Square wave / transient response reproduction
Bill Poletti
Hi all,
In recognition of the very low distortion of digital, I'd appreciate
it if a digital expert would be kind enough to try a little
experiment....
Would someone please run a 12 khz square wave into a A to D converter
and run ots output into a D to A converter. Please run the resulting
analog signal into a scope and distortion analyzer. I know I'd be
interested in the results.
After that, it might be interesting to try different frequencies: 10
khz, 16 khz, 5 khz, and others. Since square wave reproduction help
define transient response characteristics, it might be interesting to
hear about the transient response characteristics of the current state
of the art digital.
And please, no theory. let's just deal with the actual reproduction.
Thanks,
[I asked Bill for clarification before approving the above; here's his
reply. -- rgd ]
The point is that reproduction of music (arguably made up of mostly
transients) is more distorted than claimed by the promoters of
digital. The arguments put forth regarding "perfect sound forever"
and "vanishingly low distortion" seem to break down when it comes to
reproduction of waveforms associated with music.
One of the points of discussion within the analog vs. digital
discussion is that digital follows the signal closer that other forms
of source. Not that I'm saying that LP is better than CD or vice
versa. The course I'm suggesting has to do with increased sampling
rates and improvements in digital reproduction.
Bill
Tom Morley
In article <59pi89$j...@tolstoy.lerc.nasa.gov>, bi...@i1.net (Bill Poletti)
wrote:
> In recognition of the very low distortion of digital, I'd appreciate
> it if a digital expert would be kind enough to try a little
> experiment....
> Would someone please run a 12 khz square wave into a A to D converter
> and run ots output into a D to A converter. Please run the resulting
> analog signal into a scope and distortion analyzer. I know I'd be
> interested in the results.
>
Ignoring for a second distortion products the non-idealness of a
physical AD/DA --- I think is quite clear what you would get.
Bracketting the audibility issues, provided the digital filter does
not do something funny (as some do....) you would get a classical
picture of a band limited square wave. Let's see.. how to do this in
ASCII?
/\/\-----/\/\
o o
o o
o o
o o
I.e., "ringing" on both sides of the top (and bottom). Also the left
and right "vertical" parts will not be quite vertical. This is known
in the Fourier analysis world as the Gibbs phemonena.
Distorion? (square wave minus above wave) has frequency components
only above 22.1 kHz.
If the frequency is high enough ( > 12.05 kHz), then the AD-> DA
process, cannot,of course distinguish a squre wave from sine wave.
This is well known and pictures can be seen in most any undergradute
textook on Boundary Value Problems. (I'm parial to Powers's book,
becouse he refernces me ..)
Listening to "The Life of Mary" Cantigas for the Feasts of Holy
Mary by Alfonso X "The Wise" 1211-1284, Ediardo Paniagua --
Sony S2K 62284
Tom Morley | I shall now therefore humbly
mor...@math.gatech.edu | propose my own thoughts, which
tmo...@bmtc.mindspring.com | I hope will not be liable to
| the least objection -- J. Swift
Richard D Pierce
In article <59pi89$j...@tolstoy.lerc.nasa.gov>,
Bill Poletti <bi...@i1.net> wrote:
>In recognition of the very low distortion of digital, I'd appreciate
>it if a digital expert would be kind enough to try a little
>experiment....
>
>Would someone please run a 12 khz square wave into a A to D converter
>and run ots output into a D to A converter. Please run the resulting
>analog signal into a scope and distortion analyzer. I know I'd be
>interested in the results.
Let's generalize the experiment (and I have actually performed the
experiment in such a case):
Let's put a 12 kHz square wave into ANY system whose bandwidth is less
than 30 kHz, that is ANY system. What comes out is a 12 kHz sine wave.
I put a 12 kHz square wave into a DAT deck, out comes a nice 12 kHz
sine wave. I put the same 12 kHz square wave into a Revox 1/2 track
deck, guess what: no sqaure wave comes out: out comes a sine
wave. Noisier, maybe, but a sine wave nonetheless.
>After that, it might be interesting to try different frequencies: 10
>khz, 16 khz, 5 khz, and others. Since square wave reproduction help
>define transient response characteristics, it might be interesting to
>hear about the transient response characteristics of the current state
>of the art digital.
>
>And please, no theory. let's just deal with the actual reproduction.
Fine, but if you are going to point a finger at digital for not
reproducing a 16 kHz square wave, you had better be prepared to point
that same figure at EVERY available reproduction system around,
digital, analog or potherwise, you had better be prepared to point it
at EVERY studio microphone in existance, because the limitation you
observe IS NOT A PROPERTY OF DIGITAL, IT'S A PROPERTY OF ANY
BANDWIDTH-LIMITED SYSTEM IN EXISTANCE.
In order to produce a square wave at 16 kHz, the badwidth has to be
exceptionally wide. It's simply because a square wave is the sume of a
series of harmonically related sine waves extending over a very wide
bandwidth. Specifically, a 16 kHz square wave has 1 part sine at 16
kHz, 1/3 part sine at 48 kHz, 1/5 part sine wave at 80 kHz, 1/7 part
sine at 112 kHz, 1/9 part sine wave at 144 kHz, 1/11 part sine wave at
176 kHz and son on, literally to infinity. The relation is:
1 1 1 1 1
F(t) = - sin(ft) + - sin(3ft) + - sin(5ft) + - sin(7ft) + - sin(9ft) + ...
1 3 5 7 9
If your system (and it, once agains, makes NO difference if that
system is analog or digital) cannot pass the total of those
components, it cannot pass a square wave. And NO music reproduction
system can do that.
>The point is that reproduction of music (arguably made up of mostly
>transients) is more distorted than claimed by the promoters of
>digital.
The further point is that reproduction of music (not necessarily made
up mostly of transients) is more distorted by ALL reproduction
systems.
>The arguments put forth regarding "perfect sound forever"
For once and for all, PUHLEESE, that claim was made by marketing
people. Leave it alone, for god's sake.
>and "vanishingly low distortion" seem to break down when it comes to
>reproduction of waveforms associated with music.
There are MANY kinds of distortion. It is demonstrable that distortion
in the amplitude doain is far better in current digital systems than
in pretty much any analog storage medium. Distortions in the frequency
domain, as measured by frequency and phase response, is demonstrably
superior in current digital system than in all but a few analog
systems (if you have data to the contrary, present it for
evaluation). And if you are looking at transient distortion, well,
like it or not, both the theory and practice BOTH demonstrate that
frequency response, phase response and transient response are ALL THE
SAME THING.
If you are going to claim that, because of a limited bandwith, the
transient response of a system is less than ideal, NO ONE will take
issue with the assertion. However, if you are going to take a 16 kHz
sqaure wave through a digital system alone and claim that that
"proves" digital is inferior, then your assertion will be shown to be
incorrect.
Let;'s takes something more realistic. Let's look at a 2 kHz square
wave through a low-end DAT machine and a high-end cassette (amd I have
done both very recently in my lab, so I HAVE done the precise
experiment you are looking for). In the DAT (a standard TASCAM DA-30),
the square wave shows an almost theoretically perfect picture of a
band-limited square wave expected from a 20 kHz system some overshoot
and ringing, nothing other than expected Higgs behavior. The high-end
cassette (a Nakamichi) shows far poorer rise time and a great deal of
droop. It also shows a square wave performance that is VERY level
dependent. At a record level of 0 VU, it's simply awful, and
deteriorates to a lopsided sine wave. AT -20 VU, it's better, but not
as good as the DAT at ALL levels. And, interestingly enough, the
badnwidths of the two machines are nearly identical. The difference is
that tha DAT has about 15 degrees of phase shift at 20 kHz, while the
Nakamichi has several hundred.
Now, look at a 20 Hz square wave (just did that too). IN the DAT, it's
damned near perfect. In the cassette, it;'s not even recoginizable as
anything even approaching a square wave.
>One of the points of discussion within the analog vs. digital
>discussion is that digital follows the signal closer that other forms
>of source.
If the signal is confined to the bandwidth of the device, yes, it
does.
Despite claims of a lot of people, though, a square wave is a poor
signal because it first of all does not give a good handle on
transient response (it real;y isn't a transient signal, it's quite
periodic) and its bandwidth requirements FAR exceed ANY musically
relevant signals.
>The course I'm suggesting has to do with increased sampling
>rates and improvements in digital reproduction.
If that is the claim, that translates simply to wider bandwidth
requirements, plain and simple. If you want to argue for wider
bandwidth, that's an interesting discussion in and of itself. But it
is not a sampling question, it's a bandwidth question. It's not a
digital vs analog question, it's a bandwidth question.
--
| Dick Pierce |
| Loudspeaker and Software Consulting |
| 17 Sartelle Street Pepperell, MA 01463 |
| (508) 433-9183 (Voice and FAX) |
Bill Poletti
>Distorion? (square wave minus above wave) has frequency components
>only above 22.1 kHz.
>
>If the frequency is high enough ( > 12.05 kHz), then the AD-> DA
>process, cannot,of course distinguish a squre wave from sine wave.
[ cut -- rgd ]
Then perhaps should we start looking at a sampling rate higher than
that currently used for CD?
If we look at the requirements of reproducing transients that occur at
high frequencies (say above 15khz) then should we be looking at
sampling rates that reflect the need to reproduce a reasonable (?)
undistorted signal at that frequency.
To get a sample to comply with the rise, we would need at least a
third order component added in to approximate a square wave at the
limits of audibility. Therefore, if we wanted to reproduce a square
wave at 20khz, we would need a system that is capable of reproducing a
60khz analog signal. That would require a sampling rate of 120khz.
If we really wanted to reduce distortion of transients and be better
able to very accurately follow the original analog signal, we might
consider doubling that rate. Rounding off, that would get us around
256khz, about 6 times that in current digital.
Storage requirements don't seem to be an issue and considering current
processing power, D to A converters would probably be close to the
same price (after covering R&D costs).
Perhaps this might be a path we could head to correct the percieved
shortcomings and criticisms of digital reproduction. Maybe this might
to the answer to the criticisms put forth by some favoring analog.
Bill
Steven Abrams
bi...@i1.net (Bill Poletti) writes:
> Would someone please run a 12 khz square wave into a A to D converter
> and run ots output into a D to A converter. Please run the resulting
> analog signal into a scope and distortion analyzer. I know I'd be
> interested in the results.
Whoever does this experiment please do one other thing.
Take your original "12kHz square wave" and process it through a
low-pass filter with a cutoff at around 20kHz. Pass this through a
'scope, too, and a frequency analyzer. (Is that what you mean by a
"distortion analyzer?")
You *should* notice that there is no difference; both will look like a
12kHz SINE WAVE. If you band-limit a 12kHz square wave to anything
under 24kHz, you end up with a 12khz sine. If there is a difference,
your AD/DA chain, your low-pass filter, or your scope is broken. This
is an easy one.
You question indicates a lack of understanding of some basic
principles. If you have a system band-limited to 20-20kHz, be it
analog or digital, and you try to pump a 12kHz square wave through
that system, it ain't gonna work. You'll end up with a 12kHz sine
wave, because when all of the harmonics which make it look square are
outside of the bandwidth of the system. Get rid of them, and you're
left with only the fundamental. Which is a sine wave.
> After that, it might be interesting to try different frequencies: 10
> khz, 16 khz, 5 khz, and others. Since square wave reproduction help
> define transient response characteristics, it might be interesting to
> hear about the transient response characteristics of the current state
> of the art digital.
The classic mistake you are making is that 'frequency response'
defines transient response. Square waves are made up of combinations
of sine waves. Not just in theory, but in practice.
> The point is that reproduction of music (arguably made up of mostly
> transients) is more distorted than claimed by the promoters of
> digital. The arguments put forth regarding "perfect sound forever"
> and "vanishingly low distortion" seem to break down when it comes to
> reproduction of waveforms associated with music.
Flush the perfect sound forever comments down the toilet; it was made
by marketting morons, not engineers.
~~~Steve
--
Steven Abrams abr...@cs.columbia.edu
Living is easy with eyes closed, misunderstanding all you see.
-Lennon/McCartney
brad sanders
bi...@i1.net (Bill Poletti) sez:
>Would someone please run a 12 khz square wave into a A to D converter
>and run ots output into a D to A converter. Please run the resulting
>analog signal into a scope and distortion analyzer. I know I'd be
>interested in the results.
If so, then step to your local library and look through back issues of
AUDIO or Stereophile (or, I'm sure, others). What you propose is a
fairly standard test; I know I've seen square wave tests in these mags
- and I don't even BUY them. I've seen exactly eight issues of
Stereophile in my life, and yet I can recall seeing at least one such
test result.
Look it up: the data's there. If you don't trust it, do the research
yourself; all it takes is a DAT machine and a 'scope. If you wanna get
REALLY fancy, buy a soundcard for your PC and learn to use it.
Oh: and while you're at it, make sure you also have a very good
*analog* deck handy (preferably one considered, in its day, very
"accurate"). And, just for the heck of it, record a signal to DAT,
then transfer to analog tape. Then, do the reverse.
And pay careful attention to the results on the 'scope...
Per the rest of it: Nothing personal here, but I started to reply, and
then just dumped it. I ditched the rest of the reply because, frankly,
this debate is old; it's a dead horse...done to death...pushin' up
daisies...pinin' for the fjords.
It is a late
debate.
And I'm bored with it.
---------------------------------- Inspired by actual events ----
> http://radioactive.home.ml.org <-------------------------------
Richard D Pierce
In article <59uhg9$j...@tolstoy.lerc.nasa.gov>,
Bill Poletti <bi...@i1.net> wrote:
>>Distorion? (square wave minus above wave) has frequency components
>>only above 22.1 kHz.
>>If the frequency is high enough ( > 12.05 kHz), then the AD-> DA
>>process, cannot,of course distinguish a squre wave from sine wave.
>Then perhaps should we start looking at a sampling rate higher than
>that currently used for CD?
Maybe, but then you state:
>Perhaps this might be a path we could head to correct the percieved
>shortcomings and criticisms of digital reproduction. Maybe this might
>to the answer to the criticisms put forth by some favoring analog.
By this, you are then implicitly claiming that analog MUST have a
substantially wider bandwidth than current digital technology.
Sorry, but despite the claims of many, including utterly irrlevant
claims about old CD-4 records, such is absolutely not the case. When
you consider the typical rolloff points and rates of high-quelity
studio microphones (in the realm of at least 2nd order at 16-24 kHz)
AND the rollof rates of record amplifiers in high-quality tape
machines, the head losses, the rollof rate of the playback amps, the
rolloff rate and frequency of high-quality cutter heads and cutter
electronics, the rollof rate due to the stylus-groove resonance, and
so on, you are facing a VERY sloppy very higher-order filter that is
there for LP's at not a significantly different frequency than current
digital. And, furthjer, if you look at the phase response of the
combined effects above 5 kHz, it's simply awful, at the very best,
severl orders of magnitude worse than even the lousiest of digital.
Now, you may claim that, to you, LP is better. That's fine, I won't
argue with you in the least. But if you are going to invoke bandwidth
as the technical explanation, you loose, because the bandwidth is not
in any way significantly better. CD-4 records notwithstanding, you DO
NOT have any substially better bandwidth with LP's, and what bandwidth
improvement you MIGHT havem, its phase response is terrible, it's
frequency response is lousy, in that possibly extended bandwidth (and
we're not talking an octave better, we're talking, AT THE VERY BEST,
1/3 octave better), and the response below 20 kHz is privably much
worse.
People are allowed to prefer whatever they want, but attempting to
advance a superior bandwidth argument as the justification fails
simply becase the the bandwidth IS NOT BETTER in actual practice.
And there are PLENTY of other reasons that are well understood why
some may like one over the other. And that's fine.
--
| Dick Pierce |
| Loudspeaker and Software Consulting |
| 336 Broadway Hanover, MA 02339 |
| (617) 826-4953 (Voice and FAX) |
sdura...@aol.com
In article <59v927$a...@agate.berkeley.edu>, DPi...@world.std.com (Richard
D Pierce) writes:
>Sorry, but despite the claims of many, including utterly irrlevant
>claims about old CD-4 records, such is absolutely not the case. When
>you consider the typical rolloff points and rates of high-quelity
>studio microphones (in the realm of at least 2nd order at 16-24 kHz)
>AND the rollof rates of record amplifiers in high-quality tape
>machines, the head losses, the rollof rate of the playback amps, the
>rolloff rate and frequency of high-quality cutter heads and cutter
>electronics, the rollof rate due to the stylus-groove resonance, and
>so on, you are facing a VERY sloppy very higher-order filter that is
>there for LP's at not a significantly different frequency than current
>digital. And, furthjer, if you look at the phase response of the
>combined effects above 5 kHz, it's simply awful, at the very best,
>severl orders of magnitude worse than even the lousiest of digital.
I remember reading somewhere about certain DAC phase anomalies
resulting from the the fact that raw digital code sequences the stereo
channels successively rather than simultaneously as in analog. Do you
know more about this?
Secondly, when you add up all the distortions occurring in analog
reproduction, it's apparent that extra-22 kHz information is
significantly compromised. All the same, I'd like to see a study that
shows frequency response roll-off rates, noise and distortion spectra
comparing typical analog and digital playback systems in the extra-22
kHz range. There may be differences in the *nature* of information
loss in this range between the two systems that may offer clues as to
why a preference for analog exists among many keen-eared music
listeners.
Siegfried
Richard D Pierce
In article <5a17v9$h...@agate.berkeley.edu>, <sdura...@aol.com> wrote:
>I remember reading somewhere about certain DAC phase anomalies
>resulting from the the fact that raw digital code sequences the stereo
>channels successively rather than simultaneously as in analog. Do you
>know more about this?
It's confusion over what some very early CD players did, where they
used a single DAC for both channels. They cloacked the data into one,
converted it, sent to a left channel sample-and-hold, then did the
same for the right. The result was a 1 sample delay between right and
left on the audio outputs.
Now, that delay is 1/44100 of a second. At a speed of 1129 ft/sec,
that delay would be precisely the same as moving one of the speakers
1129/44100 feet farther away than the other. That's about 1/40th of a
foot, or less than 1/3 inch. If it is claimed that this is audible,
then that claimed must be reconciled with far more innacurate speaker
placement.
In any case, it's hasn't been an issue for a VERY long time, since
nobody uses single DACS and analog multiplexing anymore. The data from
the left is clocked into the left DAC, then the data from the right is
clocked into the right and then both are commanded to convert
simultaneously.
>Secondly, when you add up all the distortions occurring in analog
>reproduction, it's apparent that extra-22 kHz information is
>significantly compromised. All the same, I'd like to see a study that
>shows frequency response roll-off rates, noise and distortion spectra
>comparing typical analog and digital playback systems in the extra-22
>kHz range. There may be differences in the *nature* of information
>loss in this range between the two systems that may offer clues as to
>why a preference for analog exists among many keen-eared music
>listeners.
And one would have to deal with the notion that the differences in the
*nature* of information loss in this range between the two systems is
wat may offer clues as to why a preference for digital exists among
many keen eared music listeners.
And one might also be prepared to accept the possibility that neither
is the case as well.
Bill Poletti
POST
X-Newsreader: WinVN 0.92.6+
In article <59v927$a...@agate.berkeley.edu>, DPi...@world.std.com (Richard D Pierce) says:
>
>In article <59uhg9$j...@tolstoy.lerc.nasa.gov>,
>Bill Poletti <bi...@i1.net> wrote:
>>>Distorion? (square wave minus above wave) has frequency components
>>>only above 22.1 kHz.
>
>>>If the frequency is high enough ( > 12.05 kHz), then the AD-> DA
>>>process, cannot,of course distinguish a squre wave from sine wave.
>
>>Then perhaps should we start looking at a sampling rate higher than
>>that currently used for CD?
>
>Maybe, but then you state:
>
>>Perhaps this might be a path we could head to correct the percieved
>>shortcomings and criticisms of digital reproduction. Maybe this might
>>to the answer to the criticisms put forth by some favoring analog.
>
>By this, you are then implicitly claiming that analog MUST have a
>substantially wider bandwidth than current digital technology.
>
I am suggesting that the original signal - an analog signal, the
movement of air - not comparing digital to analog recording devices.
I am looking for shortcomings in sound reproduction - NOT, as you read
inot my words, a comparison between digital and analog.
>Sorry, but despite the claims of many, including utterly irrlevant
>claims about old CD-4 records, such is absolutely not the case. When
>you consider the typical rolloff points and rates of high-quelity
>studio microphones (in the realm of at least 2nd order at 16-24 kHz)
>AND the rollof rates of record amplifiers in high-quality tape
>machines, the head losses, the rollof rate of the playback amps, the
>rolloff rate and frequency of high-quality cutter heads and cutter
>electronics, the rollof rate due to the stylus-groove resonance, and
>so on, you are facing a VERY sloppy very higher-order filter that is
>there for LP's at not a significantly different frequency than current
>digital. And, furthjer, if you look at the phase response of the
>combined effects above 5 kHz, it's simply awful, at the very best,
>severl orders of magnitude worse than even the lousiest of digital.
>
Though I feel you are wrong in your assessment of equipment quality,
it might be based on the quality of equipment you are using and not
indicative of the equipment being used by others.
>Now, you may claim that, to you, LP is better.
There are no words in this post / thread that indicate that I favor LP
over anything or feel that CD is inferior. My original post and
follow- up are intended to suggest an improvement in music
reproduction.
>That's fine, I won't
>argue with you in the least. But if you are going to invoke bandwidth
>as the technical explanation, you loose, because the bandwidth is not
>in any way significantly better. CD-4 records notwithstanding, you DO
>NOT have any substially better bandwidth with LP's, and what bandwidth
>improvement you MIGHT havem, its phase response is terrible, it's
>frequency response is lousy, in that possibly extended bandwidth (and
>we're not talking an octave better, we're talking, AT THE VERY BEST,
>1/3 octave better), and the response below 20 kHz is privably much
>worse.
>
The observations you state as fact are not. Many systems are capable
of reproducing wider bandwidth than what you claim. Equipment is not
as limited as claimed. Correctly designed equipment is capable of
following a waveform VERY closely. Even speaker induced distortion
isn't quite as bad as you may think.
Upon what are you basing your "1/3 octave better"? If we begin to
take a hard look at transient response, the signal is SIGNIFICANTLY
distorted at the 44.1 khz sampling. Blending of high frequency
harmonics are not completely accurately reproduced IN MUSIC using
current sampling rates.
Has anyone actually tried to use sampling rates above 200khz and taken
a hard look at the results? Have these results been piped through a
system capable of high definition?
>People are allowed to prefer whatever they want, but attempting to
>advance a superior bandwidth argument as the justification fails
>simply becase the the bandwidth IS NOT BETTER in actual practice.
>
Compared to what? Tape cassettes? Upon what are you basing that
statement? Or is this just YOUR opinion?
Again, I feel wide bandwidth is better. Better playback equipment
utilizes wide bandwidth. Steep transients require it.
Distortion-free reproduction of high frequency transients, by
definition requires MUCH more bandwidth than is available in current
implementations of commercial digital sampling / reproduction.
>And there are PLENTY of other reasons that are well understood why
>some may like one over the other. And that's fine.
>
Not "understood" but claimed to be understood. Surely not agreed by
everyone....
And its OK if you're wrong, again, Dick. This is nothing new.
Richard D Pierce
In article <5a19g3$i...@agate.berkeley.edu>, Richard D Pierce
<DPi...@world.std.com> wrote:
> In article <5a17v9$h...@agate.berkeley.edu>, <sdura...@aol.com> wrote:
>> I remember reading somewhere about certain DAC phase anomalies
>> resulting from the the fact that raw digital code sequences the
>> stereo channels successively rather than simultaneously as in
>> analog. Do you know more about this?
> It's confusion over what some very early CD players did, where they
> used a single DAC for both channels.
>
> Now, that delay is 1/44100 of a second. At a speed of 1129 ft/sec,
> that delay would be precisely the same as moving one of the speakers
> 1129/44100 feet farther away than the other. That's about 1/40th of
> a foot, or less than 1/3 inch. If it is claimed that this is
> audible, then that claimed must be reconciled with far more
> innacurate speaker placement.
Sorry, the delay is not 1/44100 of a second, but 1/44100/2 second, or
1/88200 of a second, or slightly more than 1/8 of an inch effective
path length difference.
Again, in any case, NO ONE ever advanced ANY credible evidence that
such a delay has any audible sugnificance. If they do, they must show
that it is possible to locate speakers and ears to similar or better
accuracy.
Robert Orban
In article <59uhhe$g...@tolstoy.lerc.nasa.gov>, abr...@cs.columbia.edu says...
>Take your original "12kHz square wave" and process it through a
>low-pass filter with a cutoff at around 20kHz. Pass this through a
>'scope, too, and a frequency analyzer. (Is that what you mean by a
>"distortion analyzer?")
>You *should* notice that there is no difference; both will look like a
>12kHz SINE WAVE. If you band-limit a 12kHz square wave to anything
>under 24kHz, you end up with a 12khz sine.
To pick a slight nit...square waves have only odd harmonics, so the
correct number is 36kHz, not 24kHz.
Richard D Pierce
Robert Orban
Stewart Pinkerton
sdura...@aol.com writes:
>I remember reading somewhere about certain DAC phase anomalies
>resulting from the the fact that raw digital code sequences the stereo
>channels successively rather than simultaneously as in analog. Do you
>know more about this?
The serial datastream is indeed sequentially encoded and VERY early
single-DAC players sometimes had a one sample left-right phase error
(about the same as moving one speaker closer by about 3mm), BUT the
channels were originally digitised simultaneously and modern players
and DACs (since 1985 or thereabouts) are so designed that this
interchannel timing difference is essentially zero. In other words,
this was never a real problem and hasn't been measurable for a decade.
>Secondly, when you add up all the distortions occurring in analog
>reproduction, it's apparent that extra-22 kHz information is
>significantly compromised. All the same, I'd like to see a study that
>shows frequency response roll-off rates, noise and distortion spectra
>comparing typical analog and digital playback systems in the extra-22
>kHz range. There may be differences in the *nature* of information
>loss in this range between the two systems that may offer clues as to
>why a preference for analog exists among many keen-eared music
>listeners.
You might find it interesting to subscribe to Gabe Wieners mailing
list pro-...@pgm.com where you may read the contributions of many
dedicated professional recording engineers on the subject of whether
96kHz sampling would be beneficial. A sub-text of this discussion is
comments from several highly competent mastering engineers on how
little (if any) signal-related content there ever was on LP above
15kHz, never mind 22! Loud guffaws tend to accompany references to
CD-4, and this is amongst the guys who actually cut records for a
living, not 'armchair quarterbacks'.
--
Stewart Pinkerton | If you can't measure what you're making,
A S P Consulting | how do you know when you've got it made?
(44) 1509 880112 |
"I canna change the laws o' physics" - the other Scotty
sdura...@aol.com
In article <5a19g3$i...@agate.berkeley.edu>, DPi...@world.std.com (Richard
D Pierce) writes:
>In any case, it's hasn't been an issue for a VERY long time, since
>nobody uses single DACS and analog multiplexing anymore. The data from
>the left is clocked into the left DAC, then the data from the right is
>clocked into the right and then both are commanded to convert
>simultaneously.
Thanks for your reply. Couple of thoughts. In order to convert
simultaneously, wouldn't the left channel still need to be held for
1/88,000 of a second? Would a sample-and-hold circuit take care of
this?
Secondly, my understanding of frequency delays suggest phase
distortions may create interference patterns at frequencies *below*
that of the primary. Is this incorrect?
[ cut -- rgd ]
>And one would have to deal with the notion that the differences in the
>*nature* of information loss in this range between the two systems is
>wat may offer clues as to why a preference for digital exists among
>many keen eared music listeners.
Oh, absolutely, yes, that goes without saying. Forgive me for putting
the cart before the horse. Who'd be capable of conducting a careful
study of this?
Siegfried
Richard D Pierce
In article <5a8v5k$h...@tolstoy.lerc.nasa.gov>, <sdura...@aol.com> wrote:
>In article <5a19g3$i...@agate.berkeley.edu>, DPi...@world.std.com (Richard
>D Pierce) writes:
>
>>In any case, it's hasn't been an issue for a VERY long time, since
>>nobody uses single DACS and analog multiplexing anymore. The data from
>>the left is clocked into the left DAC, then the data from the right is
>>clocked into the right and then both are commanded to convert
>>simultaneously.
>
>Thanks for your reply. Couple of thoughts. In order to convert
>simultaneously, wouldn't the left channel still need to be held for
>1/88,000 of a second?
It's an irrelevant issue...
>Would a sample-and-hold circuit take care of this?
Nobody uses sample and hold circuits any more, so it's irrelevant.
It's irrelevant because the order of transmission and relative delays
between channels DURING TRANSMISSION has NO EFFECT upon presentation
out of the DACS as long as the output from the DACS is within
reasonable tolerances. And if there IS a difference between the output
of the DACS, it's a simple delay issue, EXACTLY equivalent to having
your head slightly (on the order of millimeters) closer to one speaker
than the other.
>Secondly, my understanding of frequency delays suggest phase
>distortions may create interference patterns at frequencies *below*
>that of the primary. Is this incorrect?
These are NOT frequency delays: the term itself is somewhat
meaningless. They are simply time delays. And there will be
interference patterns at ANY frequency at ANY time delays between
multiple sources: the presence of an 11 millionths of a second delay
between left and right channels in a stereo system WILL NOT make the
resultant interference pattern any worse in the audio band than
already exists simply because you are already using a stereo speaker
pair.
Those interference patterns are complex and completely unavaoidable
simply because of the very nature of stereo reproduction, delay or
not. The effect of the delay is to simply shift the axis of the
pattern slightly towards the speaker with the earlier signal. Again,
and the point is VERY important, it is EXACTLY equivalent to shifting
the position of that loudspeaker by a couple of millimeters. The delay
DOES NOT appreciably change the pattern at all, it simly turns it
slightly in the room.
In fact, the anomoloies of the pattern are by far controlled by the
design (or misdesign) of the speaker itself and what's placed around
it in the room. The radiation pattern of a single speakers at high
frequencies are, pretty much without exception, god-awful messes
simply as a result of the physical inevitability of cabinets and
drivers and driver separations being large compared to the wavelengths
involved.
Again, the presence of such a delay, which, for the record, DOES NOT
EXIST IN MODERN PLAYERS, does not make those awful patterns any worse
or any better.
To get back to the underlying principles of the original question,
it's very simple: the A/D converters DO NOT convert one channel, then
the other. The capture both channels simultaneously to a very high
accuracy (small fractions of a microsecond). The order and delay
between samples in the transmission and storage media have NO
relevance whatsoever. In fact, you could transmit ALL of the left
channel of a performance on Tuesday, then, the following Saturday,
transmit ALL of the right channel BACKWARDS. And that fact of
traansmission anomoly has NO relevance if, on Sunday, you sit down,
push a button, and have the audio from both channels show up at the
right time. Assembling the data in the right order at the right time
and presenting it to the listener is the job of the DAC, and the order
in the media and on the cables has NO relevance.
jj, curmudgeon and all-around grouch
In article <59uhg9$j...@tolstoy.lerc.nasa.gov> bi...@i1.net (Bill Poletti) writes:
>If we look at the requirements of reproducing transients that occur at
>high frequencies (say above 15khz) then should we be looking at
>sampling rates that reflect the need to reproduce a reasonable (?)
>undistorted signal at that frequency.
Well, only if you are asserting that supra 20khz signals are somehow
perceptable, or that the processing to get down to 20kHz has some
other audible effect.
>Perhaps this might be a path we could head to correct the percieved
>shortcomings and criticisms of digital reproduction. Maybe this might
>to the answer to the criticisms put forth by some favoring analog.
It's always possible, but why do such criticisms often disappear when
some purely digital distortions are added to the digital signal when
listening under controlled conditions?
--
Copyright alice!jj 1996, all rights reserved, except transmission by USENET
and like facilities granted. This notice must be included. Any use by a
provider charging in any way for the IP represented in and by this article
and any inclusion in print or other media are specifically prohibited.
jj, curmudgeon and all-around grouch
In article <5a3s6j$e...@eyrie.graphics.cornell.edu> DPi...@world.std.com (Richard D Pierce) writes:
>Again, in any case, NO ONE ever advanced ANY credible evidence that
>such a delay has any audible sugnificance. If they do, they must show
>that it is possible to locate speakers and ears to similar or better
>accuracy.
Careful, Dick, there is evidence that, in headphones, under some
circumstances, there are subjects who can resolve down to somewhere
around 5 microsecond differences in 20Khz click signals
binaurally. (no intraaural mixing)
Of course, this doesn't say much about any acoustically recorded
signal, I do agree.
Bob Myers
Bill Poletti (bi...@i1.net) wrote:
> I am suggesting that the original signal - an analog signal, the
> movement of air - not comparing digital to analog recording devices.
> I am looking for shortcomings in sound reproduction - NOT, as you read
> inot my words, a comparison between digital and analog.
Ummm, Bill? The orignal signal IS the movement of air - a pressure
wave in a gaseous medium, to be technical about it. But that ISN'T an
"analog" signal.
Once again, "analog" is an ENCODING PROCESS. This word does NOT mean
the same as "continuous" or "time-varying" or any of the other things
that usually (and unfortunately) get caught up with it.
Bob Myers KC0EW Hewlett-Packard Co. |Opinions expressed here are not
O- Workstations Systems Div.|those of my employer or any other
my...@fc.hp.com Fort Collins, Colorado |sentient life-form on this planet.
Craig Luna
>Thanks for your reply. Couple of thoughts. In order to convert
>simultaneously, wouldn't the left channel still need to be held for
>this?
This is taken care of by the SP/DIF input receiver as it decodes the
stream of data and recovers the data and clock. The (the actual dac
not the box with the same name) DACs realize the data for the right
and left channels simultaneously as the master clock(or the
over-sampling clock) triggers the conversion.
Craig Luna
Scott Frankland
Stewart Pinkerton wrote (in response to SDuraybito):
> A sub-text of this discussion is comments from several highly
> competent mastering engineers on how little (if any) signal-related
> content there ever was on LP above 15kHz, never mind 22! Loud
> guffaws tend to accompany references to CD-4, and this is amongst
> the guys who actually cut records for a living, not 'armchair
> quarterbacks'.
That's right, Sig. It's probably most pertinent to look at the output
of musical instruments themselves (e.g., Olson: *Music, Physics, and
Engineering*). Very little fundamental energy is produced by symphonic
instruments above about 4kHz. Above this, all is partials. The energy
of the partials generally falls off with frequency, and is fully
exhausted at about 16kHz. 20kHz is used as an upper limit for audio
equipment because this is the limit of human hearing. But what's the
point if the instruments themselves quit at 16kHz?
~SF~
Steven Abrams
ror...@slip.net (Robert Orban) writes:
> To pick a slight nit...square waves have only odd harmonics, so the
> correct number is 36kHz, not 24kHz.
Well, I've never been accused of thinking too hard while typing here.
It is, afterall, only USENET.
[ See what having a child will do to your perspective on life. Come
on Steve, get a grip. -- jwd ]
jj, curmudgeon and all-around grouch
In article <5ac4u9$g...@canyon.sr.hp.com> bi...@i1.net (Bill Poletti)
writes:
> I am suggesting that the original signal - an analog signal, the
> movement of air - not comparing digital to analog recording devices.
Um, the "movement of air" is actually the movement of many small
particles, each bouncing off the others and the walls (that's what's
called pressure). This "analog" signal has its own noise level, a
noise level not far from the best sensitivity of human hearing, by the
way, and a very wide bandwidth, since each of those small "bounces" in
fact an "impulse" of some amplitude, and what we hear is the sum of
those "impulses".
Now, in the MACROSCOPIC sense, does this matter? Yes. It's a problem
in horns (distortion mechanisms due to pressure not being
approximately constant any more), it can be a problem in some kinds of
tweeters, although not usually any more, and the nonlinearities of air
have more than once socked a box design in the chops (ask Dick about
people desiging boxes with too-small area ports :-).
> I am looking for shortcomings in sound reproduction - NOT, as you
> read inot my words, a comparison between digital and analog.
Well, the first is that stereo is really rotten illusion of the
original soundfield.
> Though I feel you are wrong in your assessment of equipment quality,
> it might be based on the quality of equipment you are using and not
> indicative of the equipment being used by others.
The rolloff of most equipment is rather like what he's stated, Bill,
and most LP's don't have much information PUT ON TO THEM above 15k.
Now, READ OFF of them is another story, you can find a lot of HF
energy simply thanks to grove tracing problems, with any stylus.
> The observations you state as fact are not. Many systems are
> capable of reproducing wider bandwidth than what you claim.
But the energy reproduced wasn't there in the original recording in
most cases, and may or may not have been there in the original venue
(it can go either way).
So, what's the "trVth"? Good question.
Problem is, the only way to find out is to go on experimenting,
changing bandwidths, trying distortion mechanisms at the digital
output, etc. Yes, people are doing that.
sdura...@aol.com
In article <5ah348$3...@eyrie.graphics.cornell.edu>, Scott Frankland
<audi...@ix.netcom.com> writes:
>That's right, Sig. It's probably most pertinent to look at the output
>of musical instruments themselves (e.g., Olson: *Music, Physics, and
>Engineering*). Very little fundamental energy is produced by symphonic
>instruments above about 4kHz. Above this, all is partials. The energy
>of the partials generally falls off with frequency, and is fully
>exhausted at about 16kHz. 20kHz is used as an upper limit for audio
>equipment because this is the limit of human hearing. But what's the
>point if the instruments themselves quit at 16kHz?
Yes, I'm aware of the direct-radiated spectrum from musical
instruments and I agree that there's precious little information by
the time you get to 20 kHz (though I wonder about cymbals). I believe
things get a little more interesting when you consider reverberation
and localization. Fundamentals and harmonics get divided down pretty
quickly from bouncing off surfaces, while still maintaining high
levels of amplitude. The ability to perceive instruments accurately
in the soundfield relates to localization, but of even greater
significance in the perception of an instrument's shape, which, when
you consider the distances involved between instrument and microphone,
certainly point to an angle of resolution that correlates to
frequencies in the top octave and beyond.
In an HP interview with Keith Johnson of Pacific Microsonics, Johnson
has brought up the issue of shotgun studies conducted at Stanford
suggesting the ear has an "early-warning" system that detects supra-20
kHz signals. Of course, this has been roundly refuted here on
r.a.h-e, mostly by the pro-digital crowd. I also recently re-read an
interview with Joachim Gerhard of Audio Physic wherein he suggests
that certain psychoacoustic studies relating to "event identification"
are not always consistent with our traditional understanding of aural
acuity from a strict frequency response/dynamic range perspective.
In that digital has not yet been able to fully set well with me, these
peripheral reports that suggest the need for supra-22kHz information
bear further investigation. What I'd really like to see is a direct
comparison of frequency response, harmonic distortion, noise and other
parameters between analog and digital. There have been several
concurrent releases of the same material on CD and LP (Reference
Recordings comes to mind). Why not measure several of these pairs and
see what shows up?
Siegfried
Vladimir Kuznetsov
> In an HP interview with Keith Johnson of Pacific Microsonics,
> Johnson has brought up the issue of shotgun studies conducted at
> Stanford suggesting the ear has an "early-warning" system that
> detects supra-20 kHz signals. Of course, this has been roundly
> refuted here on r.a.h-e, mostly by the pro-digital crowd.
. . . This is an outrageous distortion of
facts. This discussion of gunshot studies was going here for several
months but nobody saw it. When Siegfrid eventually was asked to give
reference to an article containing results all that he was able to
produce was a reference to somebody who believed that such study
exists. Sounded like urban myth to me. Few other people browsed
through Stanford data bases in a search of this report and found
nothing. And in Stanford they are archiving their papers since 1974.
So the final result was absolute zero. Nobody heard about it, but
Siegfrid tortured everybody with references to this article alleging
it as a proven fact. Now when some time passed, Siegfried obviously
tries to resurrect this report. It was not refuted by "pro-digital
crowd"! Siegfried used unsubstantiated rumors presenting them as
established evidence.
Sigfried, please, apologize and stop misinformation!
vlad
Vladimir Kuznetsov
Nousaine
Mr Meyers notes that "analog" ain't what it used to be....
"Ummm, Bill? The orignal signal IS the movement of air - a pressure
wave in a gaseous medium, to be technical about it. But that ISN'T an
"analog" signal.......Once again, "analog" is an ENCODING PROCESS.
This word does NOT mean the same as "continuous" or "time-varying" or
any of the other things that usually (and unfortunately) get caught up
with it."
It is perhaps unfortunate that a basic technique used to describe a
the digital process is the stair step quantization of a sine wave. A
basic problem is that we tend to confuse the sine wave picture with
the original sound when in fact it just one way to conceptualize the
original "signal." A stair-stepped sine wave graph looks
horrible....but the sine wave picture wasn't the sound in the first
place just another way to think about it. To my way of thinking sound
is a bunch of air molecules being bumped into each other which seems
much more "digital" that analog in nature.
Robert Orban
In article <5ak5u1$l...@canyon.sr.hp.com>, sdura...@aol.com says...
>
>
>Yes, I'm aware of the direct-radiated spectrum from musical
>instruments and I agree that there's precious little information by
>the time you get to 20 kHz (though I wonder about cymbals). I believe
>things get a little more interesting when you consider reverberation
>and localization. Fundamentals and harmonics get divided down pretty
>quickly from bouncing off surfaces, while still maintaining high
>levels of amplitude.
If, by this, you mean that reverberation increases the bandwidth of
the signal, this isn't the case. Reverberation is basically a linear
process, meaning that it doesn't add spectral components to the
sound. In fact, high frequencies usually decay far more quickly than
low frequencies.
The only exceptions you will find to the 'reverb is linear' paradigm
are cases where the reflecting surfaces behave in a non-linear way. A
wood panel that buzzes when hit by strong bass is a good example.
Richard D Pierce
In article <5atorj$j...@eyrie.graphics.cornell.edu>,
Vladimir Kuznetsov <vl...@best.com> wrote:
>In article <5ak5u1$l...@canyon.sr.hp.com>, <sdura...@aol.com> wrote:
>> In an HP interview with Keith Johnson of Pacific Microsonics,
>> Johnson has brought up the issue of shotgun studies conducted at
>> Stanford suggesting the ear has an "early-warning" system that
>> detects supra-20 kHz signals. Of course, this has been roundly
>> refuted here on r.a.h-e, mostly by the pro-digital crowd.
>. . . This is an outrageous distortion of facts. This discussion of
>gunshot studies was going here for several months but nobody saw
>it. When Siegfrid eventually was asked to give reference to an article
>containing results all that he was able to produce was a reference to
>somebody who believed that such study exists. Sounded like urban myth
>to me. Few other people browsed through Stanford data bases in a
>search of this report and found nothing.
In fact, several people have made direct queries to Pacific
Microsonics for references to these alledged gunshot studies, one of
whom, in fact, called PM, asking for the reference. In ALL cases, any
attempts to gain ANY information about these alledged studies have
been either ignored or actively rebuffed by PM.
Seigfried SPECIFICALLY chastised us to get a hold of these studies
from PM. At least 6 attempts were made in good faith to do so. Every
single request resulted in NO information on ANY gunshot studies from
PM.
PM has refused to publish this information or provide ANY references
to this alledged study. We are left with the conclusion that NO SUCH
STUDY EXISTS, and that PM fabricated the information.
The burden is NOW on PM to provide the results, as it ALWAYS was. THEY
made the claim, and despite good faith efforts on the part of others,
no such study has been forthcoming. IUt's now time for Keith Johnson,
PM and their apologist to either put up or retract the claim.
>And in Stanford they are
>archiving their papers since 1974. So the final result was absolute
>zero. Nobody heard about it, but Siegfrid tortured everybody with
>references to this article alleging it as a proven fact. Now when some
>time passed, Siegfried obviously tries to resurrect this report. It
>was not refuted by "pro-digital crowd"!
NOBODY on this group refuted such a study, BECAUSE THERE IS NO SUCH
STUDY TO REFUTE. We ahve asked for the study, for references to the
study, for a citation, ANYTHING. And the PM crowd and their apologist
have utterly failed to respond. It's that simple. There is no
refutation. There is merely a complete lack of such a study from the
very person who claimed the study existed in the first place.
>Siegfried used unsubstantiated rumors presenting them as established
>evidence.
Old news, unfortunately. Sigh. Seigfried digs up the rumor as
"evidence" then utterly fails to provide the "evidence," chastising US
to get the report. we attempt to do so from the very person claiming
the report, and that person provides nothing, absolutely nothing.
It is interesting to note that the last time this discussion came
around in rec.audio.tech, Seigfried himself speculated that the report
simply did not exist.
Martin Leese - OMG
On 2 Jan 1997 14:43:04 -0500 Scott Frankland (audi...@ix.netcom.com) wrote:
...
>> 20kHz is used as an upper limit for audio
>> equipment because this is the limit of human hearing. But what's the
>> point if the instruments themselves quit at 16kHz?
I would suggest an element of caution here. The quoted upper limit of
human hearing is for continuous tones. The perception of transients
is not nearly as well understood.
An unrelated note on the "gunshot" studies. I remember that
discussion and considered it Usenet at its best. Many people
cooperated to track down the study, and so assist the original poster,
but were unsuccessful. A cornerstone of science is the need to get
back to source documents. Unless somebody can point to these, I
suggest the topic is quietly ignored - without source documents for
study it is a dead-end.
I know of only two references on supersonic hearing:
Ohashi, T., Nishina, E., Kawai, N., Fuwamoto, Y. and Imai, H.
"High Frequency Sound Above the Audible Range Affects Brain
Electrical Activity and Sound Perception", 91st AES Convention,
New York, preprint 3207 (Oct. 1991)
Ohashi, T., Nishina, E., Fuwamoto, Y. and Kawai, N.
"On the Mechanism of Hypersonic Effect", Proceedings Int'l
Computer Music Conference, Tokoyo, 432-434 (1933)
I have not read either.
Regards,
Martin
E-mail: mle...@omg.unb.ca
WWW: http://www.omg.unb.ca/~mleese/
______________________________________________________________________
Want to know how Ambisonics can improve the sound of your LPs and CDs?
Read the Ambisonic Surround Sound FAQ. Version 2.7 now on my WWW page.
sdura...@aol.com
In article <5audhg$r...@agate.berkeley.edu>, DPi...@world.std.com (Richard
D Pierce) writes:
>Seigfried SPECIFICALLY chastised us to get a hold of these studies
>from PM. At least 6 attempts were made in good faith to do so. Every
>single request resulted in NO information on ANY gunshot studies from
>PM.
I am sorry. I was unaware of six separate attempts to get more
information from PM. I guess you're right, the ball is in their
court. Even worse, I can't believe everything I read in a magazine
(TAS) that I write for. Such is life...
I find it curious, though, that the literature of aural acuity of
supra-22kHz signals, particularly non-continuous, is very scarce,
indeed. It was a great leap of faith for the digital fathers to
establish CD's bandwidth standard with so few studies available. When
you consider the great number of serious music listeners who can't
quite embrace digital sound, I think it makes sense to get a better
handle on our perception of the top octave and beyond.
Siegfried
jj, curmudgeon and all-around grouch
>I would suggest an element of caution here. The quoted upper limit of
>human hearing is for continuous tones. The perception of transients
>is not nearly as well understood.
Well, the frequency content of transients is well understood. The
non-linear effects on the ear are not that well understood, but one
must consider the level of the stimulus as well as the "impulsive"
nature. One must also realize that people haver used broadband
stimulii in calculating the upper limit of human hearing, but have
gone to tones since it resulted in less masking and a HIGHER answer.
Still, it's not a well-answered question, but I'd suggest that there
is quite a bit more evidence than you seem to mention.
--
Copyright alice!jj 1997, all rights reserved, except transmission by USENET
Lee Meador
Martin Leese - OMG wrote:
> I know of only two references on supersonic hearing:
>
> Ohashi, T., Nishina, E., Kawai, N., Fuwamoto, Y. and Imai, H.
> "High Frequency Sound Above the Audible Range Affects Brain
> Electrical Activity and Sound Perception", 91st AES Convention,
> New York, preprint 3207 (Oct. 1991)
>
> Ohashi, T., Nishina, E., Fuwamoto, Y. and Kawai, N.
> "On the Mechanism of Hypersonic Effect", Proceedings Int'l
> Computer Music Conference, Tokoyo, 432-434 (1933)
Have these four people actually been studing this topic for 58 years?
Additionally, I would suspect that a computer music conference was
sparsely attended in 1933. ( :') <-- note tongue in cheek)
[ Sheesh, everyone is a critic these days. -- jwd ]
-- Lee Meador
Martin Leese - OMG
On 9 Jan 1997 14:22:07 -0500 jj, curmudgeon and all-around grouch (j...@research.att.com) wrote:
>> In article <5audbr$r...@agate.berkeley.edu> mle...@omg.unb.ca writes:
>> >I would suggest an element of caution here. The quoted upper limit of
>> >human hearing is for continuous tones. The perception of transients
>> >is not nearly as well understood.
>> Well, the frequency content of transients is well understood. The
>> non-linear effects on the ear are not that well understood, but one
>> must consider the level of the stimulus as well as the "impulsive"
>> nature. One must also realize that people haver used broadband
>> stimulii in calculating the upper limit of human hearing, but have
>> gone to tones since it resulted in less masking and a HIGHER answer.
>> Still, it's not a well-answered question, but I'd suggest that there
>> is quite a bit more evidence than you seem to mention.
The original post for this thread was about what happens when a square
wave is passed through an AtoD / DtoA chain of converters. It has
digressed into human perception, specifically supersonic hearing.
Supersonic hearing keeps cropping up and I would like to try to nail
this down a little firmer.
Digital audio reproduction systems are band limited. (Yes, I know, so
are analogue, but lets just stick with digital.) Because of the
sampling theorum, this band limiting is the same for the reproduction
of continuous tones and for transients.
The human ear/brain seems to use different mechanisms to perceive
continuous tones and transients. While each of these mechanisms must
be band limited, we cannot say that the limits are the same.
Is the above a reasonable summary of what is going on? Are there any
estimates for the band limiting of the perception of transients?
(I have chosen to ignore the fact that recording systems are also band
limited and if you don't record it, it ain't there to reproduce.)
conn...@gol.com
sdura...@aol.com wrote:
> DPi...@world.std.com (Richard D Pierce) writes:
> >Seigfried SPECIFICALLY chastised us to get a hold of these studies
> >from PM. At least 6 attempts were made in good faith to do so. Every
> >single request resulted in NO information on ANY gunshot studies from
> >PM.
> I am sorry. I was unaware of six separate attempts to get more
> information from PM. I guess you're right, the ball is in their
> court.
I have talked to Keith Johnson in the past, and I think that his
tenure at Stanford predates 1974 by quite a few years (more likely
time would the late 50s or early 60s). Incidentally, knowing how many
businesses operate and prioritize issues (as opposed to educational or
research institutions), another possibility is that PM simply doesn't
care what rahe thinks. They seem to being doing fairly well
commercially, and they are gaining the support of quite a few big and
influential companies, so I would be quite surprised if they felt any
need whatsoever to justify themselves to an internet newsgroup.
> I find it curious, though, that the literature of aural acuity of
> supra-22kHz signals, particularly non-continuous, is very scarce,
> indeed. It was a great leap of faith for the digital fathers to
> establish CD's bandwidth standard with so few studies available. When
> you consider the great number of serious music listeners who can't
> quite embrace digital sound, I think it makes sense to get a better
> handle on our perception of the top octave and beyond.
I have heard of similar studies from Europe purporting to show that
the human ear can hear transients well above 20kHz, although I do not
know if they had anything to do with gunshots. I believe that the
Mangar bending-wave transducer from Germany uses some of this
research. I may try to see if I can locate a copy of this study,
although it will likely take some time (as I might have to go to
Germany to look for it).
No promises, though. This is mainly for my own personal curiosity.
Best regards
Jonathan C
--
Without comprehension, knowledge is nothing.