[Tobiah]

Wikipedia says they will offer music in 24bit/192kHz format.
There has been endless debate here over whether 96kHz is better
than 44.1, most admitting that they can't hear the difference.
192kHz however, seems like blatant overkill to me.

Toby

[John Williamson]

On 09/09/2014 18:34, Tobiah wrote:
Wikipedia says they will offer music in 24bit/192kHz format.
There has been endless debate here over whether 96kHz is better
than 44.1, most admitting that they can't hear the difference.
192kHz however, seems like blatant overkill to me.


It has its uses when you are clearing artifacts such as scratches and
noise from digital copies of vinyl and tape recordings.

Other than that, I agree with you, except that for some material with
extended HF, I would record at double the sample frequency of the item
to be published, especially now that storage is so cheap and converters
are so much better than 20 years ago.

--
Tciao for Now!

John.

[Mike Rivers[_2_]]

On 9/9/2014 1:34 PM, Tobiah wrote:
There has been endless debate here over whether 96kHz is better
than 44.1, most admitting that they can't hear the difference.
192kHz however, seems like blatant overkill to me.

It is, but that's one method of marketing - making you believe that
overkill is better than what's good enough for most people.

The thing they're counting on is that when you have this high fidelity
playback gizmo, you'll plop yourself down in front of your $20,000
speakers driven by a pair of $5000 single triode amplifiers and spend
half an hour paying attention to how good your investment sounds instead
of listening to the same songs through earbuds connected to your iPhone
while exercising in the gym on your lunch hour.

--
For a good time, visit http://mikeriversaudio.wordpress.com

[Les Cargill[_4_]]

Tobiah wrote:
Wikipedia says they will offer music in 24bit/192kHz format.
There has been endless debate here over whether 96kHz is better
than 44.1, most admitting that they can't hear the difference.
192kHz however, seems like blatant overkill to me.

Toby

For production material, it may well be worth it.

For distribution... not so much.

--
Les Cargill

[Jay Ts[_4_]]

On Tue, 09 Sep 2014 10:34:23 -0700, Tobiah wrote:

There has
been endless debate here over whether 96kHz is better than 44.1, most
admitting that they can't hear the difference.

Why please most people when you can satisfy everyone?

If you don't want to use 192 KHz, no one has a gun to your head.

192kHz however, seems like blatant overkill to me.

That is also a good reason to make it a standard, along with supporting
other formats, of course. If there is debate over 48 KHz vs. 92 KHz, then
why not support at least one higher than 96 KHz, just to make sure?

If you have both 96 KHz and 192 KHz available, then people can try both
to see which they prefer. Then no one needs debates by a few who make
choices for everyone.

Even if few use it, I think it's a good idea to support a standard that
no one can find fault with.

If you were driving over a bridge that said "Maximum 1 ton" in a car that
was almost a ton, would you feel as comfortable as if it said "2 tons"?
Or more?

As I see it, supporting 192 KHz is simply a good engineering practice.

[Sean Conolly]

"Tobiah" wrote in message
...
Wikipedia says they will offer music in 24bit/192kHz format.
There has been endless debate here over whether 96kHz is better
than 44.1, most admitting that they can't hear the difference.
192kHz however, seems like blatant overkill to me.

For distribution it is. Even if you ignore the accepted limits of human
hearing, good luck finding speakers that go two octaves over human hearing,
especially with low distortion.

But I still want to be able to record even higher than that for measuring
both the response and the harmonic distortion of gear. If there's a
prominant 3rd harmonic of a 15K sine wave, I'd like to know that to figure
out why.

Sean

[hank alrich]

Jay Ts wrote:

As I see it, supporting 192 KHz is simply a good engineering practice.

Based on something other than assumtions that more is automatically
better?

http://lavryengineering.com/pdfs/lav...ing-theory.pdf


--
shut up and play your guitar * HankAlrich.Com
HankandShaidriMusic.Com
YouTube.Com/WalkinayMusic

[Scott Dorsey]

Jay Ts wrote:
On Tue, 09 Sep 2014 10:34:23 -0700, Tobiah wrote:

There has
been endless debate here over whether 96kHz is better than 44.1, most
admitting that they can't hear the difference.

Why please most people when you can satisfy everyone?

If you don't want to use 192 KHz, no one has a gun to your head.

Right, and I can just take my 44.1 recordings and upsample them to 192 for
sale online and nobody will ever know the difference.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Mike Rivers[_2_]]

On 9/10/2014 8:21 AM, Scott Dorsey wrote:
Right, and I can just take my 44.1 recordings and upsample them to 192 for
sale online and nobody will ever know the difference.

When you do that, does it interpolate or re-sample to get the additional
sample values, or just stuff in three more samples of the same value? If
you had four samples of the same value followed by four samples of
another value, and so on, it would be a fair guess that it was
up-sampled rather than re-sampled, interpolated, or actually recorded at
the higher sample rate.



--
For a good time, visit http://mikeriversaudio.wordpress.com

[Scott Dorsey]

In article , Mike Rivers wrote:
On 9/10/2014 8:21 AM, Scott Dorsey wrote:
Right, and I can just take my 44.1 recordings and upsample them to 192 for
sale online and nobody will ever know the difference.

When you do that, does it interpolate or re-sample to get the additional
sample values, or just stuff in three more samples of the same value?

If you are smart and do it properly, it interpolates and resamples, it does
not just duplicate samples. Duplicating samples will give you accurate
response in the original passband but a whole lot of weird correlated noise
above it. Properly interpolating, you get no noise above the maximum
frequency of the original recording.

If
you had four samples of the same value followed by four samples of
another value, and so on, it would be a fair guess that it was
up-sampled rather than re-sampled, interpolated, or actually recorded at
the higher sample rate.

Just looking at a spectrum and seeing an abrupt cutoff at 22.05 kc will make
it pretty obvious what happened.... but nobody will bother.
--scott

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

[William Sommerwerck]

"Tobiah" wrote in message ...

Wikipedia says they will offer music in 24bit/192kHz format.
There has been endless debate here over whether 96kHz is
better than 44.1, most admitting that they can't hear the
difference. 192kHz however, seems like blatant overkill to me.

If human beings cannot hear beyond 22kHz, then 44.1kHz sampling doesn't lose
any audible information. That doesn't mean higher sampling rates aren't
desirable.

To me, 44.1kHz is slightly inferior to DSD. * Part of this might be due to
fact that musical instruments have ultrasonic components. (I'm assuming mics
capable of picking them up are used.) It's reasonable to wonder whether even
the "brick wall" filtering used blocks them sufficiently to prevent audible
aliasing artifacts.

Of course, no one has ever bothered to test this, because it's easier to
//assume// rather than experiment.

To retain any theoretical advantages of a higher sampling rate, you have to
maintain the higher rate. Down sampling requires the same sharp filtering
needed if you'd originally recorded at a lower rate.

* I can't hear above 12kHz, so the apparent loss of "quality" is something
that extends below the top octave.

[Dave Plowman (News)]

In article ,
William Sommerwerck wrote:
To me, 44.1kHz is slightly inferior to DSD. * Part of this might be due
to fact that musical instruments have ultrasonic components. (I'm
assuming mics capable of picking them up are used.)

Which common high quality mics go above 22K?

--
*A fine is a tax for doing wrong. A tax is a fine for doing well*

Dave Plowman London SW
To e-mail, change noise into sound.

[Don Pearce[_3_]]

On Wed, 10 Sep 2014 18:21:45 +0100, "Dave Plowman (News)"
wrote:

In article ,
William Sommerwerck wrote:
To me, 44.1kHz is slightly inferior to DSD. * Part of this might be due
to fact that musical instruments have ultrasonic components. (I'm
assuming mics capable of picking them up are used.)

Which common high quality mics go above 22K?

Oh they all go above 22k - in a mess of peaks and troughs that you
absolutely have to get rid of.

d

[Scott Dorsey]

In article ,
William Sommerwerck wrote:
"Tobiah" wrote in message ...

Wikipedia says they will offer music in 24bit/192kHz format.
There has been endless debate here over whether 96kHz is
better than 44.1, most admitting that they can't hear the
difference. 192kHz however, seems like blatant overkill to me.

If human beings cannot hear beyond 22kHz, then 44.1kHz sampling doesn't lose
any audible information. That doesn't mean higher sampling rates aren't
desirable.

It also doesn't mean they aren't desirable. If your recording chain can
accurately record ultrasonics, you're now dumping ultrasonics into your
playback chain which increases the chance of getting distortion products
down in the audible range. Your playback chain now not only needs to be
linear across the audible range, it needs to be linear beyond it.

To me, 44.1kHz is slightly inferior to DSD. * Part of this might be due to
fact that musical instruments have ultrasonic components. (I'm assuming mics
capable of picking them up are used.) It's reasonable to wonder whether even
the "brick wall" filtering used blocks them sufficiently to prevent audible
aliasing artifacts.

Try taking a 44.1 recording and resampling it as DSD and playing it back.
The differences between converters these days is greater than the differences
between sampling methods and rates. I have indeed heard converters that
sounded better at one rate than another, but that wasn't due to the rate
better better or worse, that was an artifact of the converter.

Of course, no one has ever bothered to test this, because it's easier to
//assume// rather than experiment.

Actually, some folks did a respectable test on audibility in the JAES a couple
years ago. It wasn't perfect but it was not badly conducted.

To retain any theoretical advantages of a higher sampling rate, you have to
maintain the higher rate. Down sampling requires the same sharp filtering
needed if you'd originally recorded at a lower rate.

There are no theoretical advantages, though, aside from wider bandwidth.
There may well be some practical avantages, though.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Scott Dorsey]

Dave Plowman (News) wrote:
In article ,
William Sommerwerck wrote:
To me, 44.1kHz is slightly inferior to DSD. * Part of this might be due
to fact that musical instruments have ultrasonic components. (I'm
assuming mics capable of picking them up are used.)

Which common high quality mics go above 22K?

A lot of them, even the SM-57. It's down a lot compared with the 1kc nominal
level, but it's not down enough to prevent aliasing if you omit proper
filtration.

One manufacturer is selling a special microphone with flat response out to
40kc... which is fact just their normal microphone with an equalizer built
into the electronics. Which is shameful.
--scott

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

[William Sommerwerck]

"Dave Plowman (News)" wrote in message ...
In article ,
William Sommerwerck wrote:

To me, 44.1kHz is slightly inferior to DSD. * Part of this might be due
to fact that musical instruments have ultrasonic components. (I'm
assuming mics capable of picking them up are used.)

Which common high quality mics go above 22K?

I used to own Pearl mics. All went past 20kHz, and at least one made it to
24kHz.

[William Sommerwerck]

"Scott Dorsey" wrote in message ...
In article ,
William Sommerwerck wrote:

If human beings cannot hear beyond 22kHz, then 44.1kHz sampling
doesn't lose any audible information. That doesn't mean higher
sampling rates aren't desirable.

It also doesn't mean they aren't [sic] desirable. If your recording chain
can
accurately record ultrasonics, you're now dumping ultrasonics into your
playback chain which increases the chance of getting distortion products
down in the audible range. Your playback chain now not only needs to be
linear across the audible range, it needs to be linear beyond it.

But there's nothing "magical" about any particular bandwidth. Your amplifier
doesn't "know" it's trying to reproduce inaudible ultrasonics, and out of
spite, screws up the sound.


To retain any theoretical advantages of a higher sampling rate, you
have to maintain the higher rate. Down sampling requires the same
sharp filtering needed if you'd originally recorded at a lower rate.

There are no theoretical advantages, though, aside from wider bandwidth.
There may well be some practical avantages, though.

Which, ultimately, is the issue.

[Scott Dorsey]

William Sommerwerck wrote:
"Scott Dorsey" wrote in message ...
In article ,
William Sommerwerck wrote:

If human beings cannot hear beyond 22kHz, then 44.1kHz sampling
doesn't lose any audible information. That doesn't mean higher
sampling rates aren't desirable.

It also doesn't mean they aren't [sic] desirable. If your recording chain
can
accurately record ultrasonics, you're now dumping ultrasonics into your
playback chain which increases the chance of getting distortion products
down in the audible range. Your playback chain now not only needs to be
linear across the audible range, it needs to be linear beyond it.

But there's nothing "magical" about any particular bandwidth. Your amplifier
doesn't "know" it's trying to reproduce inaudible ultrasonics, and out of
spite, screws up the sound.

There IS something magical about 20 KHz, because we can't hear anything above
it. So if you are going to bandlimit a signal to reduce intermodulation
effects, bandlimiting it to 20 KHz is a good plan.

To retain any theoretical advantages of a higher sampling rate, you
have to maintain the higher rate. Down sampling requires the same
sharp filtering needed if you'd originally recorded at a lower rate.

There are no theoretical advantages, though, aside from wider bandwidth.
There may well be some practical avantages, though.

Which, ultimately, is the issue.

Maybe, but the problem is that the advantages are reduced more and more as
converters are improved. Remember the Panasonic SV3700 where you could tell
dramatic differences between 44.1 and 48 ksamp/sec recordings? Those days
are gone.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Tobiah]

If you don't want to use 192 KHz, no one has a gun to your head.

If I'm forced to download an album of that size, then convert to
my preferred format, that's at least a pea shooter.


Sure, _support_. The way I read it, the distribution is only in 24/192.
We'll see what they end up with.

Even if few use it, I think it's a good idea to support a standard that
no one can find fault with.

If you were driving over a bridge that said "Maximum 1 ton" in a car that
was almost a ton, would you feel as comfortable as if it said "2 tons"?
Or more?

Yeah, but if it was a small town 2 lane bridge and it said is supported 25
tons and it hit the town financially to build it, then there is a better way.

[geoff]

On 11/09/2014 3:02 a.m., Mike Rivers wrote:
On 9/10/2014 8:21 AM, Scott Dorsey wrote:
Right, and I can just take my 44.1 recordings and upsample them to 192
for
sale online and nobody will ever know the difference.

When you do that, does it interpolate or re-sample to get the additional
sample values, or just stuff in three more samples of the same value? If
you had four samples of the same value followed by four samples of
another value, and so on, it would be a fair guess that it was
up-sampled rather than re-sampled, interpolated, or actually recorded at
the higher sample rate.





Any converter that wasn't a POS would interpolate.

geoff

[Mike Rivers[_2_]]

On 9/10/2014 8:45 PM, geoff wrote:
Any converter that wasn't a POS would interpolate.

Name five and tell me how you know that they interpolate.

Thank you.

--
For a good time, visit http://mikeriversaudio.wordpress.com

[Neil Gould]

Tobiah wrote:
Wikipedia says they will offer music in 24bit/192kHz format.
There has been endless debate here over whether 96kHz is better
than 44.1, most admitting that they can't hear the difference.
192kHz however, seems like blatant overkill to me.

I don't care if it's 384kHz... it won't help Neil Young's tonal quality one
byte. That, and he's far too old to hear 16kHz accurately, so this smells
like snake oil to me. ;-)
--
best regards,

Neil

[Sean Conolly]

"William Sommerwerck" wrote in message
...
"Scott Dorsey" wrote in message ...
It also doesn't mean they aren't [sic] desirable. If your recording
chain can
accurately record ultrasonics, you're now dumping ultrasonics into your
playback chain which increases the chance of getting distortion products
down in the audible range. Your playback chain now not only needs to be
linear across the audible range, it needs to be linear beyond it.

But there's nothing "magical" about any particular bandwidth. Your
amplifier doesn't "know" it's trying to reproduce inaudible ultrasonics,
and out of spite, screws up the sound.

Not out of spite, but in general linearity decreases as frequency increases.
Phase shift also tends to increase which can can cause instability as it
approaches 180 degrees.

I would expect that good designs would have a low pass filter to prevent
anything really weird or harmful occuring, at least those that have a high
enough slew rate for it to be a concern.

Now what do tweeters do when given a signal higher than their design
bandwidth?

Sean

[Sean Conolly]

"Neil Gould" wrote in message
...
Tobiah wrote:
Wikipedia says they will offer music in 24bit/192kHz format.
There has been endless debate here over whether 96kHz is better
than 44.1, most admitting that they can't hear the difference.
192kHz however, seems like blatant overkill to me.

I don't care if it's 384kHz... it won't help Neil Young's tonal quality
one
byte. That, and he's far too old to hear 16kHz accurately, so this smells
like snake oil to me. ;-)

You'll need to buy Pono Certified speakers and cables, I'm sure. An
oxygen-free power cords for the amps.

Sean

[Jay Ts[_4_]]

On Tue, 09 Sep 2014 22:02:35 -0700, hank alrich wrote:

Jay Ts wrote:

As I see it, supporting 192 KHz is simply a good engineering practice.

Based on something other than assumtions that more is automatically
better?

More is not automatically better, but often is in ways that are
unexpected. This happens frequently in technological development,
including electronics and especially computers.

http://lavryengineering.com/pdfs/lav...ing-theory.pdf

I saw that about a year ago, and disregarded it due to its non-scientific
nature (weasel words, dogma, and denial). If you can, please provide
something better, based on good scientific thought and attitude.

[Don Pearce[_3_]]

On Thu, 11 Sep 2014 06:02:09 GMT, Jay Ts wrote:

On Tue, 09 Sep 2014 22:02:35 -0700, hank alrich wrote:

Jay Ts wrote:

As I see it, supporting 192 KHz is simply a good engineering practice.

Based on something other than assumtions that more is automatically
better?

More is not automatically better, but often is in ways that are
unexpected. This happens frequently in technological development,
including electronics and especially computers.

http://lavryengineering.com/pdfs/lav...ing-theory.pdf

I saw that about a year ago, and disregarded it due to its non-scientific
nature (weasel words, dogma, and denial). If you can, please provide
something better, based on good scientific thought and attitude.

That paper is a very thorough treatment of the question. I very much
doubt you will find anything better.

d

[geoff]

On 11/09/2014 1:09 p.m., Mike Rivers wrote:
On 9/10/2014 8:45 PM, geoff wrote:
Any converter that wasn't a POS would interpolate.

Name five and tell me how you know that they interpolate.

Thank you.


Sound Forge and Wavelab are the only two I have direct experience of.

Nothing claiming to be even vaguely profession would have the temetiy
rto do anything but interpolate.

Surely not even the cheesiest amateur apps would have the cheek to
simmply insert 'same' samples.

geoff

[geoff]

On 11/09/2014 2:56 p.m., Neil Gould wrote:
Tobiah wrote:
Wikipedia says they will offer music in 24bit/192kHz format.
There has been endless debate here over whether 96kHz is better
than 44.1, most admitting that they can't hear the difference.
192kHz however, seems like blatant overkill to me.

I don't care if it's 384kHz... it won't help Neil Young's tonal quality one
byte. That, and he's far too old to hear 16kHz accurately, so this smells
like snake oil to me. ;-)



If he could actually hear how crappy his later music has been, maybe he
would have tried harder....


geoff

[[email protected]]

Tobiah wrote: "Sep 9Tobiah
Wikipedia says they will offer music in 24bit/192kHz format.
There has been endless debate here over whether 96kHz is better
than 44.1, most admitting that they can't hear the difference.
192kHz however, seems like blatant overkill to me.

Toby "

Those bit depths/resolutions do have a place: In the production environment.. But unless dogs and other animals start buying music, they are pointless as consumer carriers. 16/44.1 more than adequately accommodates the vast majority of human hearing circumstances.

[Scott Dorsey]

In article , Mike Rivers wrote:
On 9/10/2014 8:45 PM, geoff wrote:
Any converter that wasn't a POS would interpolate.

Name five and tell me how you know that they interpolate.

r8brain does because I have seen the source code. sox does also, and I
can say so for the same reason.

The AD1890 and AD1894 hardware ones do also, and you can see the description
of how they work on the datasheet. However, they use somewhat cruder filters
than some of the software solutions do.

I don't know a fifth one offhand.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Scott Dorsey]

Sean Conolly wrote:
"William Sommerwerck" wrote in message
...
"Scott Dorsey" wrote in message ...
It also doesn't mean they aren't [sic] desirable. If your recording
chain can
accurately record ultrasonics, you're now dumping ultrasonics into your
playback chain which increases the chance of getting distortion products
down in the audible range. Your playback chain now not only needs to be
linear across the audible range, it needs to be linear beyond it.

But there's nothing "magical" about any particular bandwidth. Your
amplifier doesn't "know" it's trying to reproduce inaudible ultrasonics,
and out of spite, screws up the sound.

Not out of spite, but in general linearity decreases as frequency increases.
Phase shift also tends to increase which can can cause instability as it
approaches 180 degrees.

This is generally true.

I would expect that good designs would have a low pass filter to prevent
anything really weird or harmful occuring, at least those that have a high
enough slew rate for it to be a concern.

They don't necessarily, and that's where the problems occur. Which is why
if I were releasing a 96 ksamp/sec recording I would want to look carefully
on a spectrum analyzer to make sure there isn't any junk up there that
might be detrimental, if not just filter before releasing.

Now what do tweeters do when given a signal higher than their design
bandwidth?

Dick Pierce has a great story about this.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Neil Gould]

Jay Ts wrote:
On Tue, 09 Sep 2014 22:02:35 -0700, hank alrich wrote:

Jay Ts wrote:

As I see it, supporting 192 KHz is simply a good engineering
practice.

Based on something other than assumtions that more is automatically
better?

More is not automatically better, but often is in ways that are
unexpected. This happens frequently in technological development,
including electronics and especially computers.

http://lavryengineering.com/pdfs/lav...ing-theory.pdf

I saw that about a year ago, and disregarded it due to its
non-scientific nature (weasel words, dogma, and denial). If you can,
please provide something better, based on good scientific thought and
attitude.

I read this expecting to find something that supports your assetions
"weasel words..." etc., but did not. What did you find objectionable about
the content in this writing?
--
best regards,

Neil

[Mike Rivers[_2_]]

On 9/11/2014 3:12 AM, geoff wrote:
Sound Forge and Wavelab are the only two I have direct experience of.

Nothing claiming to be even vaguely profession would have the temetiy
rto do anything but interpolate.

Surely not even the cheesiest amateur apps would have the cheek to
simmply insert 'same' samples.

I should have been clearer. Any converter chip that claims 192 kHz
performance should actually produce the correct number of samples. You
answered the question relative to software (off-line) conversion. I'd
trust Wavelab and Sound Forge to do the right thing, and any honest
music supplier who offered high sample rate versions of things not
originally sampled at the high rate would probably convert them
responsibly.

But there are ways to do it wrong, and you know someone will, and won't
know it. You can't assume that just because it's not difficult for you
to find the right software or the right hardware if you're building an
up-sampling D/A converter box, that someone who thinks he knows what
he's doing won't be selling a deficient product. And the more popular
the concept gets, the more people who don't really know what they're
doing will be doing it.


--
For a good time, visit http://mikeriversaudio.wordpress.com

[hank alrich]

Jay Ts wrote:

On Tue, 09 Sep 2014 22:02:35 -0700, hank alrich wrote:

Jay Ts wrote:

As I see it, supporting 192 KHz is simply a good engineering practice.

Based on something other than assumtions that more is automatically
better?

More is not automatically better, but often is in ways that are
unexpected. This happens frequently in technological development,
including electronics and especially computers.

http://lavryengineering.com/pdfs/lav...ing-theory.pdf

I saw that about a year ago, and disregarded it due to its non-scientific
nature (weasel words, dogma, and denial). If you can, please provide
something better, based on good scientific thought and attitude.

Weasel worry all you want. The man makes some of the best convertors in
the world. He knows more about this stuff than most of us.

--
shut up and play your guitar * HankAlrich.Com
HankandShaidriMusic.Com
YouTube.Com/WalkinayMusic

[hank alrich]

Neil Gould wrote:

Jay Ts wrote:
On Tue, 09 Sep 2014 22:02:35 -0700, hank alrich wrote:

Jay Ts wrote:

As I see it, supporting 192 KHz is simply a good engineering
practice.

Based on something other than assumtions that more is automatically
better?

More is not automatically better, but often is in ways that are
unexpected. This happens frequently in technological development,
including electronics and especially computers.

http://lavryengineering.com/pdfs/lav...ing-theory.pdf

I saw that about a year ago, and disregarded it due to its
non-scientific nature (weasel words, dogma, and denial). If you can,
please provide something better, based on good scientific thought and
attitude.

I read this expecting to find something that supports your assetions
"weasel words..." etc., but did not. What did you find objectionable about
the content in this writing?

The main problem people appear to face with that paper is that it runs
directly counter to their assumptions. No response that I have read so
far directly takes Dan on with countering, supportable information.

--
shut up and play your guitar * HankAlrich.Com
HankandShaidriMusic.Com
YouTube.Com/WalkinayMusic

[Scott Dorsey]

hank alrich wrote:
Jay Ts wrote:

On Tue, 09 Sep 2014 22:02:35 -0700, hank alrich wrote:

Jay Ts wrote:

As I see it, supporting 192 KHz is simply a good engineering practice.

Based on something other than assumtions that more is automatically
better?

More is not automatically better, but often is in ways that are
unexpected. This happens frequently in technological development,
including electronics and especially computers.

http://lavryengineering.com/pdfs/lav...ing-theory.pdf

I saw that about a year ago, and disregarded it due to its non-scientific
nature (weasel words, dogma, and denial). If you can, please provide
something better, based on good scientific thought and attitude.

Weasel worry all you want. The man makes some of the best convertors in
the world. He knows more about this stuff than most of us.

The article isn't a scientific paper, it's a basic introduction to how sampling
theory works in the real world.

I'm curious what you mean, though, by "weasel words, dogma, and denial."
As an introductory tutorial it's pretty good, and if it does some handwaving
over details and proofs, that's the nature of a tutorial.

If you want proofs, I recommend "Introduction to Shannon Sampling and
Interpolation."

Note that he is going over the theory in a perfect world, and that this being
an imperfect world, we sometimes have to deal with imperfect hardware. High
rate sampling and oversampling techniques can help with this, but it's
important to know where they can help and where they can't.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[geoff]

On 12/09/2014 12:20 a.m., wrote:
Tobiah wrote: "Sep 9Tobiah
Wikipedia says they will offer music in 24bit/192kHz format.
There has been endless debate here over whether 96kHz is better
than 44.1, most admitting that they can't hear the difference.
192kHz however, seems like blatant overkill to me.

Toby "

Those bit depths/resolutions do have a place: In the production environment. But unless dogs and other animals start buying music, they are pointless as consumer carriers. 16/44.1 more than adequately accommodates the vast majority of human hearing circumstances.



And just to think of all those great-sounding recordings made on gear
that barely had any response over17KHz !


geoff

[Jay Ts[_4_]]

On Thu, 11 Sep 2014 08:28:31 -0500, Neil Gould wrote:
Jay Ts wrote:
On Tue, 09 Sep 2014 22:02:35 -0700, hank alrich wrote:

Jay Ts wrote:

As I see it, supporting 192 KHz is simply a good engineering
practice.

Based on something other than assumtions that more is automatically
better?

More is not automatically better, but often is in ways that are
unexpected. This happens frequently in technological development,
including electronics and especially computers.

http://lavryengineering.com/pdfs/lav...ing-theory.pdf

I saw that about a year ago, and disregarded it due to its
non-scientific nature (weasel words, dogma, and denial). If you can,
please provide something better, based on good scientific thought and
attitude.

I read this expecting to find something that supports your assetions
"weasel words..." etc., but did not. What did you find objectionable
about the content in this writing?

I'll have to retract the term "weasel words" because just now I re-read
the article, and it seems my memory of it from the previous reading was
inaccurate. Sorry about that one. (I could have sworn that I at least
encountered the phrase, "It is generally recognized that..." in the
article, but when I searched for it just now, it wasn't there! I may be
confused with something else I read on the subject recently.)

Anyway, here is what I find offensive about the article and how it is
being used (or maybe really mis-used) by people online. I have more time
today so I'll use some of it to explain my reactive attitude on this
topic. (Maybe over-reactive, but that's normal for me.

To quote,

"Research shows that musical instruments may produce energy above 20 KHz,
but there is little sound energy at above 40KHz. Most microphones do not
pick up sound at much over 20KHz. Human hearing rarely exceeds 20KHz."

I find all of these statements offensive. Although true, they mislead the
reader into thinking that there is absolutely no reason to save anything
above 20 KHz, and the rest of the article continues off of that idea,
which is not even literally present in the above quote! If you read
carefully and think a little, you might realize that there *is* content
above 20 KHz, and that some people miss out on that if it is not present
in the recording.

The reason I'm offended by this is not that I think I can hear above 20
KHz. I know I can't. The best I've ever heard at a conscious level is
about 17 KHz, and that was a long time ago. I'm not as good today. But I
did meet someone who could hear up to 30 KHz in a very informal blind
test. In science, all it takes to disprove any theory is one
counterexample, and that incident was enough to change my thinking.

I'm really concerned that if a few people (it might be only 1%) have
exceptional hearing at a conscious level, then maybe many more have
subconscious senses in the 20KHz range that may be significant somehow.
I don't know of any method of scientific research yet developed that can
test for that, so my attitude is to remain conservative and wait
patiently for more study. I'm waiting for higher resolution realtime
brain scanning technology and other things that probably haven't been
invented yet, and until then, I don't put much faith in listening tests.
In the meantime, I think it's best to design for the possibility that the
current assumptions may be wrong, and include support for higher
frequencies.

I was watching videos with Rupert Neve recently, and in one he told a
story about a recording engineer who complained about a fault in just one
channel of the studio's Neve console. Mr. Neve went in to investigate, at
first found nothing, but then studied more and found a fault at about 50
KHz in that channel. After fixing the circuit, the engineer was
satisfied. This is more anecdotal "evidence" that proves nothing, but I
think it's significant, and I wish I knew more of what happened. I think
things like this should be followed up on more to discern more about
what's happening. For that to happen, people must get beyond the dogma
that people can't hear anything above 20 KHz.

Mr. Neve's response to the incident was to improve his circuit designs,
and extend the frequency response to 120 KHz (I'm looking at his current
designs to get that number, since I don't remember what he said in the
video). He did not just say, "Well, I can't hear beyond 12 KHz so I don't
care about that."

This to me shows a good attitude towards engineering, with a good
scientific attitude of never being sure about anything, and being open-
minded. I've seen specs on many high-end "pro audio" products with
extended high-frequency response, and it seems that Mr. Neve is not alone
in thinking it is important to do so.

It really bothers me that if I have products with that quality, I may
have it cut back to a bandwidth of only 20 KHz, or even 40 KHz, at any
later stage in the signal chain, especially the last one at the recording.

Another reason I don't like the quote is that it assumes a lot about
things like musical instruments and sound reproduction equipment. Maybe
someone tomorrow will invent a microphone and speaker that can accurately
reproduce sound at much higher frequencies. Who knows? I think if the
industry standards don't support using them, that would really suck.
Let's not make assumptions about the future based on things from the past.

The real point is that the standards for digital recording, processing
and distribution have the effect of setting the status quo. Once a
standard is set, everything in the future is limited by it. So rather
than set the standard to a minimum, isn't it better to have them set
higher? At least some provision needs to be made for applications that
don't fit the norm, to allow further development to happen as it is
needed.

One example I keep thinking about is in the area of scientific research.
In recent years, scientists have studied various species of animals and
found that their vocalizations have volcabulary and grammar. Many animals
have the ability to make and perceive sounds with frequencies above 20
KHz, and if scientists want to study them, they will need equipment that
supports frequencies much higher than humans can hear. Of course it's
possible to design and build custom equipment to do that, but it can be
prohibitively expensive, time-consuming, and beyond the capability of
many good biologists to figure out how to manage a research project that
requires it. Another issue is that there is a problem in science now
because the research results of not too long ago are being lost because
the data was recorded in weird ways that cannot be maintained. Using
industry standard data representations would help avoid that kind of
thing.

Another area is ultrasonics. I keep thinking that if sound cards were
commonly available that had good support for high frequencies, some
clever nerds might create some cool new applications for it. Maybe even
some kind of disruptive technologies or something that helps save the
world.

Years ago I learned of software-defined radio technology. A quote from
the Wikipedia article on this topic:

"Software-defined radio (SDR) is a radio communication system where
components that have been typically implemented in hardware (e.g. mixers,
filters, amplifiers, modulators/demodulators, detectors, etc.) are
instead implemented by means of software on a personal computer or
embedded system.[1] While the concept of SDR is not new, the rapidly
evolving capabilities of digital electronics render practical many
processes which used to be only theoretically possible."

At the time, I was simply amazed that someone thought of that idea. They
realized that computers had become fast enough that parts of the
electronic circuitry in radio systems could be replaced by software
driving the computer's sound card, even though the sound card was limited
to audio frequencies. This has allowed radio systems to be controlled in
new ways and do more things. If sound cards supported higher frequencies,
I wonder what more might be possible.

I think the response of some people here is "Who cares? that is another
profession, and that's their problem. It has nothing to do with this." To
head that off, I think it is very unprofessional for anyone to be so
selfish as to limit things for people in other professions, or other
people in general. I'm trying to give just one example of how I think it
would be beneficial to the world to think in terms that go beyond the
traditional pro audio profession. Audio technology is used for a lot more
than just music and voice, and those applications are important, too. I
used this example to point out that whales and dolphins talk and sing
too, and humans are only starting to appreciate the value of what they
talk and sing about. In the future, there may be many other things humans
become aware of and want to do with sound. Let's not limit ourselves.

Another quote:

"Sampling at 192KHz produces larger files requiring more storage space
and slowing down the transmission. Sampling at 192KHz produces a huge
burden on the computational processing speed requirements"

The copyright shown in the document is 2004, a full 10 years ago. The
statement was a lot more true then than it is now!

If not for economic and political limitations, by all rights in the USA
today we should commonly have at least 1 Gbit/sec Internet connections by
now, with more on the way. The technology is all there, and there's
plenty of speed above that.

Also, I read in the news recently that Western Digital announced a 10 TB
hard drive, and Seagate announced an 8 GB model. Drives of 1 TB or more
are now commonly available and inexpensive.

A stereo 24bit 96KHz album in flac format takes up only about 1 GB. That
is not a lot anymore.

If anyone wants to continue to argue that we should continue to use
smaller files for efficiency, I just don't want to hear it. It will just
remind me of when Bill Gates proclaimed that no one would ever need more
than 640KB of main memory. (I had already needed megabytes years before
that!)

Fortunately, the rest of the computer industry continued to develop. I
saw in one of the Pono videos where Neil Young was saying how our other
technology has advanced, but digital audio is stuck, and if anything
(because of mp3s) has gotten worse in the same time period. I don't agree
with everything Neil has to say, but I do agree that it's time to move
things forward.

I really don't see what anyone would have anything against the pono. If
you don't think you need it, can't you just ignore it? Even if the pono
takes over the market, it still plays mp3s and CD-quality files, or you
can save some money and use technology that supports only that. I'm
confident that CD and mp3 quality will continue to be available to
satisfy people who don't want anything more.

In general, even if the current 24/96 and 24/192 formats seem silly, why
not just allow them to exist and watch as people experiment with them to
see if they are good for anything practical? Even if they are not, the
explorations may later lead to something cool.

One final note about the Sampling Theory For Digital Audio article: I
have nothing against the Nyquist Theorem, pure mathematics, or how the
author explained it. My problem is with the limited thinking in the audio
industry, and too many times, I've seen that article used to support and
maintain that dogma. I think that is inappropriate; that's all. I think
my negative statements earlier were more about that than anything in the
article itself.

(I know this was long, but I realized that I wasn't putting enough time
and attention into my posts again, and I was messing up. I hope this
helps clarify my position and doesn't cause too much trouble.)

[Les Cargill[_4_]]

Jay Ts wrote:
On Thu, 11 Sep 2014 08:28:31 -0500, Neil Gould wrote:
snip

"Research shows that musical instruments may produce energy above 20 KHz,
but there is little sound energy at above 40KHz. Most microphones do not
pick up sound at much over 20KHz. Human hearing rarely exceeds 20KHz."

I find all of these statements offensive. Although true, they mislead the
reader into thinking that there is absolutely no reason to save anything
above 20 KHz, and the rest of the article continues off of that idea,
which is not even literally present in the above quote! If you read
carefully and think a little, you might realize that there *is* content
above 20 KHz, and that some people miss out on that if it is not present
in the recording.

The reason I'm offended by this is not that I think I can hear above 20
KHz. I know I can't. The best I've ever heard at a conscious level is
about 17 KHz, and that was a long time ago. I'm not as good today. But I
did meet someone who could hear up to 30 KHz in a very informal blind
test. In science, all it takes to disprove any theory is one
counterexample, and that incident was enough to change my thinking.


But the 20-20kHz limit is not a theory, it is a normative assumption.
It's a pretty well empirically teste3d normative assumption. Getting
beyond it will be a big project. Nobody's done that yet.

I'm really concerned that if a few people (it might be only 1%) have
exceptional hearing at a conscious level, then maybe many more have
subconscious senses in the 20KHz range that may be significant somehow.
I don't know of any method of scientific research yet developed that can
test for that, so my attitude is to remain conservative and wait
patiently for more study. I'm waiting for higher resolution realtime
brain scanning technology and other things that probably haven't been
invented yet, and until then, I don't put much faith in listening tests.
In the meantime, I think it's best to design for the possibility that the
current assumptions may be wrong, and include support for higher
frequencies.

I was watching videos with Rupert Neve recently, and in one he told a
story about a recording engineer who complained about a fault in just one
channel of the studio's Neve console. Mr. Neve went in to investigate, at
first found nothing, but then studied more and found a fault at about 50
KHz in that channel. After fixing the circuit, the engineer was
satisfied. This is more anecdotal "evidence" that proves nothing, but I
think it's significant, and I wish I knew more of what happened. I think
things like this should be followed up on more to discern more about
what's happening. For that to happen, people must get beyond the dogma
that people can't hear anything above 20 KHz.

Mr. Neve's response to the incident was to improve his circuit designs,
and extend the frequency response to 120 KHz (I'm looking at his current
designs to get that number, since I don't remember what he said in the
video). He did not just say, "Well, I can't hear beyond 12 KHz so I don't
care about that."


I'd expect a fault at 50KHz to be more likely an oscillation - too much
signal - rather than a deficit in the ultrasonic. Perhaps Mr. Neve
updated his test regime as well as his designs to account
for those ranges.

But that's just a conjecture - as you say, we kinda don't know.

There is absolutely nothing wrong with guardband, so long
as cost isn't an issue.

This to me shows a good attitude towards engineering, with a good
scientific attitude of never being sure about anything, and being open-
minded. I've seen specs on many high-end "pro audio" products with
extended high-frequency response, and it seems that Mr. Neve is not alone
in thinking it is important to do so.


This is true.

It really bothers me that if I have products with that quality, I may
have it cut back to a bandwidth of only 20 KHz, or even 40 KHz, at any
later stage in the signal chain, especially the last one at the recording.


I have lowpassed stuff @ 10Khz and listened to what's left. It isn't
pretty Of course, that's a goofy thing to do and for
all I know what I heard was artifacts from the filtering.

I really should translate that to a couple octaves down and see what it
sounds like.

The human ear does 1000Hz best for semi-physics reasons - we mainly
evolved hearing for .... predation and defense purposes, but then we
laid language on it.

Distorting and bandlimiting speech may improve intelligibility. So the
4-5 octaves above 1000 Hz are very likely about all we'll ever get.
Anything beyond that is likely an evolutiuonary process.

My understanding is that animals that hear beyond 20Khz don't have
what *we* would call *hearing* in that range. It's either "for"
echolocation or sometimes young animals pinging their mothers.

And that the dividing line is ( very curiously ) always 20 KHz.

snip
Years ago I learned of software-defined radio technology. A quote from
the Wikipedia article on this topic:

"Software-defined radio (SDR) is a radio communication system where
components that have been typically implemented in hardware (e.g. mixers,
filters, amplifiers, modulators/demodulators, detectors, etc.) are
instead implemented by means of software on a personal computer or
embedded system.[1] While the concept of SDR is not new, the rapidly
evolving capabilities of digital electronics render practical many
processes which used to be only theoretically possible."

At the time, I was simply amazed that someone thought of that idea. They
realized that computers had become fast enough that parts of the
electronic circuitry in radio systems could be replaced by software
driving the computer's sound card, even though the sound card was limited
to audio frequencies. This has allowed radio systems to be controlled in
new ways and do more things. If sound cards supported higher frequencies,
I wonder what more might be possible.


SDR isn't primarily about using soundcards. Indeed, there are
"Raspberry Pi" type SDR kits available now.

Because it's RF, you really want the physical frequency well out of the
audible range - but SFAIK, it's the IF that's in the "soundcard" range -
the signal is then translated - mixed - to an RF band.

I think the response of some people here is "Who cares? that is another
profession, and that's their problem. It has nothing to do with this." To
head that off, I think it is very unprofessional for anyone to be so
selfish as to limit things for people in other professions, or other
people in general. I'm trying to give just one example of how I think it
would be beneficial to the world to think in terms that go beyond the
traditional pro audio profession. Audio technology is used for a lot more
than just music and voice, and those applications are important, too. I
used this example to point out that whales and dolphins talk and sing
too, and humans are only starting to appreciate the value of what they
talk and sing about. In the future, there may be many other things humans
become aware of and want to do with sound. Let's not limit ourselves.


So what has happened is that hardware specialized to SDR is emerging.

Another quote:

"Sampling at 192KHz produces larger files requiring more storage space
and slowing down the transmission. Sampling at 192KHz produces a huge
burden on the computational processing speed requirements"

The copyright shown in the document is 2004, a full 10 years ago. The
statement was a lot more true then than it is now!


In a way.

If not for economic and political limitations, by all rights in the USA
today we should commonly have at least 1 Gbit/sec Internet connections by
now, with more on the way. The technology is all there, and there's
plenty of speed above that.


I don't think that's all that realistic, really. The cost shear between
100 mbit stuff and 1 GBit stuff is pretty profound - even if you have a
1 GBit NIC on your computer, you won't see sustained
throughput rates of 1 GBit on it and you probably can't afford the sort
of networking equipment that provides 1 GBit sustained throughput from
node to node.

A joke has been that it takes 1 GBit stuff to get 100 MBit performance.

You can get FIOS (maybe) and it has astounding speed. But that will be
more likely used to provide more and more channels.

Also, I read in the news recently that Western Digital announced a 10 TB
hard drive, and Seagate announced an 8 GB model. Drives of 1 TB or more
are now commonly available and inexpensive.


Yep.

A stereo 24bit 96KHz album in flac format takes up only about 1 GB. That
is not a lot anymore.


It's still a lot.

snip

In general, even if the current 24/96 and 24/192 formats seem silly, why
not just allow them to exist and watch as people experiment with them to
see if they are good for anything practical? Even if they are not, the
explorations may later lead to something cool.


I'd personally have no objections to it at all. I no longer
buy or even much listen to music.

The background to pono is pretty specific - Neil Young has this immense
.... "box set" he wants to release as a sort of musical
monument to ... his career ( which is fine ) and pono is
him trying to invent a box to put it in.

I am a Neil Young fan, but sometimes he does Doc Brown things, and this
looks like one. I find it endearing.
I'm not gonna laugh at a guy who's trying to be a bit of a renaissance
man. But his comments lead us to think maybe he's not that
grounded in the theory.

But God bless him for his efforts - like you say, who knows? But
the golden age of big music is probably past us. Once, Berry Gordy
was able to move Motown to LA to escape the trap of singles, but ...
they're baaaack.

One final note about the Sampling Theory For Digital Audio article: I
have nothing against the Nyquist Theorem, pure mathematics, or how the
author explained it. My problem is with the limited thinking in the audio
industry, and too many times, I've seen that article used to support and
maintain that dogma. I think that is inappropriate; that's all. I think
my negative statements earlier were more about that than anything in the
article itself.


So 20KHz is not an absolute limit. It's a normative limit. So ... yep.

I will be unlikely to ever buy it.

(I know this was long, but I realized that I wasn't putting enough time
and attention into my posts again, and I was messing up. I hope this
helps clarify my position and doesn't cause too much trouble.)


--
Les Cargill

[PStamler]

On Thursday, September 11, 2014 7:09:47 PM UTC-6, Jay Ts wrote:
I was watching videos with Rupert Neve recently, and in one he told a
story about a recording engineer who complained about a fault in just one
channel of the studio's Neve console. Mr. Neve went in to investigate, at
first found nothing, but then studied more and found a fault at about 50
KHz in that channel. After fixing the circuit, the engineer was
satisfied. This is more anecdotal "evidence" that proves nothing, but I
think it's significant, and I wish I knew more of what happened.

The engineer was Geoff Emerick.

I've met both Mr. Emerick and Mr. Neve, and can round out the story a little. Neve (the company) had just delivered a new console, and Mr. Emerick heard one of the channels sounding different from the others. Mr. Neve pulled the module, and found that a termination resistor for one of the transformers had been left out, producing a resonance around 50kHz. He installed the resistor, and the channel worked right.

But that doesn't prove that Mr. Emerick can hear things happening at 50kHz. Leaving out that terminating resistor would very likely disrupt the channel's frequency response below 20kHz, and would certainly cause that channel's phase response to be different from that of a properly-made channel, probably well into the audio range.

So the event doesn't necessarily prove that Mr. Emerick (or other people) can hear things happening around 50kHz -- not when those things can also affect response within what is conventionally considered the audio band. What it does show, I suspect, is that Geoff Emerick has really excellent abilities for hearing things within the audio band.

But we knew that...

Peace,
Paul