>> >> > reaction
>> >> > to the music, and perhaps to live music. When you get 100-150
>> >> > people
>> >> > (minimum) doing this for each of the variables (minimum two) then
>> >> > you
>> >> > can
>> >> > apply statistical difference measurements to determine what if
>> >> > anything
>> >> > is
>> >> > different, at what level of statistical significance.
>> >> >
>> >> > IMO this type of test represents the "gold standard" of audio
>> >> > testing.
>> >
>> >> Since each of those hundreds of people would apply their own
>> >> definition
>> >> and
>> >> standard for said "variable", as well as the thresholds for values on
>> >> your
>> >> "semantic scales", the results of such a test would be absolutely
>> >> meaningless.
>> >
>> > Agreed. It's marketing bat**** "science."
>
>> Sorry, John, it isn't. It is a pretty standard "social science" type of
>> testing. Do you consider all the social sciences "marketing"? Do you
>> consider them all "bat****"?
>
> I certainly consider marketing as bat****.

Well, in a perfect free-enterprise world we wouldn't need "marketing"
because we'd hae perfect markets (that has a very specific connotation to
economists....it means a marketplace where every buyer knows everything
available and knows everything he/she needs to know about the product). If
you want to know what an imperfect market looks like, try to select a
primary care physician. Or the best personal injury lawyer.

Most products in this world function in an imperfect market, that is not all
possibilities are known to or available to the buyer, and they buyer doesn't
have perfect information to make the purchase. That's where marketing
enters in....it trys to draw perspective purchasers to an available product,
and to offer persuasive reasons for that buyer to purchase that particular
product. It can be done highly ethically....and it can be done in a
sleazeball fashion. Frankly the standards used to be higher...back in the
sixties when I got out of business school, if I want'ed to market a "new,
improved" Jell-O pudding, I had to show the company lawyers the test results
providing the statistical "proof". Then I (the ad agency, actually) had to
do the same with the lawyers for all the networks we wanted to use. I even
had to run the promotional materials we used with our own sales force past
our internal lawyers. Frankly, if those standards held today Microsoft and
half the remainder of the computer industry would be out of business, and
John Kerry would be President. You can draw your own conclusions about the
audio industry. But I can guarantee you that Audio Research, McIntosh, and
Thiel Audio would still be in business.

Chris Hornbeck

April 18th 07, 03:12 AM

On Tue, 17 Apr 2007 07:48:07 -0400, "Arny Krueger" >
wrote:

>>>Name something that is generally agreed-upon to adversly affect sonic
>>>performance that doesn't adversly affect dynamic range and bandwidth.
>
>> Monotonicity. Crossover distortion. Small signal threshold
>> effects. Quantization. Off the top of my head.
>
>They all adversely affect dynamic range by creating spurious responses. This
>effectively raises the noise floor.

Not really universally true. I (and you) can give specific
counter-examples in each case.

>Dynamic Range is *not* the same thing as SNR.

Of course. And, dynamic range can be defined pretty much the
way anyone wishes, for any specific purpose. And so can SNR.
They're just boiled-down numbers, and the boiling point
is adjustable.

My point is not to argue against classical technical
measurement simplifications per se; rather to contest
the equivalence of the simplifications to the total
measurement.

We live in a world of unimaginable complexity; we survive
and prosper by simplifying it. This is called technology.

But the unimaginable world still lurks around us. Attempts
to tackle that unimaginable world are called science. It
ain't pretty, and it ain't convenient.

Much thanks, as always,

Chris Hornbeck

"Physics is like sex: sure, it may give some practical results, but
that's not why we do it," - Richard Feynman.

Arny Krueger

April 18th 07, 03:52 PM

"Chris Hornbeck" > wrote in message
...
> On Tue, 17 Apr 2007 07:48:07 -0400, "Arny Krueger" >
> wrote:
>
>>>>Name something that is generally agreed-upon to adversly affect sonic
>>>>performance that doesn't adversly affect dynamic range and bandwidth.
>>
>>> Monotonicity. Crossover distortion. Small signal threshold
>>> effects. Quantization. Off the top of my head.

>>They all adversely affect dynamic range by creating spurious responses.
>>This
>>effectively raises the noise floor.

> Not really universally true. I (and you) can give specific
> counter-examples in each case.

Then do it. I'm willing to be educated.

Steven Sullivan

April 18th 07, 11:31 PM

Harry Lavo > wrote:

> "Jim Gilliland" > wrote in message
> ...
> > Harry Lavo wrote:
> >> .
> >> That's the whole point of considering monadic testing as a better
> >> alternative, Steven. All the listener has to do is listen to one
> >> variable and after having enough exposure to feel confident in his
> >> ability to "rate", then "rating" his listening experience on a series of
> >> semantic scales. That's it. No comparison to anything other than his own
> >> subjective reaction to the music, and perhaps to live music. When you
> >> get 100-150 people (minimum) doing this for each of the variables
> >> (minimum two) then you can apply statistical difference measurements to
> >> determine what if anything is different, at what level of statistical
> >> significance.
> >>
> >> IMO this type of test represents the "gold standard" of audio testing.
> >
> > Since each of those hundreds of people would apply their own definition
> > and standard for said "variable", as well as the thresholds for values on
> > your "semantic scales", the results of such a test would be absolutely
> > meaningless.

> Sorry...the groups are randomly matched and the law of averages which is
> inherent in statistical evaluation comes into play. It is a tried and true
> appoach in the social sciences for many purposes.

Has it been used to test perception of difference?

___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

Steven Sullivan

April 19th 07, 12:00 AM

Harry Lavo > wrote:

> "Mike Rivers" > wrote in message
> ups.com...
> > On Apr 14, 7:37 am, "Natalie Drest" >
> > wrote:
> >

> >snip<

> >
> > I did learn something interesting about wide bandwidth recording from
> > the people who make a monstrous sample library with which they
> > simulate large pipe organs with multi-channel installations. They
> > record each pipe individually rather than recording the blended sound
> > of the organ from a distance, then they play back those pipes and let
> > them combine naturally in air. What they claim (and can prove by
> > looking at the frequency content of their recordings) is that there's
> > ultrasonic energy coming off a pipe, and when two pipes are playing
> > together, the beating of those ultrasonic frequencies is down in the
> > audible range. What you hear sounds different with and without that
> > ultrasonic energy.
> >

> >snip<

> Interesting. Mike, my hunch is that this is what Oohashi's subjects
> experienced with the recorded gamelan music they used in his ultrasonic
> test. The frequency overtones of the gongs, xylophones, and drums extended
> to very high frequencies, and so have the potential to create lots of
> "beats". As you recall, in that test of both hearing and physiology
> subjects recorded (at a statistically significant level) that the music
> sounded more pleasant (and their pleasure centers were activated) when the
> ultrasonics were included but not when they were excluded. And yet the
> ultrasonics reproduced separately from the audible music were not audible.

So let's see...Mike's talking about an effect 'down in the audible range'. Thus if we set up a
mike in the room where those pipes are playing, and faithfully capture only what's in the
audible range (20-20kHz give or take a few) ...aren't we recording the effect? Why record the
'cause' as well?

___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason

Chris Hornbeck

April 19th 07, 01:27 AM

On Wed, 18 Apr 2007 10:52:25 -0400, "Arny Krueger" >
wrote:

>>>>>Name something that is generally agreed-upon to adversly affect sonic
>>>>>performance that doesn't adversly affect dynamic range and bandwidth.
>>>
>>>> Monotonicity. Crossover distortion. Small signal threshold
>>>> effects. Quantization. Off the top of my head.
>
>>>They all adversely affect dynamic range by creating spurious responses.
>>>This
>>>effectively raises the noise floor.
>
>> Not really universally true. I (and you) can give specific
>> counter-examples in each case.
>
>Then do it. I'm willing to be educated.

Monotonicity: make some bits too big, some bits too little
(the usual case, more-or-less)

Crossover distortion: can actually increase measured dynamic
range, depending on definition

Small signal threshold effects: chemical capacitors and
B/H threshold flat spots in iron-cored inductors

Quantization: has no effect on classically measured dynamic
range

But, actually, this thread is getting terminably boring.
See ya in another.

Chris Hornbeck
"Ghidorah is a space monster; the monsters from Earth cannot win.
I will get in touch with you when you feel like giving up."
- Kiraaku Queen, "Destroy All Monsters"

Chris Hornbeck

April 19th 07, 01:27 AM

On Wed, 18 Apr 2007 10:33:31 +0200, Chel van Gennip
> wrote:

>> My point is not to argue against classical technical measurement
>> simplifications per se; rather to contest the equivalence of the
>> simplifications to the total measurement.

> At that
>point errors become not audible anymore, and the logic of discriminating
>between errors disappears.

An a priori assumption that's at the heart of the discussion.

Anyway, this is boring. See ya in another thread,

Chris Hornbeck
"Ghidorah is a space monster; the monsters from Earth cannot win.
I will get in touch with you when you feel like giving up."
- Kiraaku Queen, "Destroy All Monsters"

Arny Krueger

April 19th 07, 02:41 AM

"Steven Sullivan" > wrote in message
...
> Harry Lavo > wrote:
>
>> "Jim Gilliland" > wrote in message
>> ...
>> > Harry Lavo wrote:
>> >> .
>> >> That's the whole point of considering monadic testing as a better
>> >> alternative, Steven. All the listener has to do is listen to one
>> >> variable and after having enough exposure to feel confident in his
>> >> ability to "rate", then "rating" his listening experience on a series
>> >> of
>> >> semantic scales. That's it. No comparison to anything other than his
>> >> own
>> >> subjective reaction to the music, and perhaps to live music. When you
>> >> get 100-150 people (minimum) doing this for each of the variables
>> >> (minimum two) then you can apply statistical difference measurements
>> >> to
>> >> determine what if anything is different, at what level of statistical
>> >> significance.
>> >>
>> >> IMO this type of test represents the "gold standard" of audio testing.
>> >
>> > Since each of those hundreds of people would apply their own definition
>> > and standard for said "variable", as well as the thresholds for values
>> > on
>> > your "semantic scales", the results of such a test would be absolutely
>> > meaningless.
>
>> Sorry...the groups are randomly matched and the law of averages which is
>> inherent in statistical evaluation comes into play. It is a tried and
>> true
>> appoach in the social sciences for many purposes.
>
> Has it been used to test perception of difference?

One of the major features of monadic testing, the avoidance of close
comparisons, is well-known to produce insensitive results when equipment
closely approaches ideal sonics. That's a major reason why the AES only
recommends monadic testing for speakers, which generally sound quite a bit
different from the ideal.

Arny Krueger

April 19th 07, 02:44 AM

"Chris Hornbeck" > wrote in message
...
> On Wed, 18 Apr 2007 10:52:25 -0400, "Arny Krueger" >
> wrote:
>
>>>>>>Name something that is generally agreed-upon to adversly affect sonic
>>>>>>performance that doesn't adversly affect dynamic range and bandwidth.
>>>>
>>>>> Monotonicity. Crossover distortion. Small signal threshold
>>>>> effects. Quantization. Off the top of my head.
>>
>>>>They all adversely affect dynamic range by creating spurious responses.
>>>>This
>>>>effectively raises the noise floor.
>>
>>> Not really universally true. I (and you) can give specific
>>> counter-examples in each case.
>>
>>Then do it. I'm willing to be educated.

> Monotonicity: make some bits too big, some bits too little
> (the usual case, more-or-less)

Nope, that produces spurious responses when there is a signal present.

> Crossover distortion: can actually increase measured dynamic
> range, depending on definition

Nope, crossover distortion produces spurious responses when there is a
signal present.

> Small signal threshold effects: chemical capacitors and
> B/H threshold flat spots in iron-cored inductors

Nope, they produce spurious responses when there is a signal present.

> Quantization: has no effect on classically measured dynamic
> range

I don't know what a classic measurement of dynamic range is. But, I do know
what the current industry standard practice is. There is a signal present.

Harry Lavo

April 19th 07, 03:59 AM

"Steven Sullivan" > wrote in message
...
> Harry Lavo > wrote:
>
>> "Jim Gilliland" > wrote in message
>> ...
>> > Harry Lavo wrote:
>> >> .
>> >> That's the whole point of considering monadic testing as a better
>> >> alternative, Steven. All the listener has to do is listen to one
>> >> variable and after having enough exposure to feel confident in his
>> >> ability to "rate", then "rating" his listening experience on a series
>> >> of
>> >> semantic scales. That's it. No comparison to anything other than his
>> >> own
>> >> subjective reaction to the music, and perhaps to live music. When you
>> >> get 100-150 people (minimum) doing this for each of the variables
>> >> (minimum two) then you can apply statistical difference measurements
>> >> to
>> >> determine what if anything is different, at what level of statistical
>> >> significance.
>> >>
>> >> IMO this type of test represents the "gold standard" of audio testing.
>> >
>> > Since each of those hundreds of people would apply their own definition
>> > and standard for said "variable", as well as the thresholds for values
>> > on
>> > your "semantic scales", the results of such a test would be absolutely
>> > meaningless.
>
>> Sorry...the groups are randomly matched and the law of averages which is
>> inherent in statistical evaluation comes into play. It is a tried and
>> true
>> appoach in the social sciences for many purposes.
>
> Has it been used to test perception of difference?
..
It is used to record perceptions in a quantifiable way. Then statistical
analysis takes over to tell you whether or not their is a difference. That
is the beauty of it...you don't have to "compare" it to anything....all you
have to do is listen, listen, listen and then afterwards decided how you
felt about your experience/the sound as you fill out the semantic
differential scales. It is as close as you can come to normal listening in
a test environment.

Harry Lavo

April 19th 07, 04:01 AM

"Steven Sullivan" > wrote in message
...
> Harry Lavo > wrote:
>
>> "Mike Rivers" > wrote in message
>> ups.com...
>> > On Apr 14, 7:37 am, "Natalie Drest" >
>> > wrote:
>> >
>
>> >snip<
>
>> >
>> > I did learn something interesting about wide bandwidth recording from
>> > the people who make a monstrous sample library with which they
>> > simulate large pipe organs with multi-channel installations. They
>> > record each pipe individually rather than recording the blended sound
>> > of the organ from a distance, then they play back those pipes and let
>> > them combine naturally in air. What they claim (and can prove by
>> > looking at the frequency content of their recordings) is that there's
>> > ultrasonic energy coming off a pipe, and when two pipes are playing
>> > together, the beating of those ultrasonic frequencies is down in the
>> > audible range. What you hear sounds different with and without that
>> > ultrasonic energy.
>> >
>
>> >snip<
>
>> Interesting. Mike, my hunch is that this is what Oohashi's subjects
>> experienced with the recorded gamelan music they used in his ultrasonic
>> test. The frequency overtones of the gongs, xylophones, and drums
>> extended
>> to very high frequencies, and so have the potential to create lots of
>> "beats". As you recall, in that test of both hearing and physiology
>> subjects recorded (at a statistically significant level) that the music
>> sounded more pleasant (and their pleasure centers were activated) when
>> the
>> ultrasonics were included but not when they were excluded. And yet the
>> ultrasonics reproduced separately from the audible music were not
>> audible.
>
>
>
> So let's see...Mike's talking about an effect 'down in the audible range'.
> Thus if we set up a
> mike in the room where those pipes are playing, and faithfully capture
> only what's in the
> audible range (20-20kHz give or take a few) ...aren't we recording the
> effect? Why record the
> 'cause' as well?

Because the effect is created in your room, from the "cause". That's why
they recorded each pipe separately.

Arny Krueger

April 19th 07, 11:15 AM

"Harry Lavo" > wrote in message
. ..
>
> "Steven Sullivan" > wrote in message
> ...
>> Harry Lavo > wrote:
>>
>>> "Jim Gilliland" > wrote in message
>>> ...
>>> > Harry Lavo wrote:
>>> >> .
>>> >> That's the whole point of considering monadic testing as a better
>>> >> alternative, Steven. All the listener has to do is listen to one
>>> >> variable and after having enough exposure to feel confident in his
>>> >> ability to "rate", then "rating" his listening experience on a series
>>> >> of
>>> >> semantic scales. That's it. No comparison to anything other than his
>>> >> own
>>> >> subjective reaction to the music, and perhaps to live music. When
>>> >> you
>>> >> get 100-150 people (minimum) doing this for each of the variables
>>> >> (minimum two) then you can apply statistical difference measurements
>>> >> to
>>> >> determine what if anything is different, at what level of statistical
>>> >> significance.
>>> >>
>>> >> IMO this type of test represents the "gold standard" of audio
>>> >> testing.

>>> > Since each of those hundreds of people would apply their own
>>> > definition
>>> > and standard for said "variable", as well as the thresholds for values
>>> > on your "semantic scales", the results of such a test would be
>>> > absolutely
>>> > meaningless.

Semantic scales can work. They are a big part of the loudspeaker evaluation
methodology that is known as AES 20. The major difference between what AES
20 attempts to accomplish and what ABX is commonly used for is that AES20 is
designed to be used when it is known that there are large audible
differences.

>>> Sorry...the groups are randomly matched and the law of averages which is
>>> inherent in statistical evaluation comes into play. It is a tried and
>>> true appoach in the social sciences for many purposes.

Note being a practitioner of audio engineering, Harry doesn't know that what
works for sociologists doesn't necessarily work for engineers. One problem
with his approach is the necessary time scale. It's possible to do several
ABX tests with a reasonble number of skilled listeners in an afternoon.
Harry is describing processes that are done over a period of days and weeks.

>> Has it been used to test perception of difference?

> It is used to record perceptions in a quantifiable way. Then statistical
> analysis takes over to tell you whether or not their is a difference.

Using statistics and a number of trials and listeners has always been a
standard feature of ABX tests. However, a key to this process is having
listeners that can actually hear a difference. By eliminating the
possibility of close comparisons, Harry destroys the ability of his
listeners to hear small differences.

> That is the beauty of it...you don't have to "compare" it to
> anything....all you have to do is listen, listen, listen and then
> afterwards decided how you felt about your experience/the sound as you
> fill out the semantic differential scales. It is as close as you can come
> to normal listening in a test environment.

When it comes to reliably detecting subtle differences, this is a well-known
recipie for an absolute and total disaster. I challenge Harry to cite even
one sucessful test where his proposed methodology was sucessful in
establishing an audible JND at the thresholds of audibility.

William Sommerwerck

April 19th 07, 11:34 AM

> It is used to record perceptions in a quantifiable way. Then statistical
> analysis takes over to tell you whether or not their is a difference. That
> is the beauty of it... you don't have to "compare" it to anything... all
you
> have to do is listen, listen, listen and then afterwards decided how you
> felt about your experience/the sound as you fill out the semantic
> differential scales. It is as close as you can come to normal listening
> in a test environment.

Thank you, Harry.

This is the way people tend to listen over the long term (qv, my previously
posted example with the Acoustat system, or John Atkinson's report of his
experience with the QUAD 405). The non-neurotic audiophile (yes, Arny, there
are such people), doesn't generally upgrade his equipment until he has a
good reason to do so.

The purpose of long-term listening is to get a feeling for what a particular
component "sounds like". The theory is that if a component has perceptible
distortions or colorations, these will begin to reveal themselves over a
period of time, because they're audible on all (or most) recordings. If
multiple listeners, particularly when listening blind, report that the
Plotztron 68 sounds grainy and brittle, one can reasonably conclude that it
does.

Yes, such testing is (theoretically) less-sensitive to errors near the
threshold of perceptibility. So what? Why change your equipment to get rid
of a problem you're not aware of?

William Sommerwerck

April 19th 07, 11:35 AM

> Those tests are nice, but have nothing to do with Hi(gh)Fi(dellity)
> what aims at a _minimal_difference_ with the original.

Read my preceding post. Also consider that there are multiple ways to test
equipment, and multiple reasons for doing so.

William Sommerwerck

April 19th 07, 11:41 AM

> Note being a practitioner of audio engineering, Harry doesn't know
> that what works for sociologists doesn't necessarily work for engineers.

???

> One problem with his approach is the necessary time scale. It's
> possible to do several ABX tests with a reasonble number of skilled
> listeners in an afternoon. Harry is describing processes that are done
> over a period of days and weeks.

Who cares how long a test takes, if it reveals useful information? We're
looking for validity, not vitesse.

> When it comes to reliably detecting subtle differences, this is a
well-known
> recipe for an absolute and total disaster. I challenge Harry to cite even
> one sucessful test where his proposed methodology was sucessful in
> establishing an audible JND at the thresholds of audibility.

Again, see my other post. Not all listeners are interested in JNDs.

If you compared a new amp with one you owned, and the sonic differences,
though in favor of the new amp, were just barely noticeable, would you sell
the old amp and buy the new one? Probably not.

I've been reviewing a doctoral thesis about ambience extraction in which
monadic testing was used, and even recommended. The candidate's advisors
apparently did not advise him about this dangerous protocol.

Arny Krueger

April 19th 07, 11:51 AM

"William Sommerwerck" > wrote in message
...

>> Note being a practitioner of audio engineering, Harry doesn't know
>> that what works for sociologists doesn't necessarily work for engineers.

> ???

Well, you're not one either, William.

>> One problem with his approach is the necessary time scale. It's
>> possible to do several ABX tests with a reasonble number of skilled
>> listeners in an afternoon. Harry is describing processes that are done
>> over a period of days and weeks.

> Who cares how long a test takes, if it reveals useful information?

Ironic given that one common complaint with ABX tests is that they take too
long to be practical for many things. (true!)

> We're looking for validity, not vitesse.

No problem here with finding valid tests.

>> When it comes to reliably detecting subtle differences, this is a
> well-known
>> recipe for an absolute and total disaster. I challenge Harry to cite even
>> one sucessful test where his proposed methodology was sucessful in
>> establishing an audible JND at the thresholds of audibility.

> Again, see my other post. Not all listeners are interested in JNDs.

Actually, they are. There's not a big market for differences that can't be
noticed. I don't see a lot of equipment reviews that talk about differences
that can't be noticed.

> If you compared a new amp with one you owned, and the sonic differences,
> though in favor of the new amp, were just barely noticeable, would you
> sell
> the old amp and buy the new one? Probably not.

I'd like to make that decision based on reliable information.

> I've been reviewing a doctoral thesis about ambience extraction in which
> monadic testing was used, and even recommended. The candidate's advisors
> apparently did not advise him about this dangerous protocol.

It's about using the right tool for the job. I've already cited AES20 which
advises the use of monadic testing for loudspeakers. So the point is moot.

William Sommerwerck

April 19th 07, 12:11 PM

>> Again, see my other post. Not all listeners are interested in JNDs.

> Actually, they are. There's not a big market for differences that can't be
> noticed. I don't see a lot of equipment reviews that talk about
differences
> that can't be noticed.

I said NOT ALL. I'm not a loony audiophile who changes his equipment like
socks.

Over the years as a Stereophile reviewer, I gradually stopped talking about
subtle differences, mostly because I wasn't confident of my ability to hear
them reliably or repeatably.

>> If you compared a new amp with one you owned, and the sonic differences,
>> though in favor of the new amp, were just barely noticeable, would you
>> sell the old amp and buy the new one? Probably not.

> I'd like to make that decision based on reliable information.

That's assumed in the question. The answer is still "probably not".

Arny Krueger

April 19th 07, 12:36 PM

"William Sommerwerck" > wrote in message
. ..

>>> Again, see my other post. Not all listeners are interested in JNDs.

>> Actually, they are. There's not a big market for differences that can't
>> be
>> noticed. I don't see a lot of equipment reviews that talk about
> differences > that can't be noticed.

> I said NOT ALL. I'm not a loony audiophile who changes his equipment like
> socks.

Excluded middle argument noted. Not all listeners are interested in much of
anything.

> Over the years as a Stereophile reviewer, I gradually stopped talking
> about
> subtle differences, mostly because I wasn't confident of my ability to
> hear
> them reliably or repeatably.

At this point in life, I'm sure I can't hear a lot of things that younger
well-trained listeners can easily hear. That's one reason why I did
www.pcabx.com - to get younger listeners involved. It worked, as the
hydrogen audio forum shows.

>>> If you compared a new amp with one you owned, and the sonic differences,
>>> though in favor of the new amp, were just barely noticeable, would you
>>> sell the old amp and buy the new one? Probably not.
>
>> I'd like to make that decision based on reliable information.
>
> That's assumed in the question.

You can't make that assumption William, because you've never done a reliable
listening test, right?

> The answer is still "probably not".

Not a given.

William Sommerwerck

April 19th 07, 12:50 PM

>>>> If you compared a new amp with one you owned, and the sonic
>>>> differences, though in favor of the new amp, were just barely
>>>> noticeable, would you sell the old amp and buy the new one?
>>>> Probably not.

>>> I'd like to make that decision based on reliable information.

>> That's assumed in the question.

> You can't make that assumption William, because you've never
> done a reliable listening test, right?

We're talking about YOU, Arny. And, in your belief system, you have a source
for reliable testing.

William Sommerwerck

April 19th 07, 12:51 PM

> >> Those tests are nice, but have nothing to do with Hi(gh)Fi(dellity)
> >> what aims at a _minimal_difference_ with the original.
> >
> > Read my preceding post. Also consider that there are multiple ways to
> > test equipment, and multiple reasons for doing so.
>
> This one?

No. The recent preceding post.

Arny Krueger

April 19th 07, 12:59 PM

"Chel van Gennip" > wrote in message
...

> Anyhow, IMHO tests that try to determin a _minimal_difference_ should
> include a comparision.

There's really no choice. Zillions of experiments show that the ability to
detect really small amounts of distortion goes away when the comparison
involves a separation of more than a few seconds.

My argument is that experiments with close juxtapositioning have practical
significance because there is a natural ability to hear similar amounts of
distortion in some long term-listening experiences. When those long term
listening experiences are dissected, we find that when long term listening
works, it works because it functions like a string of short-term listening
tests. IOW, there is repetition of situations where clean and distorted
sounds are closely juxtapositioned by the natural flow of the music.

Distortion in the real world is often conditional, so the conditions line up
one way and there is more distortion, and then they don't line up that way,
there is less distortion. That provides the required comparands.

Those put-your-teeth-on edge situations seem to always involve relatively
large amounts of distortion.

The mystery is not that people hear such small amounts of distortion, but
that circumstances conspire to generate so much distortion.

In situ-testing will reveal a lot of those cases of unexpected high
distortion, but there isn't a lot of in situ testing being done.

Most bench tests of power amps for example, are done with resistive loads.
The few reactive-load tests that are done often involve what I think is a
wimpy loudspeaker simulator, like Stereophile's. Like Stereophile I started
out with Kantor's design, and doubled it up to make it that much tougher.
Kantor designed his simulator like he designs his speakers - basically easy
to drive. That's smart for a speaker designer, but not smart for an
amplifier tester. Of course, I'm not in the business of promoting tubed amps
and particularly not SETs, so I can put the pedal to the metal.

Scott Dorsey

April 19th 07, 01:19 PM

Steven Sullivan > wrote:
>
>So let's see...Mike's talking about an effect 'down in the audible range'. Thus if we set up a
>mike in the room where those pipes are playing, and faithfully capture only what's in the
>audible range (20-20kHz give or take a few) ...aren't we recording the effect? Why record the
>'cause' as well?

It depends where the mixing products are being generated, whether they are
happening in the room or in the ear.

And note that the recording process is also going to generate mixing products
due to IMD, and they might not be the same ones the ear would. Although, it
is possible that with a little DSP they could be made to be.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Arny Krueger

April 19th 07, 01:38 PM

"Chel van Gennip" > wrote in message
...
> On Thu, 19 Apr 2007 13:59:47 +0200, Arny Krueger wrote:
>
>
>> "Chel van Gennip" > wrote in message
>> ...
>>
>>> Anyhow, IMHO tests that try to determin a _minimal_difference_ should
>>> include a comparision.
>>
>> There's really no choice. Zillions of experiments show that the ability
>> to detect really small amounts of distortion goes away when the
>> comparison involves a separation of more than a few seconds.
>
> Sometimes it is good to look at the starting point. For this discussion
> about tests the starting point was:
>
> On Sat, 14 Apr 2007 16:13:08 +0200, Chel van Gennip wrote:

>> As this is as pro(duction) group, it is possible to do simple tests:
>> compare input and output. If the output of an AD-DA chain is better or
>> worse than the input, the device is wrong.

I totally agree.

> I still think it is the best way to test _minimal_difference_ , although
> I know some people want to leave the original out of the test, as they
> don't have the original available under normal listening conditions.

I totally agree. I' e done this kind of test on a number of devices and as
in as unbiased fashion as is possible. It's also a test that can be an exact
test using the PCABX methodology. The results of a number of test-setups is
available for download here:

http://www.pcabx.com/product/soundcard/index.htm

Arny Krueger

April 19th 07, 01:44 PM

"William Sommerwerck" > wrote in message
. ..
>>>>> If you compared a new amp with one you owned, and the sonic
>>>>> differences, though in favor of the new amp, were just barely
>>>>> noticeable, would you sell the old amp and buy the new one?
>>>>> Probably not.
>
>>>> I'd like to make that decision based on reliable information.

>>> That's assumed in the question.

>> You can't make that assumption William, because you've never
>> done a reliable listening test, right?

> We're talking about YOU, Arny. And, in your belief system, you have a
> source
> for reliable testing.

Nihilists need not apply for positions in audio production. ;-)

More to the point - the golden ear strategy is perfectly clear. They
cover-up their lack of ability and dilligence by attempting to trash the
good work of people who do have the ability and are willing to expend the
effort to do good listening tests.

One means of attempted trashing is to deny the good work that has been done,
and to set illogical (see Harry's monadic tests for subtle differences) and
impossible (see William's imaginary listening test that has not even risen
to the standard of being hypothetical) for real world examples of the same.

Once again William, take this discussion to RAHE or watch your posts on RAP
go unanswered.

NOW!

Arny Krueger

April 19th 07, 02:14 PM

"Steven Sullivan" > wrote in message
...
> Harry Lavo > wrote:
>
>> "Mike Rivers" > wrote in message
>> ups.com...
>> > On Apr 14, 7:37 am, "Natalie Drest" >
>> > wrote:
>> >
>
>> >snip<
>
>> >
>> > I did learn something interesting about wide bandwidth recording from
>> > the people who make a monstrous sample library with which they
>> > simulate large pipe organs with multi-channel installations. They
>> > record each pipe individually rather than recording the blended sound
>> > of the organ from a distance, then they play back those pipes and let
>> > them combine naturally in air. What they claim (and can prove by
>> > looking at the frequency content of their recordings) is that there's
>> > ultrasonic energy coming off a pipe, and when two pipes are playing
>> > together, the beating of those ultrasonic frequencies is down in the
>> > audible range. What you hear sounds different with and without that
>> > ultrasonic energy.
>> >
>
>> >snip<
>
>> Interesting. Mike, my hunch is that this is what Oohashi's subjects
>> experienced with the recorded gamelan music they used in his ultrasonic
>> test. The frequency overtones of the gongs, xylophones, and drums
>> extended
>> to very high frequencies, and so have the potential to create lots of
>> "beats". As you recall, in that test of both hearing and physiology
>> subjects recorded (at a statistically significant level) that the music
>> sounded more pleasant (and their pleasure centers were activated) when
>> the
>> ultrasonics were included but not when they were excluded. And yet the
>> ultrasonics reproduced separately from the audible music were not
>> audible.

> So let's see...Mike's talking about an effect 'down in the audible range'.
> Thus if we set up a
> mike in the room where those pipes are playing, and faithfully capture
> only what's in the
> audible range (20-20kHz give or take a few) ...aren't we recording the
> effect?

First off, note that the effect is hypothetical. If it exists, it should be
readily measurable and quantifiable.

Most natural sounds have measurable components at ultrasonic frequencies,
but few of them have strong response above 20 KHz. The instruments listed
generally have their peak power output in a 1/3 octave that is below 20 KHz.
Their response falls at -12 dB/octave or faster above the peak. This means
that energy around 100 KHz might only be at 30-40 dB SPL. IOW, so faint as
to be benign.

> Why record the 'cause' as well?

There's a good reason to *not* record the ultraonic causes, in order to
ensure more faithful reproduction of the effect, on less-than-perfect
reproduction equipment.

There's an old saying about opening windows too far and just letting in more
dirt.

William Sommerwerck

April 19th 07, 04:10 PM

> More to the point - the golden ear strategy is perfectly clear. They
> cover-up their lack of ability and dilligence by attempting to trash the
> good work of people who do have the ability and are willing to expend the
> effort to do good listening tests.

> One means of attempted trashing is to deny the good work that has been
done,
> and to set illogical (see Harry's monadic tests for subtle differences)
and
> impossible (see William's imaginary listening test that has not even risen
> to the standard of being hypothetical) for real world examples of the
same.

> Once again William, take this discussion to RAHE or watch your posts on
RAP
> go unanswered.

> NOW!

**** ***

Harry Lavo

April 19th 07, 05:15 PM

"Chel van Gennip" > wrote in message
...
> On Thu, 19 Apr 2007 04:59:38 +0200, Harry Lavo wrote:
>
>> Then statistical analysis takes over to tell you whether or not their is
>> a difference. That is the beauty of it...you don't have to "compare" it
>> to anything....all you have to do is listen, listen, listen and then
>> afterwards decided how you felt about your experience/the sound as you
>> fill out the semantic differential scales.
>
> Those tests are nice, but have nothing to do with Hi(gh)Fi(dellity) what
> aims at a _minimal_difference_ with the original. In fact those tests show
> many people like a rather high level of even harmonics (or as you may say:
> warm tubeamps without propper feedback).

It doesn't matter...the test will tell you if there is a difference and why
(in a listening sense). If you don't want to design and market the
equipment, or use whatever is prefered, don't.

Arny Krueger

April 19th 07, 06:29 PM

"Harry Lavo" > wrote in message
. ..
>
> "Chel van Gennip" > wrote in message
> ...
>> On Thu, 19 Apr 2007 04:59:38 +0200, Harry Lavo wrote:
>>
>>> Then statistical analysis takes over to tell you whether or not their is
>>> a difference. That is the beauty of it...you don't have to "compare" it
>>> to anything....all you have to do is listen, listen, listen and then
>>> afterwards decided how you felt about your experience/the sound as you
>>> fill out the semantic differential scales.

Eliminate timely comparisons, and listener sensitivity falls down a big,
deep hole.

>> Those tests are nice, but have nothing to do with Hi(gh)Fi(dellity) what
>> aims at a _minimal_difference_ with the original. In fact those tests
>> show
>> many people like a rather high level of even harmonics (or as you may
>> say:
>> warm tubeamps without propper feedback).

Like you say Chel, fairly gross levels of distortion, such as the ones you
find in transducers, particularly speakers.

> It doesn't matter...the test will tell you if there is a difference and
> why (in a listening sense).

Monadic tests are only useful for fairly gross levels of distortion,
compared to many of the issues that audiophiles are interested in.

Statistical analysis is one of those things we included in ABX. We found
that if you have to resort to large numbers of samples and complex analysis,
the difference is so small as to be pretty moot.

Steven Sullivan

April 19th 07, 09:25 PM

Harry Lavo > wrote:

> "Steven Sullivan" > wrote in message
> ...
> > Harry Lavo > wrote:
> >
> >> "Mike Rivers" > wrote in message
> >> ups.com...
> >> > On Apr 14, 7:37 am, "Natalie Drest" >
> >> > wrote:
> >> >
> >
> >> >snip<
> >
> >> >
> >> > I did learn something interesting about wide bandwidth recording from
> >> > the people who make a monstrous sample library with which they
> >> > simulate large pipe organs with multi-channel installations. They
> >> > record each pipe individually rather than recording the blended sound
> >> > of the organ from a distance, then they play back those pipes and let
> >> > them combine naturally in air. What they claim (and can prove by
> >> > looking at the frequency content of their recordings) is that there's
> >> > ultrasonic energy coming off a pipe, and when two pipes are playing
> >> > together, the beating of those ultrasonic frequencies is down in the
> >> > audible range. What you hear sounds different with and without that
> >> > ultrasonic energy.
> >> >
> >
> >> >snip<
> >
> >> Interesting. Mike, my hunch is that this is what Oohashi's subjects
> >> experienced with the recorded gamelan music they used in his ultrasonic
> >> test. The frequency overtones of the gongs, xylophones, and drums
> >> extended
> >> to very high frequencies, and so have the potential to create lots of
> >> "beats". As you recall, in that test of both hearing and physiology
> >> subjects recorded (at a statistically significant level) that the music
> >> sounded more pleasant (and their pleasure centers were activated) when
> >> the
> >> ultrasonics were included but not when they were excluded. And yet the
> >> ultrasonics reproduced separately from the audible music were not
> >> audible.
> >
> >
> >
> > So let's see...Mike's talking about an effect 'down in the audible range'.
> > Thus if we set up a
> > mike in the room where those pipes are playing, and faithfully capture
> > only what's in the
> > audible range (20-20kHz give or take a few) ...aren't we recording the
> > effect? Why record the
> > 'cause' as well?

> Because the effect is created in your room, from the "cause". That's why
> they recorded each pipe separately.

Why not just record the effect? Do we routinely mic every string in a piano?

___
-S
"As human beings, we understand the world through simile, analogy,
metaphor, narrative and, sometimes, claymation." - B. Mason