[James T]

I've read through some archived r.a.p. threads where it was argued
that 16 bit/44.1k had an inaudible affect on music, at least as far as
the cited test went. I understand the merit in having more bits for
multitrack ops, where internal math can lose bits. But now I'm curious
about the widespread 'understanding' that 24-96 is better. Is there no
merit whatsoever in final mix?

[Les Cargill[_4_]]

James T wrote:
I've read through some archived r.a.p. threads where it was argued
that 16 bit/44.1k had an inaudible affect on music, at least as far as
the cited test went. I understand the merit in having more bits for
multitrack ops, where internal math can lose bits. But now I'm curious
about the widespread 'understanding' that 24-96 is better. Is there no
merit whatsoever in final mix?



It's an open question. Reputable people can hear differences. There
are no ... *essential* information-theoretic reasons for it.

24 bit depth is good for tracking because of the additional 8 bits
headroom.

--
Les Cargill

[Robert Morein]

"James T" wrote in message
...

I've read through some archived r.a.p. threads where it was argued
that 16 bit/44.1k had an inaudible affect on music, at least as far as
the cited test went. I understand the merit in having more bits for
multitrack ops, where internal math can lose bits. But now I'm curious
about the widespread 'understanding' that 24-96 is better. Is there no
merit whatsoever in final mix?

I think there is. This is my vote, not a statement of fact.

Bob Morein
(310) 237-6511

[Scott Dorsey]

James T wrote:
I've read through some archived r.a.p. threads where it was argued
that 16 bit/44.1k had an inaudible affect on music, at least as far as
the cited test went. I understand the merit in having more bits for
multitrack ops, where internal math can lose bits. But now I'm curious
about the widespread 'understanding' that 24-96 is better. Is there no
merit whatsoever in final mix?

Try it for yourself and see.

I don't think it has any merit, and in fact it may have some degradation
if you wind up carrying along out-of-band material that winds up causing
audble distortion products.

But try it yourself and see. Record at 24/96 then downsample to 44.1,
then upsample to 96. Can you tell the difference between the 44.1 and
the 96 ksamp/sec audio?

Some converters sound a lot better at 44.1 than 96, other converters
sound better at 96 than 44.1, but the beauty of the modern era is that
you can run the converter at whatever sample rate it sounds good at
and then resample to whatever rate you want to store it at.
--scott

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

[William Sommerwerck]

To my ears, Blu-ray audio sounds noticeably better than the common run of CD
sound. However...

Blu-ray audio disks are audiophile recordings, and may use better mics,
fewsd\\er electronics, less processing, etc, than Compact Disk recordings.
At the moment, I cannot say with any objectivity that "high resolution"
recordings are definitely superior. In theory, there's no reason for
them being better, as 16/44 presumably records everything that's audible.

[Sean Conolly]

"James T" wrote in message
...
I've read through some archived r.a.p. threads where it was argued
that 16 bit/44.1k had an inaudible affect on music, at least as far as
the cited test went. I understand the merit in having more bits for
multitrack ops, where internal math can lose bits. But now I'm curious
about the widespread 'understanding' that 24-96 is better. Is there no
merit whatsoever in final mix?

The problem is that the difference is so slight that it becomes subjective.
Some people will hear a difference, some of the time, but most will not.

Higher bit rates have a real benefit, but for me higher sample rates do not.
I have so many other factors that audibly compromise my recordings, sample
rate is the least of my concerns.

Sean

[James T]

On Sat, 16 Jun 2012 12:54:50 -0700, "William Sommerwerck"
wrote:

To my ears, Blu-ray audio sounds noticeably better than the common run of CD
sound. However...

Blu-ray audio disks are audiophile recordings, and may use better mics,
fewsd\\er electronics, less processing, etc, than Compact Disk recordings.
At the moment, I cannot say with any objectivity that "high resolution"
recordings are definitely superior. In theory, there's no reason for
them being better, as 16/44 presumably records everything that's audible.


I can understand that ('records everything that's audible") purely
with regard to capturing waveform -bandwidth- within predefined span
of human hearing (20-20k, which is very optimistic). Nyquist/Shannon
covers sample rates. But has bit depth been thoroughly explored?

Also wondering how transient and positional info relate to the sample
rate question.

[Scott Dorsey]

James T wrote:
I can understand that ('records everything that's audible") purely
with regard to capturing waveform -bandwidth- within predefined span
of human hearing (20-20k, which is very optimistic). Nyquist/Shannon
covers sample rates. But has bit depth been thoroughly explored?

Yes, in the early 1950s. Bit depth covers dynamic range.

Also wondering how transient and positional info relate to the sample
rate question.

In the same way that they relate to frequency response.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[James T]

On 16 Jun 2012 15:40:24 -0400, (Scott Dorsey) wrote:

James T wrote:
I've read through some archived r.a.p. threads where it was argued
that 16 bit/44.1k had an inaudible affect on music, at least as far as
the cited test went. I understand the merit in having more bits for
multitrack ops, where internal math can lose bits. But now I'm curious
about the widespread 'understanding' that 24-96 is better. Is there no
merit whatsoever in final mix?

Try it for yourself and see.

I may do that eventually, but I have presumed (perhaps wrongly) that
the question has been thoroughly explored by those with greater means
than my own.

I don't think it has any merit, and in fact it may have some degradation
if you wind up carrying along out-of-band material that winds up causing
audble distortion products.

Interesting side effect.

But try it yourself and see. Record at 24/96 then downsample to 44.1,
then upsample to 96. Can you tell the difference between the 44.1 and
the 96 ksamp/sec audio?

I've seen the (AES-published?) blindfold tests that introduced an A/D
and D/A into a loop with a presumably high quality source. I thought
that should be a reasonable baseline, but of course such a test is not
necessarily a good basis for judging whether the 16-44.1 conversion
-can- be perceived--just whether a statistically significant number of
the test subjects were able to perceive it.

IOW, I figured there was going to be a caveat in there somewhere.

Some converters sound a lot better at 44.1 than 96, other converters
sound better at 96 than 44.1, but the beauty of the modern era is that
you can run the converter at whatever sample rate it sounds good at
and then resample to whatever rate you want to store it at.
--scott

That makes sense as well. Given the near-universal acceptance of 2496,
and the increased storage requirements, I expect that there are test
results somewhere that 'justify' it rather than just marketing or
payola from the disk drive industry. :-)

Also wondering how transient and positional info relate, since as far
as I know, the relevant refs to Nyquist cover capture of static
waveform data. I haven't seen much tech info on auditory positional
cues and such. Nor do I know if the tests above used source material
that would be likely to expose deficiencies there.

Thanks to all who have commened!

[James T]

On Sat, 16 Jun 2012 17:10:38 -0400, James T
wrote:

Thanks to all who have commened!

And also to those who have err.... "Commented"

[James T]

On 16 Jun 2012 17:04:28 -0400, (Scott Dorsey) wrote:

James T wrote:
I can understand that ('records everything that's audible") purely
with regard to capturing waveform -bandwidth- within predefined span
of human hearing (20-20k, which is very optimistic). Nyquist/Shannon
covers sample rates. But has bit depth been thoroughly explored?

Yes, in the early 1950s. Bit depth covers dynamic range.

Also wondering how transient and positional info relate to the sample
rate question.

In the same way that they relate to frequency response.
--scott

Of course transients will correspond to ability to reproduce
equivalent frequencies. I understand that. But I was wondering to what
extent that type of information would be likely to exceed ye olde
20-20khz spec that is regarded as the baseline for human perception.
There seems to be a lot of gray area there. IOW, did the A/D/A tests
done with the 16-44.1 loop adequately test perception of that aspect
of human hearing. Do you know of any relevant test data?

[Randy Yates[_2_]]

James T writes:

I've read through some archived r.a.p. threads where it was argued
that 16 bit/44.1k had an inaudible affect on music, at least as far as
the cited test went. I understand the merit in having more bits for
multitrack ops, where internal math can lose bits. But now I'm curious
about the widespread 'understanding' that 24-96 is better. Is there no
merit whatsoever in final mix?

James,

I would not say "there is no merit whatsoever in the extra range," but
I do think the extra cost to convert to such a medium would far, far
outweigh any advantages.

Consider this thought experiment:

A keyboardist uses his 24-bit Korg SuperHighHD-1024 synth and a
two-track digital recorder to record two tracks at 24 bits perfectly.
During the first minute of his composition, the average power output
is -60 dBFS, then the finale crescendos to an average of -20 dBFS with
peaks near full-scale.

Now convert that to 16-bit using your favorite noise-shaping
algorithm, and play it back using ear buds in a quiet room such that
an ambient level 30 dBA is achieved with the ear buds in. Set levels
so that the recording's -60 dBFS section comes in at 60 dBA over the
ear bugs.

The full-scale finale would then translate to around 100 dBA SPL
average with peaks around 120 dBA - loud but within reason.

With ambient noise 10 dB quieter than the quantization noise, I think
many folks would be able to hear the quantization noise in the initial
quiet passage.

So, yeah, in those near-laboratory conditions and this very strange
composition, 24 bits would give an advantage, I believe. Otherwise, in
the vast, vast majority of practical scenarios, I believe there are very
very very few people who could tell the difference.
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

[Les Cargill[_4_]]

James T wrote:
On 16 Jun 2012 17:04:28 -0400, (Scott Dorsey) wrote:

James T wrote:
I can understand that ('records everything that's audible") purely
with regard to capturing waveform -bandwidth- within predefined span
of human hearing (20-20k, which is very optimistic). Nyquist/Shannon
covers sample rates. But has bit depth been thoroughly explored?

Yes, in the early 1950s. Bit depth covers dynamic range.

Also wondering how transient and positional info relate to the sample
rate question.

In the same way that they relate to frequency response.
--scott

Of course transients will correspond to ability to reproduce
equivalent frequencies. I understand that. But I was wondering to what
extent that type of information would be likely to exceed ye olde
20-20khz spec that is regarded as the baseline for human perception.
There seems to be a lot of gray area there. IOW, did the A/D/A tests
done with the 16-44.1 loop adequately test perception of that aspect
of human hearing. Do you know of any relevant test data?


The 20Hz-20KHz bandlimit is a pretty ... adequate defacto standard.
The problem with transients is that other things in the playback
chain and in the recording chain will have a much greater effect
than the recording medium.

From a standpoint of "what do I fix first", any 44.1/16 bit system
is probably much better than any other element in the system.

--
Les Cargill

[Mike Rivers]

On 6/16/2012 2:42 PM, James T wrote:
I understand the merit in having more bits for
multitrack ops, where internal math can lose bits.

Just what is it that you understand about this? Why do you
think that internal math can lose bits in a multitrack
recording? Can you explain?

But now I'm curious
about the widespread 'understanding' that 24-96 is better. Is there no
merit whatsoever in final mix?

Of course there's some merit, but most people don't care or
don't have the reproduction facilities to gain what might be
a small advantage.

However, it allows people who don't understand good
engineering principles or software design to be sloppier and
still get acceptable results. And it sells more storage
space media.



--
"Today's production equipment is IT based and cannot be
operated without a passing knowledge of computing, although
it seems that it can be operated without a passing knowledge
of audio." - John Watkinson

http://mikeriversaudio.wordpress.com - useful and
interesting audio stuff

[Mike Rivers]

On 6/16/2012 3:40 PM, Scott Dorsey wrote:

Try it for yourself and see.

Just because he can't hear a difference doesn't mean there
isn't one.

To some, the issue is that if it's theoretically better, if
there's the slightest possibility that someone might be able
to detect an improvement, now or in the future, than it's
worth doing. But I'm with you - good 44.1 kHz 16-bit
recordings sound just fine, and bad recordings won't be
improved by jacking up the sample rate or storing more bits.



--
"Today's production equipment is IT based and cannot be
operated without a passing knowledge of computing, although
it seems that it can be operated without a passing knowledge
of audio." - John Watkinson

http://mikeriversaudio.wordpress.com - useful and
interesting audio stuff

[Mike Rivers]

On 6/16/2012 5:10 PM, James T wrote:

I have presumed (perhaps wrongly) that
the question has been thoroughly explored by those with greater means
than my own.

There have been demonstrations and presumably well conducted
tests that prove both ways. It all boils down to human
hearing, and humans all hear differently. And equally
important, all perceive the same thing differently.

Why do you care? If you have a client that requests a high
resolution recording, by all means, don't argue with him.
But if you're recording for your own amusement, I'd spend
some time auditioning a few different A/D and D/A converters
at 44.1 kHz and pick the ones you like best.

I've seen the (AES-published?) blindfold tests that introduced an A/D
and D/A into a loop with a presumably high quality source. I thought
that should be a reasonable baseline, but of course such a test is not
necessarily a good basis for judging whether the 16-44.1 conversion
-can- be perceived--just whether a statistically significant number of
the test subjects were able to perceive it.

Several years ago my friendly local dealer set up a
demonstration. He set up a singer/guitarist with a high
quality stereo mic setup (Senheisser MKH-800s, I recall, one
of the first honest mics that demonstrated usable response
up to 40 kHz ("the one to use if you're going to record at
2x sample rate). He split the preamp output three ways using
a high grade Coleman switch box. One path was direct from
the preamp to the powered speakers (these were PMC, pretty
high grade stuff, too), one was through a pair of Lavry Blue
converters back to back running at 44.1 kHz, the other was
through another pair of the same converters running at 96
kHz. I believe both were using 24-bit resolution, or at
least both were the same.

All three sounded different. I, and most of the listeners,
had a preference for the straight-through path, some
preferred the 44.1 kHz path, some preferred the 96 kHz path,
but nobody would have always picked one sample rate over the
other every time. These were pro engineers, which doesn't
necessarily make them any more qualified than anyone else,
but they were probably more accustomed to listening for
small differences.

Also wondering how transient and positional info relate, since as far
as I know, the relevant refs to Nyquist cover capture of static
waveform data. I haven't seen much tech info on auditory positional
cues and such. Nor do I know if the tests above used source material
that would be likely to expose deficiencies there.

I think I read something about positional cues, but mostly
it was about why sample rate doesn't improve it.



--
"Today's production equipment is IT based and cannot be
operated without a passing knowledge of computing, although
it seems that it can be operated without a passing knowledge
of audio." - John Watkinson

http://mikeriversaudio.wordpress.com - useful and
interesting audio stuff

[hank alrich]

James T wrote:

On 16 Jun 2012 15:40:24 -0400, (Scott Dorsey) wrote:

James T wrote:
I've read through some archived r.a.p. threads where it was argued
that 16 bit/44.1k had an inaudible affect on music, at least as far as
the cited test went. I understand the merit in having more bits for
multitrack ops, where internal math can lose bits. But now I'm curious
about the widespread 'understanding' that 24-96 is better. Is there no
merit whatsoever in final mix?

Try it for yourself and see.

I may do that eventually, but I have presumed (perhaps wrongly) that
the question has been thoroughly explored by those with greater means
than my own.

Given the range of convertors, acoustical environments, source material,
and variation in sensitivity to different artifacts among individual
listeners, Scott's suggesstion goes right to the core of the matter.

Some convertors perform better at one or another sample rate, and it
goes on from there. So the relevant question is, what works best for you
in the context of the available parameters?

For me with the specific interface I've been using, a Metric Halo
2882+DSP, recently upgraded with the new 2d card, 24/96 has become the
default setting. That doesn't mean you should make the same decision.

--
shut up and play your guitar * http://hankalrich.com/
http://www.youtube.com/walkinaymusic
http://www.sonicbids.com/HankandShaidri

[hank alrich]

Mike Rivers wrote:

On 6/16/2012 3:40 PM, Scott Dorsey wrote:

Try it for yourself and see.

Just because he can't hear a difference doesn't mean there
isn't one.

To some, the issue is that if it's theoretically better, if
there's the slightest possibility that someone might be able
to detect an improvement, now or in the future, than it's
worth doing. But I'm with you - good 44.1 kHz 16-bit
recordings sound just fine, and bad recordings won't be
improved by jacking up the sample rate or storing more bits.

Both of you are troublemakers! g

--
shut up and play your guitar * http://hankalrich.com/
http://www.youtube.com/walkinaymusic
http://www.sonicbids.com/HankandShaidri

[hank alrich]

Randy Yates wrote:

James T writes:

I've read through some archived r.a.p. threads where it was argued
that 16 bit/44.1k had an inaudible affect on music, at least as far as
the cited test went. I understand the merit in having more bits for
multitrack ops, where internal math can lose bits. But now I'm curious
about the widespread 'understanding' that 24-96 is better. Is there no
merit whatsoever in final mix?

James,

I would not say "there is no merit whatsoever in the extra range," but
I do think the extra cost to convert to such a medium would far, far
outweigh any advantages.

I think that concern made more sense when storage space was much more
expensive. Granted if one is working with large numbers of tracks it
could still be a constraint. Now with drives costing a little as they do
I no longer think twice about using higher rates.

At this point with the landscape littered with AVID/Digi HD192's, Metric
Halo ULN-8's, and so many similar;y spec'd convertor systems, there's a
lot of 24/192 going down, even where track counts are many dozens.

--
shut up and play your guitar * http://hankalrich.com/
http://www.youtube.com/walkinaymusic
http://www.sonicbids.com/HankandShaidri

[geoff]

James T wrote:
I've read through some archived r.a.p. threads where it was argued
that 16 bit/44.1k had an inaudible affect on music, at least as far as
the cited test went. I understand the merit in having more bits for
multitrack ops, where internal math can lose bits. But now I'm curious
about the widespread 'understanding' that 24-96 is better. Is there no
merit whatsoever in final mix?

That's why internal application stuff is done in a 32 or 64-bit float. If
everything stayed at 16 bits (or better 24) then bits would be being
truncated all over the place.

geoff

[geoff]

William Sommerwerck wrote:
To my ears, Blu-ray audio sounds noticeably better than the common
run of CD sound. However...

Blu-ray audio disks are audiophile recordings, and may use better
mics, fewsd\\er electronics, less processing, etc, than Compact Disk
recordings. At the moment, I cannot say with any objectivity that
"high resolution" recordings are definitely superior. In theory,
there's no reason for them being better, as 16/44 presumably records
everything that's audible.

There is some suggestion that the difference is only in different mastering.

geoff

[geoff]

Sean Conolly wrote:
"James T" wrote in message


Higher bit rates have a real benefit, but for me higher sample rates
do not. I have so many other factors that audibly compromise my
recordings, sample rate is the least of my concerns.


Sample rate essentially *is* bit rate, in linear PCM, which is what we end
up recording and playing back for the most part. Or did you mean
"bit-depths" ?

geoff

[Randy Yates[_2_]]

(hank alrich) writes:

Randy Yates wrote:

James T writes:

I've read through some archived r.a.p. threads where it was argued
that 16 bit/44.1k had an inaudible affect on music, at least as far as
the cited test went. I understand the merit in having more bits for
multitrack ops, where internal math can lose bits. But now I'm curious
about the widespread 'understanding' that 24-96 is better. Is there no
merit whatsoever in final mix?

James,

I would not say "there is no merit whatsoever in the extra range," but
I do think the extra cost to convert to such a medium would far, far
outweigh any advantages.

I think that concern made more sense when storage space was much more
expensive. Granted if one is working with large numbers of tracks it
could still be a constraint. Now with drives costing a little as they do
I no longer think twice about using higher rates.

At this point with the landscape littered with AVID/Digi HD192's, Metric
Halo ULN-8's, and so many similar;y spec'd convertor systems, there's a
lot of 24/192 going down, even where track counts are many dozens.

Hi,

I have no argument (and no experience either!) in the multitrack realm.
I should have been more explicit in my post - my fault.

When James said "in the final mix," I took him to essentially be asking
the question, "Would a new 24 bit 96 kHz storage and delivery medium be
a good thing?" I.e., should we replace CDs with some new 24/96 format?

I certainly see the advantage in 24 bits during recording and mixing. 96
kHz - ummm, not really. At least not over 48 kHz. But my initial comment
was just about the extra bits and under the assumption we were talking
about a delivery medium.
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

[Roy W. Rising[_2_]]

James T wrote:
I've read through some archived r.a.p. threads where it was argued
that 16 bit/44.1k had an inaudible affect on music, at least as far as
the cited test went. I understand the merit in having more bits for
multitrack ops, where internal math can lose bits. But now I'm curious
about the widespread 'understanding' that 24-96 is better. Is there no
merit whatsoever in final mix?


Through the years of emerging technology I have concluded that usually it
is possible to store more information than current reproducers can recover.
There's more in the groove of an old 'record' than we could play/hear with
a sharpend nail hooked to a megaphone. Audiotape playback heads with
ever-decreasing gap sizes have discovered forgotten content in early tapes.
Videotape has demonstrated a similar evolution.

Loudspeaker accuracy's evolution lags, but the ability for us to perceive
the progress continues to improve.

I think the concept continues to apply in the digital world. I tend to
favor 24/96 tracking with the best upconverters, followed by mixing and
mastering at 24/96 with downsampling to the delivery specs as needed.
Somewhere down the road the new converters and loudspeakers will reveal
things we can not appreciate right now.

--
~ Roy
"If you notice the sound, it's wrong!"

[Trevor]

"Les Cargill" wrote in message
...
24 bit depth is good for tracking because of the additional 8 bits
headroom.

There's no converter ever made that will give 8 bits extra headroom, and
never will be even with cryogenic cooling. Of course 3 or 4 bits can still
be worthwhile.

Trevor.

[Trevor]

"Roy W. Rising" wrote in message
...
Through the years of emerging technology I have concluded that usually it
is possible to store more information than current reproducers can
recover.
There's more in the groove of an old 'record' than we could play/hear with
a sharpend nail hooked to a megaphone.

Since it was originally recorded to a disk master with a "sharpened nail" as
well, any extra "information" is likely to be just as innacurate.


Audiotape playback heads with
ever-decreasing gap sizes have discovered forgotten content in early
tapes.
Videotape has demonstrated a similar evolution.

Loudspeaker accuracy's evolution lags, but the ability for us to perceive
the progress continues to improve.

Actually most peoples hearing continues to get *worse*.


I think the concept continues to apply in the digital world. I tend to
favor 24/96 tracking with the best upconverters, followed by mixing and
mastering at 24/96 with downsampling to the delivery specs as needed.
Somewhere down the road the new converters and loudspeakers will reveal
things we can not appreciate right now.


Perhaps, when bionic "ears" improve beyond current human hearing
capabilities I guess.

Trevor.

[Trevor]

"William Sommerwerck" wrote in message
...
To my ears, Blu-ray audio sounds noticeably better than the common run of
CD
sound. However...

Blu-ray audio disks are audiophile recordings, and may use better mics,
fewsd\\er electronics, less processing, etc, than Compact Disk recordings.
At the moment, I cannot say with any objectivity that "high resolution"
recordings are definitely superior. In theory, there's no reason for
them being better, as 16/44 presumably records everything that's audible.

Simple to resample those Blu-Ray recordings to 16/44 to compare them
properly. I can't hear a difference, and never met anyone who *claims* they
can who could prove it.
You may be the exception of course, but I wouldn't bet money on it. :-)
Tracking at 24/96 is another matter of course.

Trevor.

[Trevor]

"Les Cargill" wrote in message
...
From a standpoint of "what do I fix first", any 44.1/16 bit system
is probably much better than any other element in the system.

"Probably"? Well I guess some esoteric cables are better anyway :-)
(yes it's a joke, even some power amps these days are as good, and all those
are usually *far* better than human hearing.)

Trevor.

[Trevor]

"geoff" wrote in message
...
William Sommerwerck wrote:
To my ears, Blu-ray audio sounds noticeably better than the common
run of CD sound. However...

Blu-ray audio disks are audiophile recordings, and may use better
mics, fewsd\\er electronics, less processing, etc, than Compact Disk
recordings. At the moment, I cannot say with any objectivity that
"high resolution" recordings are definitely superior. In theory,
there's no reason for them being better, as 16/44 presumably records
everything that's audible.

There is some suggestion that the difference is only in different
mastering.


And easily proved, since only a moron would compare two different releases,
rather than simply resampling from the better one to eliminate that problem.
No shortage of morons of course. :-(

Trevor.

[Trevor]

"hank alrich" wrote in message
...
At this point with the landscape littered with AVID/Digi HD192's, Metric
Halo ULN-8's, and so many similar;y spec'd convertor systems, there's a
lot of 24/192 going down, even where track counts are many dozens.

I love 24/192 when making measurements, but have absolutely no use for it in
recording music. Anyone who thinks they do is kidding themselves. But not
much harm in having vast levels of overkill any more.

Trevor.

[Trevor]

"Randy Yates" wrote in message
...
When James said "in the final mix," I took him to essentially be asking
the question, "Would a new 24 bit 96 kHz storage and delivery medium be
a good thing?" I.e., should we replace CDs with some new 24/96 format?

We already have Blu-Ray audio at higher bit rates than CD. They haven't
"replaced" CD yet of course, because most people realise it is an answer to
a problem they don't really have.

Trevor.

[James T]

On Sat, 16 Jun 2012 17:37:34 -0400, Randy Yates
wrote:

Consider this thought experiment:

A keyboardist uses his 24-bit Korg SuperHighHD-1024 synth and a
two-track digital recorder to record two tracks at 24 bits perfectly.
During the first minute of his composition, the average power output
is -60 dBFS, then the finale crescendos to an average of -20 dBFS with
peaks near full-scale.

Now convert that to 16-bit using your favorite noise-shaping
algorithm, and play it back using ear buds in a quiet room such that
an ambient level 30 dBA is achieved with the ear buds in. Set levels
so that the recording's -60 dBFS section comes in at 60 dBA over the
ear bugs.

The full-scale finale would then translate to around 100 dBA SPL
average with peaks around 120 dBA - loud but within reason.

With ambient noise 10 dB quieter than the quantization noise, I think
many folks would be able to hear the quantization noise in the initial
quiet passage.

So, yeah, in those near-laboratory conditions and this very strange
composition, 24 bits would give an advantage, I believe. Otherwise, in
the vast, vast majority of practical scenarios, I believe there are very
very very few people who could tell the difference.

Thanks, Randy. A very thoughtful thought experiment. :-)

[Trevor]

"Mike Rivers" wrote in message
...
On 6/16/2012 2:42 PM, James T wrote:
I understand the merit in having more bits for
multitrack ops, where internal math can lose bits.

Just what is it that you understand about this? Why do you think that
internal math can lose bits in a multitrack recording? Can you explain?

Well if you ever save rendered files at lower bit depths, and do further
editing, you add more rounding errors. Perhaps that's what he meant? Of
course one can actually record at lower bit depths and still save at higher
rates before any non EDL editing too.

Trevor.

[geoff]

Trevor wrote:
"Randy Yates" wrote in message
...
When James said "in the final mix," I took him to essentially be
asking the question, "Would a new 24 bit 96 kHz storage and delivery
medium be a good thing?" I.e., should we replace CDs with some new
24/96 format?

We already have Blu-Ray audio

.... and DVD for years with that capability.

geoff

[James T]

On Sat, 16 Jun 2012 22:30:49 -0400, Randy Yates
wrote:

When James said "in the final mix," I took him to essentially be asking
the question, "Would a new 24 bit 96 kHz storage and delivery medium be
a good thing?" I.e., should we replace CDs with some new 24/96 format?

I was posing a theoretical question, so in practical terms, it does
amount to "If you picked humans in the top 5% of audio acuity, would
the 24/96 format be an improvement for them?"

I certainly see the advantage in 24 bits during recording and mixing. 96
kHz - ummm, not really. At least not over 48 kHz. But my initial comment
was just about the extra bits and under the assumption we were talking
about a delivery medium.

Your reply was very relevant re the bit-depth side of the question.

[Trevor]

"geoff" wrote in message
...
Trevor wrote:
"Randy Yates" wrote in message
...
When James said "in the final mix," I took him to essentially be
asking the question, "Would a new 24 bit 96 kHz storage and delivery
medium be a good thing?" I.e., should we replace CDs with some new
24/96 format?

We already have Blu-Ray audio

... and DVD for years with that capability.

Right and SACD, which at least proves the point that you can fool some of
the people all the time, but not all the people all of the time. :-)

Trevor.

[James T]

On Sat, 16 Jun 2012 18:26:21 -0400, Mike Rivers
wrote:

On 6/16/2012 2:42 PM, James T wrote:
I understand the merit in having more bits for
multitrack ops, where internal math can lose bits.

Just what is it that you understand about this? Why do you
think that internal math can lose bits in a multitrack
recording? Can you explain?

I was just referring to scaling ops, etc. but of course those are
probably done by converting to floating point in most cases.

[James T]

On Sat, 16 Jun 2012 18:41:54 -0400, Mike Rivers
wrote:

On 6/16/2012 5:10 PM, James T wrote:

I have presumed (perhaps wrongly) that
the question has been thoroughly explored by those with greater means
than my own.

There have been demonstrations and presumably well conducted
tests that prove both ways. It all boils down to human
hearing, and humans all hear differently. And equally
important, all perceive the same thing differently.

Mike, Do you happen to have any references to those tests? That's what
I'm trying to find.

Why do you care?

?? Primarily a theoretical question. I've been curious about the
prevalence of 24/96 and associated controversy. I haven't found much
info that addresses that conclusively.

Several years ago my friendly local dealer set up a
demonstration. He set up a singer/guitarist with a high
quality stereo mic setup (Senheisser MKH-800s, I recall, one
of the first honest mics that demonstrated usable response
up to 40 kHz ("the one to use if you're going to record at
2x sample rate). He split the preamp output three ways using
a high grade Coleman switch box. One path was direct from
the preamp to the powered speakers (these were PMC, pretty
high grade stuff, too), one was through a pair of Lavry Blue
converters back to back running at 44.1 kHz, the other was
through another pair of the same converters running at 96
kHz. I believe both were using 24-bit resolution, or at
least both were the same.

All three sounded different. I, and most of the listeners,
had a preference for the straight-through path, some
preferred the 44.1 kHz path, some preferred the 96 kHz path,
but nobody would have always picked one sample rate over the
other every time. These were pro engineers, which doesn't
necessarily make them any more qualified than anyone else,
but they were probably more accustomed to listening for
small differences.

And they'd have an idea what to listen for. Useful info, Mike.

Also wondering how transient and positional info relate, since as far
as I know, the relevant refs to Nyquist cover capture of static
waveform data. I haven't seen much tech info on auditory positional
cues and such. Nor do I know if the tests above used source material
that would be likely to expose deficiencies there.

I think I read something about positional cues, but mostly
it was about why sample rate doesn't improve it.

If you remember where you saw that, please post!

[James T]

On Sat, 16 Jun 2012 18:30:36 -0400, Mike Rivers
wrote:

On 6/16/2012 3:40 PM, Scott Dorsey wrote:

Try it for yourself and see.

Just because he can't hear a difference doesn't mean there
isn't one.

Right, which is why I wouldn't be a great test subject for my own
question. My midrange hearing is fine--very sensitive in fact. But my
high and low range is not what it used to be.

To some, the issue is that if it's theoretically better, if
there's the slightest possibility that someone might be able
to detect an improvement, now or in the future, than it's
worth doing. But I'm with you - good 44.1 kHz 16-bit
recordings sound just fine, and bad recordings won't be
improved by jacking up the sample rate or storing more bits.

"Bad recordings" is a separate issue though. To take it another step,
you could say that a great recording won't improve badly composed
music. :-) But I'm just thinking in theoretical domain, so the music
was written by the best composer ever, and of course the recording was
done by the cream of r.a.p.

[William Sommerwerck]

"geoff" wrote in message
...
William Sommerwerck wrote:

To my ears, Blu-ray audio sounds noticeably better than the common
run of CD sound. However...

Blu-ray audio disks are audiophile recordings, and may use better
mics, fewsd\\er electronics, less processing, etc, than Compact Disk
recordings. At the moment, I cannot say with any objectivity that
"high resolution" recordings are definitely superior. In theory,
there's no reason for them being better, as 16/44 presumably records
everything that's audible.

There is some suggestion that the difference is only in different
mastering.

That's only one possible explanation. There are others.

Could you define what you mean by "mastering"?