No Difference Between High-Resolution Audio and CDs

[JackA]

"In double-blind tests conducted by Levitin and others -- some results of which were published in the Journal of the Audio Engineering Society -- listeners cannot tell the difference between high-resolution audio and CD-quality audio".

"But many audio professionals, including Bob Ludwig and NYU's Jim Anderson, say they can hear an improvement over CD quality, and they prefer the higher frequencies and sample rates. Anderson even teaches a class at NYU in which he instructs students on how to listen for the differences".

That's right, you have to be trained to hear the difference!

Not sure why Bob Ludwig is mentioned. I mean, boy, his John Cougar CD was LOUD!!

Jack
whoops...
http://www.laweekly.com/music/why-cds-may-actually-sound-better-than-vinyl-5352162

[Scott Dorsey]

JackA > wrote:
> "In double-blind tests conducted by Levitin and others -- some results of =
>which were published in the Journal of the Audio Engineering Society -- lis=
>teners cannot tell the difference between high-resolution audio and CD-qual=
>ity audio".

I suggest you read the paper referenced. It's very interesting, and the
procedure isn't flawless but it's pretty good. It was discussed extensively
here when it first came out.

I should point out that it has nothing to do with the subject of why CD and
vinyl pressings sound different; you're not going to get any real ultrasonics
on a vinyl disc either. In fact, you won't even get 18kc anywhere near
reference level and still get a disc anyone can track; the system is
slew-limited.
--scott

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

[Randy Yates[_2_]]

JackA > writes:

> "In double-blind tests conducted by Levitin and others -- some results of which were published in the Journal of the Audio Engineering Society -- listeners cannot tell the difference between high-resolution audio and CD-quality audio".
>
> "But many audio professionals, including Bob Ludwig and NYU's Jim Anderson, say they can hear an improvement over CD quality, and they prefer the higher frequencies and sample rates. Anderson even teaches a class at NYU in which he instructs students on how to listen for the differences".
>
> That's right, you have to be trained to hear the difference!

Do you think the AES results did not use "trained" listeners?

This is the typical pro audio / sales / marketing BS to sell another
round of media (CD/DVD-Audio/SACD/... on and on).
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

[[email protected]]

9:31 AM Randy Yates wrote:
"JackA > writes:

> "In double-blind tests conducted by Levitin and others -- some results of which were published in the Journal of the Audio Engineering Society -- listeners cannot tell the difference between high-resolution audio and CD-quality audio".
>
> "But many audio professionals, including Bob Ludwig and NYU's Jim Anderson, say they can hear an improvement over CD quality, and they prefer the higher frequencies and sample rates. Anderson even teaches a class at NYU in which he instructs students on how to listen for the differences".
>
> That's right, you have to be trained to hear the difference!

Do you think the AES results did not use "trained" listeners?

This is the typical pro audio / sales / marketing BS to sell another
round of media (CD/DVD-Audio/SACD/... on and on).
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com "



THANK YOU Randy!! And Scott Dorsey may it serve
you well.


Of course High-Res can sound different than traditional
lossless/CD: Just EQ or otherwise process the master
for high-res DIFFERENTLY than how you did so for
lossless! duhhh..

Then charge US$10 more for it than you would for the
CD version. smh.. Goes on daily, though no one has
the guts to admit it.

[[email protected]]

>
> Of course High-Res can sound different than traditional
> lossless/CD: Just EQ or otherwise process the master
> for high-res DIFFERENTLY than how you did so for
> lossless! duhhh..
>
> Then charge US$10 more for it than you would for the
> CD version. smh.. Goes on daily, though no one has
> the guts to admit it.

there is a lot of hype here to be sure but there is at least one nugget of truth

conventional CDs require 20 kHz low pass filters for anti-aliasing and
reconstruction. Oversampling does not change that, it simply allows th filters to be implementedin the digital domain, but the filters are still needed.

Steep low pass filters __can__ produce ringing around 20 kHz that __may__ be audible to some people. (not me)

It takes the right (wrong) combination of source material and mics and speakers and ears to be able to hear this.

Even phase linear filters digital can do this.

Anytime there is a steep change in the ampliture response, artifacts can be created.

In practice, I don't think this is a real issue, but in theory it can be.


Mark

[JackA]

On Wednesday, March 18, 2015 at 12:27:39 PM UTC-4, wrote:
> >
> > Of course High-Res can sound different than traditional
> > lossless/CD: Just EQ or otherwise process the master
> > for high-res DIFFERENTLY than how you did so for
> > lossless! duhhh..
> >
> > Then charge US$10 more for it than you would for the
> > CD version. smh.. Goes on daily, though no one has
> > the guts to admit it.
>
> there is a lot of hype here to be sure but there is at least one nugget of truth
>
> conventional CDs require 20 kHz low pass filters for anti-aliasing and
> reconstruction. Oversampling does not change that, it simply allows th filters to be implementedin the digital domain, but the filters are still needed.
>
> Steep low pass filters __can__ produce ringing around 20 kHz that __may__ be audible to some people. (not me)
>
> It takes the right (wrong) combination of source material and mics and speakers and ears to be able to hear this.
>
> Even phase linear filters digital can do this.
>
> Anytime there is a steep change in the ampliture response, artifacts can be created.
>
> In practice, I don't think this is a real issue, but in theory it can be.
>
>
> Mark

Mark, on YouTube, someone basically proved what you wrote.

Jack

[geoff]

On 19/03/2015 11:42 a.m., JackA wrote:

>
> Mark, on YouTube, someone basically proved what you wrote.

Through Youtube-massacred audio ?!!!

geoff

[JackA]

On Thursday, March 19, 2015 at 4:15:31 AM UTC-4, geoff wrote:
> On 19/03/2015 11:42 a.m., JackA wrote:
>
> >
> > Mark, on YouTube, someone basically proved what you wrote.
>
> Through Youtube-massacred audio ?!!!

Oh, silly, it's probably because your current headphones distort and this, mere $9,534 ear bud, would sound so much better!...

http://www.amazon.com/OneGood-Stereo---mono-Isolating-fabric-wrapped/dp/B004F0BK5U/ref=sr_1_3?s=electronics&ie=UTF8&qid=1426767157&sr=1-3&keywords=earbuds

Even one reviewer is deaf in one ear, but can hear stereo again with this amazing ear bud!! Noise cancelling, too, would work well here!!

Jack :)

>
> geoff

[BBQ]

On Tue, 17 Mar 2015 19:06:29 -0700 (PDT), JackA
> wrote:

> "In double-blind tests conducted by Levitin and others -- some results of which were published in the Journal of the Audio Engineering Society -- listeners cannot tell the difference between high-resolution audio and CD-quality audio".
>
>"But many audio professionals, including Bob Ludwig and NYU's Jim Anderson, say they can hear an improvement over CD quality, and they prefer the higher frequencies and sample rates. Anderson even teaches a class at NYU in which he instructs students on how to listen for the differences".
>
>That's right, you have to be trained to hear the difference!
>
>Not sure why Bob Ludwig is mentioned. I mean, boy, his John Cougar CD was LOUD!!
>
>Jack
>whoops...
>http://www.laweekly.com/music/why-cds-may-actually-sound-better-than-vinyl-5352162

Indeed high-resolution audio is wonderful!

Tested in a friend's studio and it's a lot different. Exactly, highs
are really brilliant, and there's something about harmonics but I
can't tell you exactly.

while he was playing a synth I could feel the sound was totally
different, like more detailed.

It's like when you play an organ on a keyboard and then you hear the
real cathedral organ in front of you.
--

http://GasGrills.biz | The Best Grills Website for the Serious Griller!

[Scott Dorsey]

BBQ > wrote:
>Indeed high-resolution audio is wonderful!
>
>Tested in a friend's studio and it's a lot different. Exactly, highs
>are really brilliant, and there's something about harmonics but I
>can't tell you exactly.
>
>while he was playing a synth I could feel the sound was totally
>different, like more detailed.

Usually when I hear something like this, I start suspecting that there is
a source of high-order even harmonic distortion going on.

Not that this is necessarily the issue, but if you're hearing a difference
due to bandwidth extension, it's probably a good idea to look and figure
out why.

It _is_ worth looking up the JAES paper being referenced. It's a decently
conducted experiment. It's not perfect, but it's done well.

Back in the nineties the difference between 44.1 and 48 ksamp/sec on the
SV3700 was very obvious. Turns out.... there's a little bitty problem
with the anti-aliasing filter design...
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Luxey]

When you say "high-resolution audio" what exactly you speak about, what is
The Resolution? Also, what did you compare it to?

[BBQ]

On Sat, 21 Mar 2015 18:02:06 -0700 (PDT), Luxey >
wrote:

>When you say "high-resolution audio" what exactly you speak about, what is
>The Resolution? Also, what did you compare it to?

ah, 24bit output (but I think it was 32 processed anyway) 96khz and
192khz..... we tested both.

the difference between 96 and 192 is.... well, hard to explain, it's
more feelings.

while from 44khz the difference is very clear and as I explained in
the other message.
--

http://GasGrills.biz | The Best Grills Website for the Serious Griller!

[Mike Rivers[_2_]]

On 3/21/2015 7:55 PM, BBQ wrote:
> Indeed high-resolution audio is wonderful!
>
> Tested in a friend's studio and it's a lot different. Exactly, highs
> are really brilliant, and there's something about harmonics but I
> can't tell you exactly.

Different master, so it's going to sound different. There are a lot of
things that you can do to improve the surface quality of a 25 year old
recording. Stuff just works better today, and given the challenge to
"make it sound better to people who are paying extra for the high
resolution version" and they'll make it sound better, by gosh.

If you want a real comparison of how it sounds better, dig up a 25 year
old A/D converter and feed something to it and a modern converter at the
same sample rate, then listen to the difference.

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

[Mike Rivers[_2_]]

On 3/21/2015 9:55 PM, BBQ wrote:
> the difference between 96 and 192 is.... well, hard to explain, it's
> more feelings.

Yeah, the singers just sing better when they're being recorded at 192
kHz. It's been demonstrated that in some instances, what goes into our
ear above 20 kHz can make a difference in how we perceive what's within
the normal hearing range, so a recording that starts out at 2x sample
rate will preserve whatever makes it though the microphone. But
re-sampling a 44.1 kHz file to 192 kHz doesn't put back anything that
was cut off by the anti-alias filtering. It only make the D/A converter
work better.

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

[Peter Larsen[_3_]]

"BBQ" > skrev i en meddelelse
...

> On Sat, 21 Mar 2015 18:02:06 -0700 (PDT), Luxey >
> wrote:

>>When you say "high-resolution audio" what exactly you speak about, what is
>>The Resolution? Also, what did you compare it to?

> ah, 24bit output (but I think it was 32 processed anyway) 96khz and
> 192khz..... we tested both.

> the difference between 96 and 192 is.... well, hard to explain, it's
> more feelings.

The difference you are listening to is how the available DA converter does
it. It may not be a format difference per se.

> while from 44khz the difference is very clear and as I explained in
> the other message.

What other message, the one trying to proove something about audio quality
via a link to a mp3-file? - didn't bother to read it.

And yes, I have heard high sample rate audio sound better than now sample
rate audio, and the difference was easy to define, which is why I wonder
about the validity of your observation.

Kind regards

Peter Larsen

[BBQ]

On Sun, 22 Mar 2015 09:02:00 +0100, "Peter Larsen"
> wrote:

>"BBQ" > skrev i en meddelelse
...
>
>> On Sat, 21 Mar 2015 18:02:06 -0700 (PDT), Luxey >
>> wrote:
>
>>>When you say "high-resolution audio" what exactly you speak about, what is
>>>The Resolution? Also, what did you compare it to?
>
>> ah, 24bit output (but I think it was 32 processed anyway) 96khz and
>> 192khz..... we tested both.
>
>> the difference between 96 and 192 is.... well, hard to explain, it's
>> more feelings.
>
>The difference you are listening to is how the available DA converter does
>it. It may not be a format difference per se.
>
>> while from 44khz the difference is very clear and as I explained in
>> the other message.
>
>What other message, the one trying to proove something about audio quality
>via a link to a mp3-file? - didn't bother to read it.
>
>And yes, I have heard high sample rate audio sound better than now sample
>rate audio, and the difference was easy to define, which is why I wonder
>about the validity of your observation.
>
> Kind regards
>
> Peter Larsen

Hello Peter, no I didn't include a link to an mp3. It was another one
not me.

We tested the rates with synthesizers. It was not a resampled sound.
It would have been useless that way.
--

http://GasGrills.biz | The Best Grills Website for the Serious Griller!

[BBQ]

On Sat, 21 Mar 2015 23:17:35 -0400, Mike Rivers >
wrote:

>On 3/21/2015 9:55 PM, BBQ wrote:
>> the difference between 96 and 192 is.... well, hard to explain, it's
>> more feelings.
>
>Yeah, the singers just sing better when they're being recorded at 192
>kHz. It's been demonstrated that in some instances, what goes into our
>ear above 20 kHz can make a difference in how we perceive what's within
>the normal hearing range, so a recording that starts out at 2x sample
>rate will preserve whatever makes it though the microphone. But
>re-sampling a 44.1 kHz file to 192 kHz doesn't put back anything that
>was cut off by the anti-alias filtering. It only make the D/A converter
>work better.

Ok Mike, but we tested with synthesizers working with different rates
and the difference between 44 and the higher rates is very clear.

Probably re-sampling a 44.1 kHz file to 192 kHz would have been
useless.

But oscillators producing real time sounds change a lot at different
rates.
--

http://GasGrills.biz | The Best Grills Website for the Serious Griller!

[BBQ]

On Sat, 21 Mar 2015 23:13:03 -0400, Mike Rivers >
wrote:

>On 3/21/2015 7:55 PM, BBQ wrote:
>> Indeed high-resolution audio is wonderful!
>>
>> Tested in a friend's studio and it's a lot different. Exactly, highs
>> are really brilliant, and there's something about harmonics but I
>> can't tell you exactly.
>
>Different master, so it's going to sound different. There are a lot of
>things that you can do to improve the surface quality of a 25 year old
>recording. Stuff just works better today, and given the challenge to
>"make it sound better to people who are paying extra for the high
>resolution version" and they'll make it sound better, by gosh.
>
>If you want a real comparison of how it sounds better, dig up a 25 year
>old A/D converter and feed something to it and a modern converter at the
>same sample rate, then listen to the difference.

I agree, anyway I think that the resolution can make a difference if
there is a new recording made exactly for that resolution not a
resampling.

And I always believe that 24bit and 96Khz to be the right balance and
the correct goal for the industry in the next years.
--

http://GasGrills.biz | The Best Grills Website for the Serious Griller!

[Scott Dorsey]

BBQ > wrote:
>On Sat, 21 Mar 2015 23:17:35 -0400, Mike Rivers >
>wrote:
>
>>On 3/21/2015 9:55 PM, BBQ wrote:
>>> the difference between 96 and 192 is.... well, hard to explain, it's
>>> more feelings.
>>
>>Yeah, the singers just sing better when they're being recorded at 192
>>kHz. It's been demonstrated that in some instances, what goes into our
>>ear above 20 kHz can make a difference in how we perceive what's within
>>the normal hearing range, so a recording that starts out at 2x sample
>>rate will preserve whatever makes it though the microphone. But
>>re-sampling a 44.1 kHz file to 192 kHz doesn't put back anything that
>>was cut off by the anti-alias filtering. It only make the D/A converter
>>work better.
>
>Ok Mike, but we tested with synthesizers working with different rates
>and the difference between 44 and the higher rates is very clear.

I am sure it -was- very clear, but your job now is to find out why, because
whenever something that should be a subtle difference turns into a very
obvious one, it's an indication something is wrong.

I do also suggest you check out the JAES article the original poster first
referenced.

>Probably re-sampling a 44.1 kHz file to 192 kHz would have been
>useless.

That depends on what is causing the difference. What if you take a 192
ksamp/sec file, downsample it to 44.1 and then upsample it again? Is there
still a difference? What if you low-pass it at 20 KHz without changing
the sample rate? Is there still a difference?

Presumably your synthesized material doesn't actually have any ultrasonics
but it would be a good idea to verify that first, too.

It is so, so easy for things to go wrong in the conversion process that
result in everything sounding "clearer and brighter" as we sadly found
out in the 1990s. This also points out the need to compare with the original
sound, and not just playback at multiple sample rates.

And it is much harder to do conversion accurately and cleanly at 192 than
it is at 44.1. Hell, it took us thirty years to get 44.1 working properly
and plenty of manufacturers still can't get it right. We finally figured out
how to get stable clocking with low jitter at 44.1 and now everybody wants to
do it four times faster with the same accuracy.

>But oscillators producing real time sounds change a lot at different
>rates.

This depends on what the sounds are, I suspect, also.

The only difference that should exist between 44.1 and 192 ksamp/conversion
is the bandwidth. Now, this being the real world, often there are other
differences too. But your job now becomes sorting out the differences that
are due to the wider bandwidth from the differences due to converter artifacts.

The wider bandwidth is a mixed bag and may have some good attributes but can
also have some bad ones. The other converter artifacts are all bad, but they
might be more bad at 44.1 than at 192, or then again they might be more bad
at 192 than 44.1, and then when you change your clock source it can all change.
This is what makes digital fun.
--scott

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

[Don Pearce[_3_]]

On Sun, 22 Mar 2015 13:01:05 +0100, BBQ
> wrote:

>On Sat, 21 Mar 2015 23:17:35 -0400, Mike Rivers >
>wrote:
>
>>On 3/21/2015 9:55 PM, BBQ wrote:
>>> the difference between 96 and 192 is.... well, hard to explain, it's
>>> more feelings.
>>
>>Yeah, the singers just sing better when they're being recorded at 192
>>kHz. It's been demonstrated that in some instances, what goes into our
>>ear above 20 kHz can make a difference in how we perceive what's within
>>the normal hearing range, so a recording that starts out at 2x sample
>>rate will preserve whatever makes it though the microphone. But
>>re-sampling a 44.1 kHz file to 192 kHz doesn't put back anything that
>>was cut off by the anti-alias filtering. It only make the D/A converter
>>work better.
>
>Ok Mike, but we tested with synthesizers working with different rates
>and the difference between 44 and the higher rates is very clear.
>
>Probably re-sampling a 44.1 kHz file to 192 kHz would have been
>useless.
>
>But oscillators producing real time sounds change a lot at different
>rates.

Can you post a few samples at various rates and depths? I know I'd
like to hear them.

d

[BBQ]

On Sun, 22 Mar 2015 12:40:47 GMT, (Don Pearce) wrote:

>On Sun, 22 Mar 2015 13:01:05 +0100, BBQ
> wrote:
>
>>On Sat, 21 Mar 2015 23:17:35 -0400, Mike Rivers >
>>wrote:
>>
>>>On 3/21/2015 9:55 PM, BBQ wrote:
>>>> the difference between 96 and 192 is.... well, hard to explain, it's
>>>> more feelings.
>>>
>>>Yeah, the singers just sing better when they're being recorded at 192
>>>kHz. It's been demonstrated that in some instances, what goes into our
>>>ear above 20 kHz can make a difference in how we perceive what's within
>>>the normal hearing range, so a recording that starts out at 2x sample
>>>rate will preserve whatever makes it though the microphone. But
>>>re-sampling a 44.1 kHz file to 192 kHz doesn't put back anything that
>>>was cut off by the anti-alias filtering. It only make the D/A converter
>>>work better.
>>
>>Ok Mike, but we tested with synthesizers working with different rates
>>and the difference between 44 and the higher rates is very clear.
>>
>>Probably re-sampling a 44.1 kHz file to 192 kHz would have been
>>useless.
>>
>>But oscillators producing real time sounds change a lot at different
>>rates.
>
>Can you post a few samples at various rates and depths? I know I'd
>like to hear them.
>
>d

Next time we'll do something similar I'll ask him for recordings to
bring home and I'll post the links here.

Anyway, I want to search on the web for similar results. I'm sure
samplers are useless, except if the original recording is made at
higher rates and not resampled from 44.

And then there is the synths question. I'll search articles about
synthesizers with oscillators at higher rates.
--

http://GasGrills.biz | The Best Grills Website for the Serious Griller!

[Don Pearce[_3_]]

On Sun, 22 Mar 2015 17:17:16 +0100, BBQ
> wrote:

>On Sun, 22 Mar 2015 12:40:47 GMT, (Don Pearce) wrote:
>
>>On Sun, 22 Mar 2015 13:01:05 +0100, BBQ
> wrote:
>>
>>>On Sat, 21 Mar 2015 23:17:35 -0400, Mike Rivers >
>>>wrote:
>>>
>>>>On 3/21/2015 9:55 PM, BBQ wrote:
>>>>> the difference between 96 and 192 is.... well, hard to explain, it's
>>>>> more feelings.
>>>>
>>>>Yeah, the singers just sing better when they're being recorded at 192
>>>>kHz. It's been demonstrated that in some instances, what goes into our
>>>>ear above 20 kHz can make a difference in how we perceive what's within
>>>>the normal hearing range, so a recording that starts out at 2x sample
>>>>rate will preserve whatever makes it though the microphone. But
>>>>re-sampling a 44.1 kHz file to 192 kHz doesn't put back anything that
>>>>was cut off by the anti-alias filtering. It only make the D/A converter
>>>>work better.
>>>
>>>Ok Mike, but we tested with synthesizers working with different rates
>>>and the difference between 44 and the higher rates is very clear.
>>>
>>>Probably re-sampling a 44.1 kHz file to 192 kHz would have been
>>>useless.
>>>
>>>But oscillators producing real time sounds change a lot at different
>>>rates.
>>
>>Can you post a few samples at various rates and depths? I know I'd
>>like to hear them.
>>
>>d
>
>Next time we'll do something similar I'll ask him for recordings to
>bring home and I'll post the links here.
>
>Anyway, I want to search on the web for similar results. I'm sure
>samplers are useless, except if the original recording is made at
>higher rates and not resampled from 44.
>
>And then there is the synths question. I'll search articles about
>synthesizers with oscillators at higher rates.

I had a trawl around the web and found a High Res resource

http://www.findhdmusic.com/high-res-audio/free-music/

Specifically, for my examination

http://www.lindberg.no/hires/test/2L-106/2L-106_stereo_PCM-192k_MAGNIFICAT_04.flac

So I had a look at what there actually was in the areas where Hi Res
has an advantage over normal 16 bit.

First - frequency. Two things were there. The first was a couple of
screeches which undoubtedly came from switching power supplies, either
in walll warts or the equipment itself. Second was a huge increase in
noise - a rising spectrum up to the Nyquist frequency. Neither of
these was an enhancement, and the audio would be better off without
them. The spectrum can be seen here.

http://www.soundthoughts.co.uk/look/hi_res.png

The second domain where hi-res might show an advantage is in its bit
depth. 16 bits will take you to about 90dB dynamic range. 24 bits to
maybe 100dB, but nothing goes further than that as quantum noise
(Johnson noise) becomes a limiting factor. The actual dynamic range in
the piece I examined was 60dB.

http://www.soundthoughts.co.uk/look/hi_res_amplitude.png

So not only was high-res unnecessary, it fell short of 16 bit dynamic
range by about 30dB.

Unless you can show something way, way better than this I think your
ideas are in trouble. Incidentally, from the bit depth point of view,
you will never find an acoustic recording that will do substantially
better than the one I examined.

d

[geoff]

On 23/03/2015 12:57 a.m., BBQ wrote:

>
> We tested the rates with synthesizers. It was not a resampled sound.
> It would have been useless that way.

Synths would likely inherently have a rather restricted bandwidth. You
would have had more meaningful results with a (high) harmonically-rich
natural sound.

geoff

[Peter Larsen[_3_]]

"geoff" > skrev i en meddelelse
...

> On 23/03/2015 12:57 a.m., BBQ wrote:

>> We tested the rates with synthesizers. It was not a resampled sound.
>> It would have been useless that way.

> Synths would likely inherently have a rather restricted bandwidth. You
> would have had more meaningful results with a (high) harmonically-rich
> natural sound.

It is undefined whether they are analog or digital. But it is unquestinably
more to the point to record a violin. Those appear to have output to around
30 kHz, suggesting that the wise sample rate choice is 64 kHz and anything
above that is a waste of storage space.

There is however also the issue that at least the converters, I can afford,
do sound better at higher sample rates and the less bad choice consequently
is to downsample in post.

> geoff

Kind regards

Peter Larsen

[Scott Dorsey]

Don Pearce > wrote:
>Specifically, for my examination
>
>http://www.lindberg.no/hires/test/2L-106/2L-106_stereo_PCM-192k_MAGNIFICAT_04.flac
>
>So I had a look at what there actually was in the areas where Hi Res
>has an advantage over normal 16 bit.
>
>First - frequency. Two things were there. The first was a couple of
>screeches which undoubtedly came from switching power supplies, either
>in walll warts or the equipment itself. Second was a huge increase in
>noise - a rising spectrum up to the Nyquist frequency. Neither of
>these was an enhancement, and the audio would be better off without
>them. The spectrum can be seen here.
>
>http://www.soundthoughts.co.uk/look/hi_res.png

The screeching is bad. The noise, now.. some of that noise might be
correlated and some might be random.

- If it's random, that's bad.

- If it's correlated and it actually came from instruments, AND it does not
create distortion products in the audio band, then it might be a good thing
and certainly is harmless.

- If it creates distortion products in the audio band when played back, then
it is a bad thing.


So... number one... if you lowpass that thing at 20 KHz, and listen to it
before and after lowpassing (do it single blind, get your wife to flip the
switch for you), can you hear a difference between the two?

If everything is perfect, you should hear no difference, but this being the
real world, I bet you can hear a difference.

>Unless you can show something way, way better than this I think your
>ideas are in trouble. Incidentally, from the bit depth point of view,
>you will never find an acoustic recording that will do substantially
>better than the one I examined.

The end result is whether it really does sound more accurate. Not whether
it sounds better, or whether it sounds different, but whether it sounds more
accurate.

If it sounds better but it's not more accurate, then we need to make the
same thing happen so when producers demand that sound we can put it on a
lower sample-rate medium. If it sounds different but after extended listening
it turns out not actually to be better (which is what I suspect will be the
case), it's a non-issue.
--scott

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

[Scott Dorsey]

geoff > wrote:
>On 23/03/2015 12:57 a.m., BBQ wrote:
>
>>
>> We tested the rates with synthesizers. It was not a resampled sound.
>> It would have been useless that way.
>
>Synths would likely inherently have a rather restricted bandwidth. You
>would have had more meaningful results with a (high) harmonically-rich
>natural sound.

Depends on the synth... if it's something like a moog it might have weird
harmonics going way way up on some patch configurations.

A lot of early digital synthesizers had poor or no reconstruction filters
too, and consequently had an enormous ultrasonic output.

Now... just because there is a lot of ultrasonic stuff doesn't mean it's
good to reproduce it.

1. find out if there are ultrasonics.

2. find out if they are correlated with signal or are random trash.

3. find out if they are a good thing to have or not.

It's easy to make things sound different, it's hard to make them sound better.
It's easy to think that different is better when it is not.
--scott

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

[Scott Dorsey]

Peter Larsen > wrote:
>There is however also the issue that at least the converters, I can afford,
>do sound better at higher sample rates and the less bad choice consequently
>is to downsample in post.

THIS is important.

And my sneaking suspicion, having not done a complete survey, is that some
converters sound better at higher sample rates and others sound better at
lower sample rates. (And which rate some others sound better might depend
on the musical content, too.)
--scott

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

[Luxey]

недеља, 22. март 2015. 21..28.56 UTC+1, Peter Larsen је написао/ла:

> There is however also the issue that at least the converters, I can afford,
> do sound better at higher sample rates and the less bad choice consequently
> is to downsample in post.

Same thing here and that is exactly where I was aiming with my question to BBQ.
I see others have similar experience.

Nothing better in audio it self, but maybe we could say that audio of higher sample rates can provide better listening experience by enabling (forcing) hardware to work closer to it's full capabilities.

- 44.1 kHz audio will sound better if upsampled to 96 kHz and played through
96 kHz capable DA, than it would sound without upsampling through 44.1 kHz DA, where by all possible means and logic audio content is exactly the same in both scenarios, only because DA work better at that sample rate.
- 96K audio played back through 44.1 kHz DA with on the fly conversion, will likely sound (very marginaly) worse, but surely not any better than 44.1 kHz
audio, even if it is audio upsampled to 96 kHz from 44.1 kHz.

[JackA]

On Saturday, March 21, 2015 at 7:55:37 PM UTC-4, BBQ wrote:
> On Tue, 17 Mar 2015 19:06:29 -0700 (PDT), JackA
> > wrote:
>
> > "In double-blind tests conducted by Levitin and others -- some results of which were published in the Journal of the Audio Engineering Society -- listeners cannot tell the difference between high-resolution audio and CD-quality audio".
> >
> >"But many audio professionals, including Bob Ludwig and NYU's Jim Anderson, say they can hear an improvement over CD quality, and they prefer the higher frequencies and sample rates. Anderson even teaches a class at NYU in which he instructs students on how to listen for the differences".
> >
> >That's right, you have to be trained to hear the difference!
> >
> >Not sure why Bob Ludwig is mentioned. I mean, boy, his John Cougar CD was LOUD!!
> >
> >Jack
> >whoops...
> >http://www.laweekly.com/music/why-cds-may-actually-sound-better-than-vinyl-5352162
>
> Indeed high-resolution audio is wonderful!
>
> Tested in a friend's studio and it's a lot different. Exactly, highs
> are really brilliant, and there's something about harmonics but I
> can't tell you exactly.
>
> while he was playing a synth I could feel the sound was totally
> different, like more detailed.
>
> It's like when you play an organ on a keyboard and then you hear the
> real cathedral organ in front of you.

Oh, okay. I thought you were listening to something that has been reissued.

Jack
> --
>
> http://GasGrills.biz | The Best Grills Website for the Serious Griller!

[Scott Dorsey]

Luxey > wrote:
>- 44.1 kHz audio will sound better if upsampled to 96 kHz and played throug=
>h=20
>96 kHz capable DA, than it would sound without upsampling through 44.1 kHz =
>DA, where by all possible means and logic audio content is exactly the same=
> in both scenarios, only because DA work better at that sample rate.

That used to be the case, but these days we're all using sigma-delta
converters where everything is effectively upsampled to some tremendously
fast 1-bit stream. So the advantages of oversampling that you saw in the
ladder converter world have all been incorporated into converters already
and adding more SRC doesn't help.

>- 96K audio played back through 44.1 kHz DA with on the fly conversion, wil=
>l likely sound (very marginaly) worse, but surely not any better than 44.1 =
>kHz
>audio, even if it is audio upsampled to 96 kHz from 44.1 kHz.

If you believe the study in the JAES, on a reasonably well-conducted listening
test it was indistinguishable. That's not to say it will be that way under
all conditions; there are many in which playing it back at the lower rate
should be an improvement.
--scott

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

[Don Pearce[_3_]]

On 22 Mar 2015 18:13:43 -0400, (Scott Dorsey) wrote:

>Don Pearce > wrote:
>>Specifically, for my examination
>>
>>http://www.lindberg.no/hires/test/2L-106/2L-106_stereo_PCM-192k_MAGNIFICAT_04.flac
>>
>>So I had a look at what there actually was in the areas where Hi Res
>>has an advantage over normal 16 bit.
>>
>>First - frequency. Two things were there. The first was a couple of
>>screeches which undoubtedly came from switching power supplies, either
>>in walll warts or the equipment itself. Second was a huge increase in
>>noise - a rising spectrum up to the Nyquist frequency. Neither of
>>these was an enhancement, and the audio would be better off without
>>them. The spectrum can be seen here.
>>
>>http://www.soundthoughts.co.uk/look/hi_res.png
>
>The screeching is bad. The noise, now.. some of that noise might be
>correlated and some might be random.
>
>- If it's random, that's bad.
>
>- If it's correlated and it actually came from instruments, AND it does not
> create distortion products in the audio band, then it might be a good thing
> and certainly is harmless.
>
>- If it creates distortion products in the audio band when played back, then
> it is a bad thing.
>
>
>So... number one... if you lowpass that thing at 20 KHz, and listen to it
>before and after lowpassing (do it single blind, get your wife to flip the
>switch for you), can you hear a difference between the two?
>
>If everything is perfect, you should hear no difference, but this being the
>real world, I bet you can hear a difference.
>
>>Unless you can show something way, way better than this I think your
>>ideas are in trouble. Incidentally, from the bit depth point of view,
>>you will never find an acoustic recording that will do substantially
>>better than the one I examined.
>
>The end result is whether it really does sound more accurate. Not whether
>it sounds better, or whether it sounds different, but whether it sounds more
>accurate.
>
>If it sounds better but it's not more accurate, then we need to make the
>same thing happen so when producers demand that sound we can put it on a
>lower sample-rate medium. If it sounds different but after extended listening
>it turns out not actually to be better (which is what I suspect will be the
>case), it's a non-issue.
>--scott

The screeching and noise lift are definitely unwanted artefacts, not
instrument-related, as they are there at the same level during both
loud and quiet passages. I tried lowpass filter at 20kHz, and I'm
afraid I can't hear a scrap of difference. I also downsampled to
44.1kHz and again heard no difference.

d

[Luxey]

Thanks for having such good attention, yes, when I said "upsampling" I
actually meant resampling (SRC), only the damn word would not come to mind.

[hank alrich]

BBQ > wrote:

> We tested the rates with synthesizers. It was not a resampled sound.
> It would have been useless that way.

Bring trumpets, clarinets, violins, pianos, and human voices. Start with
real-in-nature sources.

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

[Randy Yates[_2_]]

writes:

>>
>> Of course High-Res can sound different than traditional
>> lossless/CD: Just EQ or otherwise process the master
>> for high-res DIFFERENTLY than how you did so for
>> lossless! duhhh..
>>
>> Then charge US$10 more for it than you would for the
>> CD version. smh.. Goes on daily, though no one has
>> the guts to admit it.
>
> there is a lot of hype here to be sure but there is at least one nugget of truth
>
> conventional CDs require 20 kHz low pass filters for anti-aliasing and
> reconstruction. Oversampling does not change that, it simply allows th filters to be implementedin the digital domain, but the filters are still needed.
>
> Steep low pass filters __can__ produce ringing around 20 kHz that
> __may__ be audible to some people. (not me)

So I guess your argument is predicated on being able to hear 20 kHz?
Because the same thing happens with high-rate (e.g, 96 kHz) systems,
just at a higher frequency (48 kHz).
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

[[email protected]]

Hi Randy,

Yes, some people may be able to hear ringing at 20 kHz.
I am not one of them.

Two things are different at higher sampling rates.
No one can hear ringing above 30 kHz.
Also, the filters do not have to be steep because
There is more guard band between the top edge of the audio
and the Nyquist freq.

So yes , i could see how there could be under some very unusual
circumstances, audible effects of sampling at 44.1 kHz.

But most of it is hype.

Mark

[Randy Yates[_2_]]

writes:
> [...]
> Also, the filters do not have to be steep because
> There is more guard band between the top edge of the audio
> and the Nyquist freq.

Isn't that an assumption? What if you have significant ultrasonic crap
coming into your analog channel?
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

[John Corbett]

On 03/18/2015 08:26 AM, Scott Dorsey wrote:
> JackA > wrote:
>> "In double-blind tests conducted by Levitin and others -- some results of =
>> which were published in the Journal of the Audio Engineering Society -- lis=
>> teners cannot tell the difference between high-resolution audio and CD-qual=
>> ity audio".
>
> I suggest you read the paper referenced. It's very interesting, and the
> procedure isn't flawless but it's pretty good. It was discussed extensively
> here when it first came out.
>

Scott, which paper are you referring to? Do you mean the King, Levitin,
and Leonard 2012 paper on sample rates (a convention paper but
apparently not in the Journal) or the Meyer and Moran 2007 JAES paper,
or something else?

--John

[[email protected]]

On Monday, March 23, 2015 at 10:48:44 PM UTC-4, Randy Yates wrote:
> writes:
> > [...]
> > Also, the filters do not have to be steep because
> > There is more guard band between the top edge of the audio
> > and the Nyquist freq.
>
> Isn't that an assumption? What if you have significant ultrasonic crap
> coming into your analog channel?
> --

I'm going to assume our goal is to reproduce the signal up to 20 kHz as faithfully as possible. When sampling at 96 ksps, the anti aliasing filter can have a gentle roll off above 20 kHz and avoid any possible ringing and still provide the required attenuation at Nyquist and above to prevent aliasing. If there happens to be some ultrasonics above 20 kHz and below Nyquist, the ultrasonics will simply be reproduced however attenuated by the filter.. This is just about the same behavior as a pure analog system that is flat up to 20 kHz and then gently rolls off above that.

Heres a thought experiment. Scale everything down by a facotr of 10. If you make a brickwall steep low pass filter at 2 kHz, you sure as He** will hear ringing no matter how you implemented the filter. If instead you used a gradual roll off filter, the ringing can be reduced. There are plenty of signal compnents at 2 kHz and all the mics and speakers and everything else in the chain can pass 2 kHz. So you will hear ringing.

The only reason this ___might___ not happen at 20 kHz, is becasue all the other parts of the chain including the source sounds, mics , speakers and EARS have some rolloff at 20 kHz.

So again, in practice there is not a lot of signal energy at 20 kHz to hit the filter to ring and if it does you probably can't hear it anyway. But it is possible in theory.

That's why it is an interesting experiment, and I could belive the results either way.


Mark

[JackA]

On Tuesday, March 24, 2015 at 9:23:40 AM UTC-4, wrote:
> On Monday, March 23, 2015 at 10:48:44 PM UTC-4, Randy Yates wrote:
> > writes:
> > > [...]
> > > Also, the filters do not have to be steep because
> > > There is more guard band between the top edge of the audio
> > > and the Nyquist freq.
> >
> > Isn't that an assumption? What if you have significant ultrasonic crap
> > coming into your analog channel?
> > --
>
> I'm going to assume our goal is to reproduce the signal up to 20 kHz as faithfully as possible. When sampling at 96 ksps, the anti aliasing filter can have a gentle roll off above 20 kHz and avoid any possible ringing and still provide the required attenuation at Nyquist and above to prevent aliasing. If there happens to be some ultrasonics above 20 kHz and below Nyquist, the ultrasonics will simply be reproduced however attenuated by the filter. This is just about the same behavior as a pure analog system that is flat up to 20 kHz and then gently rolls off above that.
>
> Heres a thought experiment. Scale everything down by a facotr of 10. If you make a brickwall steep low pass filter at 2 kHz, you sure as He** will hear ringing no matter how you implemented the filter. If instead you used a gradual roll off filter, the ringing can be reduced. There are plenty of signal compnents at 2 kHz and all the mics and speakers and everything else in the chain can pass 2 kHz. So you will hear ringing.
>
> The only reason this ___might___ not happen at 20 kHz, is becasue all the other parts of the chain including the source sounds, mics , speakers and EARS have some rolloff at 20 kHz.
>
> So again, in practice there is not a lot of signal energy at 20 kHz to hit the filter to ring and if it does you probably can't hear it anyway. But it is possible in theory.
>
> That's why it is an interesting experiment, and I could belive the results either way.
>
>
> Mark

This frequency didn't bother the majority...

"Can anyone else hear the 15,750 Hz tone generated by most TVs? My girlfriend was telling me I was crazy and that the sound didn't exist..."

Jack

[Scott Dorsey]

John Corbett > wrote:
>On 03/18/2015 08:26 AM, Scott Dorsey wrote:
>> JackA > wrote:
>>> "In double-blind tests conducted by Levitin and others -- some results of =
>>> which were published in the Journal of the Audio Engineering Society -- lis=
>>> teners cannot tell the difference between high-resolution audio and CD-qual=
>>> ity audio".
>>
>> I suggest you read the paper referenced. It's very interesting, and the
>> procedure isn't flawless but it's pretty good. It was discussed extensively
>> here when it first came out.
>>
>
>Scott, which paper are you referring to? Do you mean the King, Levitin,
>and Leonard 2012 paper on sample rates (a convention paper but
>apparently not in the Journal) or the Meyer and Moran 2007 JAES paper,
>or something else?

I assumed that the original article that JackA quoted is referring to the
Meyer and Moran paper, since it described vaguely some details of the
procedure in that paper.

The King, Levitin, and Brett paper used identical converters being run at
different rates rather than downsampling and upsampling and so there are all
kinds of converter issues that weren't controlled for. And the actual results
weren't that listeners couldn't tell the difference, really.

Both of these papers are well worth reading, though, and they describe a
lot of the difficulties with effectively evaluating high sample rate
operation.

The reference you want on pre-ripple in the audio passband caused by anti
aliasing and reconstruction filters is J. Dunn's "The Benefits of 96KHz
Sampling Rate Formats For Those Who Cannot Hear Above 20 KHz." I don't think
this is actually a significant problem but it would be interesting to see if
it really is.

Oohashi's famous paper is a good example of how an A/B comparison of sample
rates can be totally screwed by by careless procedure and intermodulation
products...
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Randy Yates[_2_]]

writes:

> On Monday, March 23, 2015 at 10:48:44 PM UTC-4, Randy Yates wrote:
>> writes:
>> > [...]
>> > Also, the filters do not have to be steep because
>> > There is more guard band between the top edge of the audio
>> > and the Nyquist freq.
>>
>> Isn't that an assumption? What if you have significant ultrasonic crap
>> coming into your analog channel?
>> --
>
> I'm going to assume our goal is to reproduce the signal up to 20 kHz
> as faithfully as possible. When sampling at 96 ksps, the anti aliasing
> filter can have a gentle roll off above 20 kHz and avoid any possible
> ringing and still provide the required attenuation at Nyquist and
> above to prevent aliasing. If there happens to be some ultrasonics
> above 20 kHz and below Nyquist, the ultrasonics will simply be
> reproduced however attenuated by the filter. This is just about the
> same behavior as a pure analog system that is flat up to 20 kHz and
> then gently rolls off above that.

Mark,

Sheesh I feel stupid. It's been too many years since I went through
this exercise. Of course you are correct. Thanks for refreshing my
memory!

> Heres a thought experiment. Scale everything down by a facotr of 10.
> If you make a brickwall steep low pass filter at 2 kHz, you sure as
> He** will hear ringing no matter how you implemented the filter. If
> instead you used a gradual roll off filter, the ringing can be
> reduced. There are plenty of signal compnents at 2 kHz and all the
> mics and speakers and everything else in the chain can pass 2 kHz. So
> you will hear ringing.

Yes, of course - 2 kHz would be audible.

> The only reason this ___might___ not happen at 20 kHz, is becasue all
> the other parts of the chain including the source sounds, mics ,
> speakers and EARS have some rolloff at 20 kHz.
>
> So again, in practice there is not a lot of signal energy at 20 kHz to
> hit the filter to ring and if it does you probably can't hear it
> anyway. But it is possible in theory.
>
> That's why it is an interesting experiment, and I could belive the
> results either way.

Attributing this difference to 20 kHz ringing means the person can hear
20~kHz. How many can? I bet not many, but I am not aware of any
statistics reporting this. (I need to check my Zwicker...) I know you've
acknowledged this, just repeating it for clarity.

My belief is that, if there is a difference, it's because of something
else (converter nonlinearities, colored noise, e.g.). However, my first
suspicion is that there is NO difference when the listener is subjected
to A/B testing. (Hey, what happened to that guy, Arnie somesuch?)
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

[Scott Dorsey]

Randy Yates > wrote:

>Attributing this difference to 20 kHz ringing means the person can hear
>20~kHz. How many can? I bet not many, but I am not aware of any
>statistics reporting this. (I need to check my Zwicker...) I know you've
>acknowledged this, just repeating it for clarity.

I can't hear 20 KHz. I can't even hear 18 KHz anymore, in spite of having
spent much of the last 30 years with earplugs in.

BUT.... if there are actual genuine sounds near 20 KHz and there is also
spurious ringing at 20 KHz, it might generate audible beat products due to
playback nonlinearity. And those beat products will fall down into the
audible region and I might be able to hear them.

Of course, the solution for this problem might be to address the high
frequency nonlinearity in the first place.

>My belief is that, if there is a difference, it's because of something
>else (converter nonlinearities, colored noise, e.g.). However, my first
>suspicion is that there is NO difference when the listener is subjected
>to A/B testing. (Hey, what happened to that guy, Arnie somesuch?)

I am pretty sure that you're right, but building a foolproof experiment to
prove it is difficult if not impossible.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Luxey]

For the beat to be in the "really" audible range, 20 kHz would have to beat over
something well over 5 kHz, and three's already not too much, as far as listening
to the "pop" music goes. Listening for tech artifacts is different story.

I know some people really care about the smoothness of decay of the triangle
during crescendo of symphonic orchestra, but I'm not that type.
I also know many people just pretend they care about it, for perceived sophistication. I'm not that type, either.

[Scott Dorsey]

Luxey > wrote:
>For the beat to be in the "really" audible range, 20 kHz would have to beat over
>something well over 5 kHz, and three's already not too much, as far as listening
>to the "pop" music goes. Listening for tech artifacts is different story.

Oh, let's say there's 19 KHz content from a violin, for instance, in the
original recording. Add 20 KHz ringing (and the ringing is not just a pure
tone, it's correlated with signal) and the combination gives you a correlated
distortion product at 1 KHz.

Significant? I don't think so because the ringing is so close to the noise
floor anyway, but you can do the math and decide.

>I know some people really care about the smoothness of decay of the triangle
>during crescendo of symphonic orchestra, but I'm not that type.

It's my job to worry about stuff like that, though, so you don't have to.

>I also know many people just pretend they care about it, for perceived sophistication. I'm not that type, either.

I don't care if the customer really cares or just pretends they care, either
way I'm going to worry about it if I am paid to worry about it.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Luxey]

среда, 25. март 2015. 19.10.00 UTC+1, Scott Dorsey је написао/ла:
> Oh, let's say there's 19 KHz content from a violin, for instance, in the
> original recording. Add 20 KHz ringing (and the ringing is not just a pure
> tone, it's correlated with signal) and the combination gives you a correlated
> distortion product at 1 KHz.

Well, yes, that's about that. Violin content at 15K, or above, if any, I don't
know, can't be more than a tiney sine wave. At 20K, whatever it is, just the same.
I mean that's what could be heard. More can be measured hiigher than 20K, but
1. It for sure is not something usual microphone could possibly capture
and pass further, therefore it has no place in recording
2. Since there's "nothing" to hear anyway, there's filter doing it's job,
just at the right spot, there at 20K, or so. And if that was not enough, it's all overssampled and so on ...

[JackA]

On Wednesday, March 25, 2015 at 7:26:48 PM UTC-4, Luxey wrote:
> среда, 25. март 2015. 19.10.00 UTC+1, Scott Dorsey је написао/ла:
> > Oh, let's say there's 19 KHz content from a violin, for instance, in the
> > original recording. Add 20 KHz ringing (and the ringing is not just a pure
> > tone, it's correlated with signal) and the combination gives you a correlated
> > distortion product at 1 KHz.
>
> Well, yes, that's about that. Violin content at 15K, or above, if any, I don't
> know, can't be more than a tiney sine wave. At 20K, whatever it is, just the same.
> I mean that's what could be heard. More can be measured hiigher than 20K, but
> 1. It for sure is not something usual microphone could possibly capture
> and pass further, therefore it has no place in recording
> 2. Since there's "nothing" to hear anyway, there's filter doing it's job,
> just at the right spot, there at 20K, or so. And if that was not enough, it's all overssampled and so on ...

If I may interject - one or more of Heart's hits has a sharp cut-off about 11kHz. Recording and/or mixing cymbals always seem(s)ed to be a problem.

Jack

[Scott Dorsey]

Luxey > wrote:
>> Oh, let's say there's 19 KHz content from a violin, for instance, in the
>> original recording. Add 20 KHz ringing (and the ringing is not just a pu=
>re
>> tone, it's correlated with signal) and the combination gives you a correl=
>ated
>> distortion product at 1 KHz.
>
>Well, yes, that's about that. Violin content at 15K, or above, if any, I do=
>n't=20
>know, can't be more than a tiney sine wave. At 20K, whatever it is, just th=
>e same.
>I mean that's what could be heard. More can be measured hiigher than 20K, b=
>ut
>1. It for sure is not something usual microphone could possibly capture
>and pass further, therefore it has no place in recording
>2. Since there's "nothing" to hear anyway, there's filter doing it's job,=

There is a lot of stuff up near and above 20 KHz with a fiddle, and that
filter is clamping very abruptly at 22.05 KHz so there is plenty of converter
response at 19 KHz. The microphone might be limiting it a lot, but then
again it might not. Depends.

It's enough to worry about. The question is whether the degree of ringing is
enough to worry about, and I am reasonably sure it's not but I'm not willing
to place money on it.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."