[Arny Krueger]

http://www.lavryengineering.com/docu...ing_Theory.pdf


"
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 ideas
is 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.
"

[TT]

"Arny Krueger" wrote in message
. ..

http://www.lavryengineering.com/docu...ing_Theory.pdf


"
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 ideas
is 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.
"


So with SACDs 2.8MHz sampling and 1 bit resolution doesn't
that therefore overcome this problem with PCM? It would
have been great if this article had of touched on that.

Regards TT

[Arny Krueger]

"TT" wrote in message

"Arny Krueger" wrote in message
. ..

http://www.lavryengineering.com/docu...ing_Theory.pdf


"
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 ideas
is 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.
"


So with SACDs 2.8MHz sampling and 1 bit resolution doesn't
that therefore overcome this problem with PCM?

Doesn't seem like SACD makes much of a difference in this regard. Remember
that the basic message is that 96 KHz sampling is already more than enough
for the best possible sounding audio.

It would have been great if this article had of touched on that.

[TT]

"Arny Krueger" wrote in message
...

: So with SACDs 2.8MHz sampling and 1 bit resolution
doesn't
: that therefore overcome this problem with PCM?
:
: Doesn't seem like SACD makes much of a difference in this
regard. Remember
: that the basic message is that 96 KHz sampling is already
more than enough
: for the best possible sounding audio.
:
From my limited understanding I gained the impression that
192kHz was losing the 24 bit resolution.

Quote: "There is a tradeoff between speed and accuracy" and
"Sampling audio signals at 192KHz is about 3 times faster
than the optimal rate. It compromises the accuracy which
ends up as audio distortions."

So surely if higher sample rates lose bits then *IF* you
only have 1 bit to start with it would be very hard to lose
it. Or if you did then that would be very sad indeed ;-)

I read this as higher sample rates are good *if* you had the
processing power not to lose bits. So like I said I would
have been more interested in the comparison with SACD/DSD as
it would seem it overcomes the problems as presented in the
article.

BTW to quote from the above again "Sampling audio signals at
192KHz is about 3 times faster than the optimal rate" so it
would appear he says the optimal rate is approx 64kHz. So
where does that leave your 44.1kHz which is perfect in your
opinion? I believe I could live with 64/24 CDs quite nicely
;-)

Regards TT

[eric]

44.1 is clearly inadequate. The harsh treble overtone structures many
listeners report from CD vis-a-vis vinyl and analog tape are more than
figments of their imaginations: they are almost certainly artifacts of
the necessity of having more bandwidth than the signal can occupy. The
oscilloscope community figured that out in the 40s and many in the
audio field-Neve et al- have demonstrated it over and over. Yet, Arny
isn't listening.

I think that it is more the 16 bits that is inadaquate. The dynamic range
this presents covers basic listening requirements if everything in the
mastering chain is done perfectly. 24 would be much better.

As far as 44.1, I think that these 'harsh treble overtones' are not due to
any flaw in the basic specification. They are either due to the rolloffs
that occur in analog reproduction wear and tear making users un-used to
hearing flat reproduction, or poor implementations of anti-alias filters.

MAYBE there is an advantage to going to 48 or 50 KHz, but anything more is
gross overkill.

The analogy to oscilloscopes is, to anyone who has owned and used
oscilloscopes, hogwash. So long as your flat frequency response covers
the range of interest there is no problem. If there is a requirement
to have a higher bandwidth scope than the signal you are measuring it
arises from the usual practice of scope manufacturer specifying the
frequency range at the -3 db response point. If you are working with 20
MHz signals a 20 MHz scope (down 3db at 20 MHz) is not going to be
satisfactory.

[Eeyore]

Bret Ludwig wrote:

We now can and should do better. And, we have, if we will but use it.

I'm sure it's no accident that many top recording studios use 24/96 and now 192
as well.

Graham

[TT]

"Eeyore" wrote in
message ...
:
:
: Bret Ludwig wrote:
:
: We now can and should do better. And, we have, if we
will but use it.
:
: I'm sure it's no accident that many top recording studios
use 24/96 and now 192
: as well.
:
: Graham
:
I understood they have been using 32 bit for some time now.
So it would be 32/96 or 32/192.

Regards TT

[dizzy]

Eeyore wrote:

Bret Ludwig wrote:

We now can and should do better. And, we have, if we will but use it.

I'm sure it's no accident that many top recording studios use 24/96 and now 192
as well.

And that's no doubt useful, in the studio. In the music-delivery
system, the silver disk, CD sounds like good enough.

[Arny Krueger]

"TT" wrote in message

"Arny Krueger" wrote in message
...

So with SACDs 2.8MHz sampling and 1 bit resolution
doesn't that therefore overcome this problem with PCM?

Doesn't seem like SACD makes much of a difference in
this regard. Remember that the basic message is that 96
KHz sampling is already more than enough for the best
possible sounding audio.

From my limited understanding I gained the impression that
192kHz was losing the 24 bit resolution.

In fact there are no practical converters operating at any sample rate that
would be appropriate for audio, that also deliver true 24 bit resolution.

The way I interpret Lavry's statement is that all other things being equal,
operations at 192 KHz will be signficicantly degraded compared to operation
at about 1/3 that rate.

Quote: "There is a tradeoff between speed and accuracy"
and "Sampling audio signals at 192KHz is about 3 times
faster than the optimal rate. It compromises the accuracy
which ends up as audio distortions."

IOW a given converter that operates at 192 KHz will not have the high
resolution it has when operating in the 64 KHz range.

So surely if higher sample rates lose bits then *IF* you
only have 1 bit to start with it would be very hard to
lose it. Or if you did then that would be very sad
indeed ;-)

Don't confuse the terminology "1 bit converter" with the effective
resolution of the converter being in the range from 14 to 20 bits for audio.
The terminology "1 bit converter" related to some internal operational
details.

I read this as higher sample rates are good *if* you had
the processing power not to lose bits.

It is not really about processing power as much as it is about the
effectiveness of various elements of the converter itself. 1 bit convertors
work with pulses. As the sample rate goes up, elements of the converter lose
accuracy, and the pulses start getting a little mangled and prone to being
slightly misinterpreted.

So like I said I
would have been more interested in the comparison with
SACD/DSD as it would seem it overcomes the problems as
presented in the article.

Unlikely. SACD uses some of the same kinds of circuit elements as a
so-called "1 Bit" ADC or SACD. As the SACD converter treis to run faster and
faster, these same circuit elements also lose accuracy in a similar fashion
as they do inside the 1-bit converter.

BTW to quote from the above again "Sampling audio signals
at 192KHz is about 3 times faster than the optimal rate"
so it would appear he says the optimal rate is approx
64kHz. So where does that leave your 44.1kHz which is
perfect in your opinion? I believe I could live with
64/24 CDs quite nicely ;-)

It is a fact that converter accuracy and price/performance are no longer the
stumbling blocks to sound quality that they once were.

There is no reliable evidence that the 16/44 data format is a stumbling
block to the sonically-accurate reproduction of music thqat is distributed
to end-users. However, not all of the market that Lavry sells to is
sufficiently aware of this. Lavry's problem is that some of the people in
the market he serves, think that very high sample rates have a practical
justification.

[Eeyore]

Arny Krueger wrote:

"TT" wrote in message

From my limited understanding I gained the impression that
192kHz was losing the 24 bit resolution.

In fact there are no practical converters operating at any sample rate that
would be appropriate for audio, that also deliver true 24 bit resolution.

None deliver true 24 bit for sure. The reason for 24 bit converters is to ensure
that the bits 'really doing the work' are accurate. 20 accurate bits is hunky
dory.

Older ( 16 bit ) converters typically had serious non-linearity problems with
the bottom few bits which were clearly audible ( and measurable ).


Graham

[Eeyore]

Bret Ludwig wrote:


BTW to quote from the above again "Sampling audio signals at
192KHz is about 3 times faster than the optimal rate" so it
would appear he says the optimal rate is approx 64kHz. So
where does that leave your 44.1kHz which is perfect in your
opinion? I believe I could live with 64/24 CDs quite nicely

64 would put the Nyquist limit at 30 kHz and that IMO would be a big
improvement over CD, and in fact, 96 really is probably enough. The
primary point is that 44 is NOT.

44.1 is at best *questionable*.

It's a real shame they didn't choose 48 since that would have made such
discussion much more academic.

Graham

[Arny Krueger]

"Bret Ludwig" wrote in message
oups.com
It's crap, because the premise fails to answer the
question, "compared to what"?

As bit throughput, storage space, and processing power
steadily increase each year, 192 kHz goes from being an
onerous requirement requiring great sacrifice to
something more and more trivially handled.

But, it serves no purpose. It distracts people from far more important
issues.

While I
suspect it is indeed way more than is actually required,
the downside, once serious, is now less and less so.

So what? Following your logic, I need to have my car upgraded to over 1,000
horsepower as compared to its current 225 horsepower, because the cost of
upgrading to over 1,000 horsepower is not as prohibitive as it once was.

44.1 is clearly inadequate.

Assertion without proof or even supporting evidence.

The harsh treble overtone
structures many listeners report from CD vis-a-vis vinyl
and analog tape are more than figments of their
imaginations: they are almost certainly artifacts of the
necessity of having more bandwidth than the signal can
occupy

No bias-controlled listening tests confirm this. It is well-known that
people's biases can cause them to perceive problems that don't really exist.

. The oscilloscope community figured that out in
the 40s and many in the audio field-Neve et al- have
demonstrated it over and over.

Neve demonstrated no such thing. If you understand what Neve said, he
basically said that circuitry that resonates at say 40 KHz can have audible
effects below 15 KHz. If you look at the corresponding frequency response
curve you see that his circuit components such as input transformers did
indeed have effects on the order a few dB below 15 KHz even though they were
resonating at several times that frequency. This is just the well-known
behavior of resonant circuits.

Yet, Arny isn't listening.

Bret apparently did not pay attention to his sophomore electrical circuits
class that covered resonant circuits, if he ever actually even took such a
class. Or maybe he can't apply what he learned to practical audio circuits.

Those CDs that sound the best are usually those of
material from a time where the treble cutoff was 10 kHz
or less, functionally.

No such thing. In fact high-sample-rate material (24/96) with strong
harmonics right up to 20 KHz are audibly unchanged by a proper job of
downsampling to 44 KHz, and even lower.

This should tell us something too.
Unless, like Arny, we are quite literally not listening.

Obviously Bret you are listening to what I say, and quite irritated by it.
Too bad you can't rise to the occasion and share some wise words.

The CD was a serious compromise made in the early 80s to
put all of Beethoven's Ninth Symphony on one single-sided
optical disc easily producible at then-current technology
at a diameter a drive accomodating it could fit in a 5
1/4" floppy drive bay. And, in all fairness, it could
have been worse-a lot worse. But to uphold it as the gold
standard is idiocy.

Show us your bias-controlled listening tests that support your claims, Bret.
My friends and I did our homework. We subjected high-quality musical signals
from live performances to 16/44 coding, in one of the finest studios in the
Detroit area, which was under the direction of Robert Dennis who is still
working professionally to this day. We used over a dozen musicians, audio
engineers, and experienced audiophiles as our listening panel. No
distinguishable differences were found.

We now can and should do better. And, we have, if we will but use it.

It is true that I have dozens of channels of converters that are capable of
running at 24/96 and 24/192. I've used them to record music from broadband
sources and compared the results to what happens when the signal is further
downsampled to 16/44. No audible difference for either myself or my friends.

Anybody with high sample rate converters, who wants to listen to examples of
this issue being played out with broadband musical sounds can do so by
downloading files from http://www.pcabx.com/technical/sample_rates/index.htm
..

[Alan S]

"Arny Krueger" wrote in message
. ..
"Bret Ludwig" wrote in message
oups.com
It's crap, because the premise fails to answer the
question, "compared to what"?

As bit throughput, storage space, and processing power
steadily increase each year, 192 kHz goes from being an
onerous requirement requiring great sacrifice to
something more and more trivially handled.

But, it serves no purpose. It distracts people from far more important
issues.

While I
suspect it is indeed way more than is actually required,
the downside, once serious, is now less and less so.

So what? Following your logic, I need to have my car upgraded to over
1,000 horsepower as compared to its current 225 horsepower, because the
cost of upgrading to over 1,000 horsepower is not as prohibitive as it
once was.

44.1 is clearly inadequate.

Assertion without proof or even supporting evidence.


My ears are evidence enough. I agree that 192 kHZ is overkill for a sampling
rate and it would just complicate an already complicated process but 44.1
kHZ at 16 bit by nature requires that a lot of information gets left out
when dithering down. Many people can hear it clearly, especially those of us
that remember 2" tape to vinyl.


The harsh treble overtone
structures many listeners report from CD vis-a-vis vinyl
and analog tape are more than figments of their
imaginations: they are almost certainly artifacts of the
necessity of having more bandwidth than the signal can
occupy

No bias-controlled listening tests confirm this. It is well-known that
people's biases can cause them to perceive problems that don't really
exist.


I hear that! (no pun intended) That's why I never mix-down with cans. If I
mix-down with a great set of headphones, it takes me twice as long because I
always hear stuff that's not there.


. The oscilloscope community figured that out in
the 40s and many in the audio field-Neve et al- have
demonstrated it over and over.

Neve demonstrated no such thing. If you understand what Neve said, he
basically said that circuitry that resonates at say 40 KHz can have
audible effects below 15 KHz. If you look at the corresponding frequency
response curve you see that his circuit components such as input
transformers did indeed have effects on the order a few dB below 15 KHz
even though they were resonating at several times that frequency. This is
just the well-known behavior of resonant circuits.

Yet, Arny isn't listening.

Bret apparently did not pay attention to his sophomore electrical circuits
class that covered resonant circuits, if he ever actually even took such a
class. Or maybe he can't apply what he learned to practical audio
circuits.

Those CDs that sound the best are usually those of
material from a time where the treble cutoff was 10 kHz
or less, functionally.

No such thing. In fact high-sample-rate material (24/96) with strong
harmonics right up to 20 KHz are audibly unchanged by a proper job of
downsampling to 44 KHz, and even lower.


This is a very important aspect that this thread that hasn't had much
address. An engineer can use the best algorythm in the world and at the end
of the day they are still tossing information in the garbage. I was
recording at 16/44 for a while simply because I figured if it was going to
get dithered down to that in the end I might as well get all the information
I could on the front side. The truth is that for whatever reason, it sounds
better to me if I record at 32/96 and dither down, though the average
listener doesn't usually notice. The CD is an unfortunate example of the
dumbing down of our society and until there is a marketable improvement in
technology, it will remain. There are a lot of people out there that have
never heard amazing music and likely never will simply because they are
never around it. They are quite happy with their iPods.


This should tell us something too.
Unless, like Arny, we are quite literally not listening.

Obviously Bret you are listening to what I say, and quite irritated by it.
Too bad you can't rise to the occasion and share some wise words.

The CD was a serious compromise made in the early 80s to
put all of Beethoven's Ninth Symphony on one single-sided
optical disc easily producible at then-current technology
at a diameter a drive accomodating it could fit in a 5
1/4" floppy drive bay. And, in all fairness, it could
have been worse-a lot worse. But to uphold it as the gold
standard is idiocy.

Show us your bias-controlled listening tests that support your claims,
Bret. My friends and I did our homework. We subjected high-quality musical
signals from live performances to 16/44 coding, in one of the finest
studios in the Detroit area, which was under the direction of Robert
Dennis who is still working professionally to this day. We used over a
dozen musicians, audio engineers, and experienced audiophiles as our
listening panel. No distinguishable differences were found.

We now can and should do better. And, we have, if we will but use it.

It is true that I have dozens of channels of converters that are capable
of running at 24/96 and 24/192. I've used them to record music from
broadband sources and compared the results to what happens when the signal
is further downsampled to 16/44. No audible difference for either myself
or my friends.

Anybody with high sample rate converters, who wants to listen to examples
of this issue being played out with broadband musical sounds can do so by
downloading files from
http://www.pcabx.com/technical/sample_rates/index.htm .

[Arny Krueger]

"Bret Ludwig" wrote in message
ups.com
BTW to quote from the above again "Sampling audio
signals at 192KHz is about 3 times faster than the
optimal rate" so it would appear he says the optimal
rate is approx 64kHz. So where does that leave your
44.1kHz which is perfect in your opinion? I believe I
could live with 64/24 CDs quite nicely


64 would put the Nyquist limit at 30 kHz and that IMO
would be a big improvement over CD, and in fact, 96
really is probably enough. The primary point is that 44
is NOT.

Baseless assertions are just cheap shots.

[Harry Lavo]

"Bret Ludwig" wrote in message
ups.com...


BTW to quote from the above again "Sampling audio signals at
192KHz is about 3 times faster than the optimal rate" so it
would appear he says the optimal rate is approx 64kHz. So
where does that leave your 44.1kHz which is perfect in your
opinion? I believe I could live with 64/24 CDs quite nicely


64 would put the Nyquist limit at 30 kHz and that IMO would be a big
improvement over CD, and in fact, 96 really is probably enough. The
primary point is that 44 is NOT.


64 was actually settled upon by the DVD-A consortium in Japan in the
early-mid '90's as sufficient. Whether they did any actual testing to
support this I don't know.

[Powell]

"Arny Krueger" wrote

While there is no up side to operation at excessive speeds, there are
further disadvantages:

In a short news article Marantz reported increases in
fidelity up to 500 KHz... a 260% increase over 192
MHz.

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

I recently read about a studio in England that discarded
master tapes into a garbage bin due to a lack of physical
storage, without giving notice to owners... but they were
scarfed up by someone in the public.

Certainly storing information (music) digitally, regardless
of the size, is far more cost effective than tape.


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.

Those are only momentary technological problems. I guess
your new video production interests won't include HD
then.

Speaking of "processing power," the current fastest computer
in the world (IBM's BlueGene) will be leapfrogged with a new
super computer (Roadrunner). It uses 16,000 cell possessors
(used in Playstation 3) and 16,000 AMD Opteron
microprocessors achieving 1.6 petaflops (1,600 trillion)
calculations per second. And the operating system you ask,
Linux of course.


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 ideas is based on the fallacy that faster
rates yield more accuracy and/or more detail.
Physics theory would suggest that sampling rates are only
excessive if they exceed the energy pack rate in which sound
travels through the air. At its most fundamental level all analog
sound is digital/packet. IMO, what's needed is a new algorithm
for trans-coding analog to digital using these even higher rates.


Weather motivated by profit or ignorance, the promoters, leading the
industry in the wrong direction, are stating the opposite of what is true.

You have never been an early user (empirical knowledge) of
new technologies, Arny.

[Arny Krueger]

"Powell" wrote in message

"Arny Krueger" wrote

While there is no up side to operation at excessive
speeds, there are further disadvantages:

In a short news article Marantz reported increases in
fidelity up to 500 KHz... a 260% increase over 192
MHz.

Yes, there is an apocryphal story about Ken Ishiwata from Marantz blowing
that kind of smoke. Of course KI is well-known for his publicity and
marketing, not his knowlege of audio technology.


snip quacking

[Powell]

"Arny Krueger" wrote

While there is no up side to operation at excessive
speeds, there are further disadvantages:

In a short news article Marantz reported increases in
fidelity up to 500 KHz... a 260% increase over 192
MHz.

Yes, there is an apocryphal story about Ken Ishiwata from
Marantz blowing that kind of smoke. Of course KI is
well-known for his publicity and marketing, not his knowlege
of audio technology.

Zzzzzz....


snip quacking

I'm enjoying my new found abilities which cause
you to run away from my threads. So you see me
as the bogeyman nowadays. Hehehe... nice hold
over you, to-be-sure.