MOTM

I love reading product comments and reviews like these.  
I belong to a wine tasting club.  I often tease other members by 
substituting subjective audiophile terms into wine reviews.
If you substituted a few words in these audiophile comments they 
could easily be wine reviews as well.
The funny thing is that most of these comments don't really say 
anything, but everyone pretends to know what they mean.

My favourite equipment review fiasco happened a number of years ago when
Julian Herst decided to debunk the esoteric cable test done in "Audiophile"
in an another test he performed for "Stereo Review". Anybody remember this
one?

In Julian's test, auto jumper cables from K-Mart outperformed every single
other cable tested, including those "golden ears only" $450 per meter
jobbies...

Nasty letters went back and forth in the two magazines for MONTHS after
that....made for some damn fine reading!



Les

>My favourite equipment review fiasco happened a number of years ago when
>Julian Herst decided to debunk the esoteric cable test done in "Audiophile"
>in an another test he performed for "Stereo Review". Anybody remember this
>one?

You are now preaching to the choir. Boy, do I remember that. I was ROTF.

>In Julian's test, auto jumper cables from K-Mart outperformed every single
>other cable tested, including those "golden ears only" $450 per meter
>jobbies...

The test I remember (mid-80's) was between 30' of zip cord and Monster
Cable. They may have also used some smaller gauge wire just for laughs. It
has been quite a while. Maybe this is the same test, maybe they did it more
than once.

As I recall the couple of pros on the test panel could actually hear a
"difference" about 60% of the time (or something similarly iffy) when they
used noise bursts and test tones, but when they played any sort of actual
MUSIC nobody could hear the difference.

 I also recall that they did the same thing with a range of amplifiers, and
again with a range of CD players--larger spread but overall similar
results. The cheap stuff did surprisingly well with actual music.


The point is that if you're using your synth in a quiet studio you can hear
crap going on, but that's because of the austere circumstances.

>Nasty letters went back and forth in the two magazines for MONTHS after
>that....made for some damn fine reading!

My favorite: "I've experienced a 50% increase in sound quality since using
Monster Cable!". I just love when laymen use such highly technical
terminology.
-- 

Terry Bowman, KA4HJH
"The Mac Doctor"

"If The Amazing Kreskin was a mentalist then what the hell is a
fundamentalist?"

I remember the cable tests quite well.
More recently there have been some double blind tests of material 
recorded at different sample rates.  I believe the panel consisted 
of highly regarded mastering engineers.

Although these individuals had no trouble recognizing the higher 
quality of 24 bits over 16 bits, they could not hear an appreciable 
difference between the standard 44.1 kHz sample rate used by CDs and 
the higher 96 kHz and even 192 kHz sample rates.

I have read some equipment reviews which raved about the improved 
quality of higher sample rates, but these were not double blind 
studies.  Personally I think the reason some higher sample rate 
converters sound better is the quality of the filters, not the rate 
of conversion.  For instance my Apogee Special Edition converters 
sound way better than my Digidesign converters both at 44.1 kHz.

The Nyquist criterion states that any sample rate of more than twice 
the highest frequency is all thats needed for accurate reproduction, 
and I refuse to believe that there are that many engineers who can 
hear over 20 kHz.  I sure can't now and I never could.

:> The Nyquist criterion states that any sample rate of more than twice
:> the highest frequency is all thats needed for accurate reproduction,
:> and I refuse to believe that there are that many engineers who can
:> hear over 20 kHz.  I sure can't now and I never could.


Isn't there some research to suggest that there is a psycho-acoustic effect
caused by the "outside" frequencies that ends up affecting the audible
portion of the material as the different frequencies interact with each
other?

I'll have to go look this up. Don't remember exactly...

I must NOT be a "golden ears" type myself. I used to have a client that
specialized in pipe organ recordings, and he'd bring his own digital (?)
cables (2 feet long - he paid close to $500 for them. I don't know what the
hell they were!) to make the digital transfers from his DAT master, into my
workstation for editing and back out to his souped-up DAT machine again. I'd
have to tear half the studio apart to get it wired *just* the way he wanted.
He SWORE he could tell the difference....

When I used my cables (not exactly Radio Shack phono cords.....), I got
comments like:

"There's not enough 'air' now"
"The soundstage has collapsed somewhat"

blah, blah.... but...he always *knew* which cables were being used at the
time.

Only way I could keep him coming back for more editing was to go "Yea, I
think you're right....", otherwise, I wasn't worthy of editing his stuff in
the first place.

I lost him when I decided to try a blind "cable test" and he got it wrong.
Swore I was lying to him about which was which and never came back. Oh
well.....

Sheesh...

> The Nyquist criterion states that any sample rate of more than twice
> the highest frequency is all thats needed for accurate reproduction,

The Nyquist theorem never stated anything about accuracy, only that to
reproduce any given frequency, it must be at least half the sample rate.
If you sampled a 10khz sine wave at 20khz, it would become a square wave -
certainly NOT accurate. Most filtration in CD players would wind up
rounding the edges off anyway, but it's still not accurate - as compared
to the source.

> and I refuse to believe that there are that many engineers who can
> hear over 20 kHz.  I sure can't now and I never could.

I hear people state this so often, but the higher sampling rate has
nothing to do with a desire to hear higher frequencies, but rather to keep
the higher frequency signals more true to the original.

-->Neil

-------------------------------------------------------------------------------
Neil Bradley            In the land of the blind, the one eyed man is not
Synthcom Systems, Inc.  king - he's a prisoner.
ICQ #29402898

> :> and I refuse to believe that there are that many engineers who can
> :> hear over 20 kHz.  I sure can't now and I never could.
> Isn't there some research to suggest that there is a psycho-acoustic effect
> caused by the "outside" frequencies that ends up affecting the audible
> portion of the material as the different frequencies interact with each
> other?

Yes, it's called "listener fatigue". ;-) Lack of upper band harmonics
can contribute to it.

> cables (2 feet long - he paid close to $500 for them. I don't know what the
> hell they were!) to make the digital transfers from his DAT master, into my
> workstation for editing and back out to his souped-up DAT machine again. I'd
> have to tear half the studio apart to get it wired *just* the way he wanted.
> He SWORE he could tell the difference....

*DIGTIAL* Transfers? Now I've heard EVERYTHING!

This reminds me of a clueless individual who thought that buying 5 foot
MIDI cables would have a noticeable effect on MIDI delay over 10 foot
cables. Little did this individual know that for even a single sample @
44.1khz delay to be induced by the wire it'd have to be about 2800 feet
long. ;-)

-->Neil

-------------------------------------------------------------------------------
Neil Bradley            In the land of the blind, the one eyed man is not
Synthcom Systems, Inc.  king - he's a prisoner.
ICQ #29402898

> The Nyquist theorem never stated anything about accuracy, only that to
> reproduce any given frequency, it must be at least half the sample rate.
> If you sampled a 10khz sine wave at 20khz, it would become a square wave
> - certainly NOT accurate. Most filtration in CD players would wind up
> rounding the edges off anyway, but it's still not accurate - as compared
> to the source.

At the risk of spiraling way off-topic, I feel compelled to address this
very common misunderstanding. The Nyquist Theorem provides the
mathematical underpinning for *exact* transformation of a continuous
representation of audio into a discrete representation. If you sample a
10kHz sine wave at 20.01kHz, you get a 10kHz sine wave coming back out.
There are no "edges" to round off, because the digital-to-analog
reconstruction is not done by connecting the dots.

Of course, this is all in theory. Paul's caution about the difference
between theory and practice is quite correct. Since perfect filters don't
exist, one has to sample at noticeably more than 2x the highest frequency
to be represented. The question is whether 44.1kHz is enough, or whether
you need 96k or even 192k, and if higher sample rates are necessary, what
type of filter is optimal.

It's worth noting that there's more than just representation of audio to
be considered. Any non-linear processing (such as compression or VA
synthesis) is going to tend to produce aliasing, and higher sampling rates
can greatly reduce the impact of that aliasing on audible frequencies.

> > rounding the edges off anyway, but it's still not accurate - as compared
> > to the source.
> At the risk of spiraling way off-topic, I feel compelled to address this
> very common misunderstanding. The Nyquist Theorem provides the
> mathematical underpinning for *exact* transformation of a continuous
> representation of audio into a discrete representation. If you sample a
> 10kHz sine wave at 20.01kHz, you get a 10kHz sine wave coming back out.
> There are no "edges" to round off, because the digital-to-analog
> reconstruction is not done by connecting the dots.

But if you sampled a 10khz sine and a 10khz square wave, it'll still come
out exactly the same. Having a higher sampling rate will yield
better/closer to the original results, which was the point of the original
post IIRC.

-->Neil

-------------------------------------------------------------------------------
Neil Bradley            In the land of the blind, the one eyed man is not
Synthcom Systems, Inc.  king - he's a prisoner.
ICQ #29402898

I actually had a guitar playing buddy in high school (great guitarist who
knew how cool some short delay would be) that asked me how much cable he
would need coiled up in the back of his amp to get the delay he wanted.  He
was somewhat surprised at the size of the reel I recommended. <snicker>
Larry

----- Original Message -----
From: Neil Bradley <nb@...>
This reminds me of a clueless individual who thought that buying 5 foot
MIDI cables would have a noticeable effect on MIDI delay over 10 foot
cables. Little did this individual know that for even a single sample @
44.1khz delay to be induced by the wire it'd have to be about 2800 feet
long. ;-)

-->Neil

This is the industry I came from and as a side note, your opinion about
96k is shared among many people and is brought to evidence by the
lackluster acceptance of DVD-A.

-----Original Message-----
From: Chris Walcott [mailto:cwalcott@...] 
Sent: Tuesday, October 29, 2002 4:18 PM
To: MOTM-l (E-mail)
Subject: RE: [motm] Re: OT: Tales from an Audiophiles Crypt


In the November issue of Electronic Musician there is an article on
"Bridging the 96k Gap".  Very interesting article.  The thing that
struck me was that two highly regarded engineers (can't remember their
names) said that there very little difference between a track recorded
entirely at 24/96 from a track that was recorded at 24/48 and upsampled
to 24/96.
 
For me this is good news as I'm not about to upgrade my studio to
support 96k.  The cost would be enourmous.  Instead I can buy a stereo
24/96 interface (edirol makes a usb one for under $300, MOTU makes one
for about a grand) for monitoring and do my pre-mastering at 96k in the
DAW.
 
- chris

-----Original Message-----
From: paulhaneberg [mailto:phaneber@...]
Sent: Tuesday, October 29, 2002 3:25 PM
To: motm@yahoogroups.com
Subject: [motm] Re: OT: Tales from an Audiophiles Crypt


Although these individuals had no trouble recognizing the higher 
quality of 24 bits over 16 bits, they could not hear an appreciable 
difference between the standard 44.1 kHz sample rate used by CDs and 
the higher 96 kHz and even 192 kHz sample rates.

I have read some equipment reviews which raved about the improved 
quality of higher sample rates, but these were not double blind 
studies.  Personally I think the reason some higher sample rate 
converters sound better is the quality of the filters, not the rate 
of conversion.  For instance my Apogee Special Edition converters 
sound way better than my Digidesign converters both at 44.1 kHz.


Yahoo! Groups Sponsor	

ADVERTISEMENT


 
<http://rd.yahoo.com/M=212804.2460941.3878106.2273195/D=egroupweb/S=1705
032277:HM/A=810373/R=0/*http://geocities.yahoo.com/ps/info?.refer=blrecs
>
<http://rd.yahoo.com/M=212804.2460941.3878106.2273195/D=egroupweb/S=1705
032277:HM/A=810373/R=1/*http://geocities.yahoo.com/ps/info?.refer=blrecs
> 	
 
<http://us.adserver.yahoo.com/l?M=212804.2460941.3878106.2273195/D=egrou
pmail/S=:HM/A=810373/rand=580387767> 	

Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service
<http://docs.yahoo.com/info/terms/> .

> But if you sampled a 10khz sine and a 10khz square wave, it'll still
> come out exactly the same.

This is just another way of saying that the maximum frequency represented
is 10kHz.

> Having a higher sampling rate will yield
> better/closer to the original results, which was the point of the
> original post IIRC.

If the human ear can't hear anything above 20kHz, then it makes no
difference at all if a 15kHz square wave looks better on the scope sampled
at 96kHz. (It'll only look a little better, anyway.)

I don't really have a strong opinion about whether 96kHz sounds better
than 44.1, except in the context of audio processing. I just didn't agree
with your statement about the Nyquist theorem.

> > But if you sampled a 10khz sine and a 10khz square wave, it'll still
> > come out exactly the same.
> This is just another way of saying that the maximum frequency represented
> is 10kHz.

That's already a given, and you missed my point anyway. If you sample a
sine wave @ 20Khz and a square wave @ 20khz, you will only get a 10khz
square wave when you go D to A. The sine wave will lose detail. Having a
higher sample rate will keep the shape of the original waveform much more
closely than the lower sample rate.

> > Having a higher sampling rate will yield
> > better/closer to the original results, which was the point of the
> > original post IIRC.
> If the human ear can't hear anything above 20kHz, then it makes no
> difference at all if a 15kHz square wave looks better on the scope sampled
> at 96kHz. (It'll only look a little better, anyway.)

It does if the original sampled sound isn't a square wave.

> I don't really have a strong opinion about whether 96kHz sounds better
> than 44.1, except in the context of audio processing. I just didn't agree
> with your statement about the Nyquist theorem.

I don't know what "statement" you're referring to, other than the quality
of the waveform has zip to do with the Nyquist Theorem.

-->Neil

-------------------------------------------------------------------------------
Neil Bradley            In the land of the blind, the one eyed man is not
Synthcom Systems, Inc.  king - he's a prisoner.
ICQ #29402898

On Tue, 29 Oct 2002, Tim Walters wrote:

> If the human ear can't hear anything above 20kHz, then it makes no
> difference at all if a 15kHz square wave looks better on the scope sampled
> at 96kHz. (It'll only look a little better, anyway.)

And there's the rub.  Most people CAN distinguish between high frequency
sines and triangles, even though the harmonics of the triangle are above
the range of their hearing.  Ultrasonic components have an effect on
perception, even if they can't be heard.  Further proof of this is the
"audio spotlight" that delivers audio using ultrasonics (see
holosonics.com).

The question is, how high do you need to go to accurately reproduce a
performance?  Horns are the acoustic instruments that produce the most
ultrasonics, and they don't do much past 50kHz.  A VCO can go as high the
air's ability to carry the vibration, but there's a point where it just
doesn't matter to humans.

JDM

PS we're beginning to sound like ANALogue HeAVEN. :P

On Tue, 29 Oct 2002, Neil Bradley wrote:

> That's already a given, and you missed my point anyway. If you sample a
> sine wave @ 20Khz and a square wave @ 20khz, you will only get a 10khz
> square wave when you go D to A.

No, you get a 10kHz SINE wave, due to the reconstruction filter following
the DAC.  That is, IF the filter is ideal, as specified by Nyquist.
Since no real world filter is ideal, in reality you'll get some of the 2nd
harmonic at 20kHz.

JDM

> > That's already a given, and you missed my point anyway. If you sample a
> > sine wave @ 20Khz and a square wave @ 20khz, you will only get a 10khz
> > square wave when you go D to A.
> No, you get a 10kHz SINE wave, due to the reconstruction filter following
> the DAC.  That is, IF the filter is ideal, as specified by Nyquist.

Good catch - very true. Let's reverse my statement - if you sample a 10khz
square wave at 20khz, assuming reconstruction filter/oversampling you'll
get a 10khz sine wave - not the original 10khz square wave. Still the
point remains - a higher sampling rate will yield a "closer to the
original" waveform than a lower one.

-->Neil

-------------------------------------------------------------------------------
Neil Bradley            In the land of the blind, the one eyed man is not
Synthcom Systems, Inc.  king - he's a prisoner.
ICQ #29402898

>That's already a given, and you missed my point anyway. If you sample a
>sine wave @ 20Khz and a square wave @ 20khz, you will only get a 10khz
>square wave when you go D to A. The sine wave will lose detail.

No, it won't. That's the whole point of the Nyquist theorem. 
Everything below the Nyquist frequency is reproduced *exactly* (given 
ideal filters etc.). A 20kHz sine wave is just as detailed when 
sampled at 44.1kHz as when sampled at 96kHz; either way, it contains 
all the information of the original wave.

The only thing increasing the sample rate does is allow you to 
represent higher frequencies (and possibly to design a better 
real-world filter).

>I don't know what "statement" you're referring to, other than the quality
>of the waveform has zip to do with the Nyquist Theorem.

That would be it.
-- 


-----------------------------------------------------------------
Tim Walters : The Doubtful Palace : http://www.doubtfulpalace.com

>And there's the rub.  Most people CAN distinguish between high frequency
>sines and triangles, even though the harmonics of the triangle are above
>the range of their hearing.

That's because most people who try it use a sine and triangle with 
the same peak-to-peak value, which means the amplitude of the 
fundamental is different. My recollection is that studies that 
correct for this show that, in fact, most people *can't* hear the 
difference.

>   Ultrasonic components have an effect on
>perception, even if they can't be heard.  Further proof of this is the
>"audio spotlight" that delivers audio using ultrasonics (see
>holosonics.com).

They use difference tones to derive audible frequencies from 
ultrasonic frequencies. This doesn't prove anything about ultrasonic 
perception, any more than a theremin does.

I'm open to the possibility that ultrasonic perception is real, but I 
have yet to see any convincing evidence. The closest thing is the 
notorious Oohashi study, which I don't find convincing, but some do.

>The question is, how high do you need to go to accurately reproduce a
>performance?  Horns are the acoustic instruments that produce the most
>ultrasonics, and they don't do much past 50kHz.

Gamelans and crash cymbals go up into the MHz, IIRC.
-- 


-----------------------------------------------------------------
Tim Walters : The Doubtful Palace : http://www.doubtfulpalace.com

> >That's already a given, and you missed my point anyway. If you sample a
> >sine wave @ 20Khz and a square wave @ 20khz, you will only get a 10khz
> >square wave when you go D to A. The sine wave will lose detail.
> No, it won't. That's the whole point of the Nyquist theorem.
> Everything below the Nyquist frequency is reproduced *exactly* (given
> ideal filters etc.). A 20kHz sine wave is just as detailed when
> sampled at 44.1kHz as when sampled at 96kHz; either way, it contains
> all the information of the original wave.

Misstated example - Replace "sine wave" with "square wave". The square
wave turns in to a sine wave.

> The only thing increasing the sample rate does is allow you to
> represent higher frequencies (and possibly to design a better
> real-world filter).

And represent the original waveshape better provided it's not a sine wave.
;-) Change the input to a sawtooth or a square wave, you'll get a sine
wave out. A higher sample rate will not yield a sine wave as the lower
sample rate will. So again, increasing the sample rate will yield a closer
to the original waveform representation.

-->Neil

-------------------------------------------------------------------------------
Neil Bradley            In the land of the blind, the one eyed man is not
Synthcom Systems, Inc.  king - he's a prisoner.
ICQ #29402898

>  > The only thing increasing the sample rate does is allow you to
>>  represent higher frequencies (and possibly to design a better
>>  real-world filter).
>
>And represent the original waveshape better provided it's not a sine wave.

Not in any way except by representing higher frequencies.

Weren't we just here?
-- 


-----------------------------------------------------------------
Tim Walters : The Doubtful Palace : http://www.doubtfulpalace.com

> >>  represent higher frequencies (and possibly to design a better
> >>  real-world filter).
> >And represent the original waveshape better provided it's not a sine wave.
> Not in any way except by representing higher frequencies.
> Weren't we just here?

Yes, and what you're stating is incorrect. ;-)

If I have a 20khz sample rate, and I have the following waveforms being
sampled (assuming PERFECT alignment of the sample point and the peaks of
each cycle of each waveform):

10Khz Sine wave
10Khz Square wave
10Khz Sawtooth wave
10Khz Pulse wave

When played back at the same 20khz sample rate, they are *ALL* going to be
sine waves (assuming an ideal filter, of course). The peaks from the
sawtooth wave are now rounded.

Now let's assume a 40khz sample rate with the same 10Khz signals above.
Each waveform looks quite a bit closer to its original. Therefore, a
higher sample rate == higher detail at the same original input frequency.

If you double the sample rate, you double the significant samples within a
waveform, making it closer to the original. Hopefully this clears it up
100%.

-->Neil

-------------------------------------------------------------------------------
Neil Bradley            In the land of the blind, the one eyed man is not
Synthcom Systems, Inc.  king - he's a prisoner.
ICQ #29402898

>Yes, and what you're stating is incorrect. ;-)

Nope. :)

>If I have a 20khz sample rate, and I have the following waveforms being
>sampled (assuming PERFECT alignment of the sample point and the peaks of
>each cycle of each waveform):

First of all, you need >2x, not 2x. So say 20.01 kHz. No perfect 
alignment necessary.

>10Khz Sine wave
>10Khz Square wave
>10Khz Sawtooth wave
>10Khz Pulse wave
>
>When played back at the same 20khz sample rate, they are *ALL* going to be
>sine waves (assuming an ideal filter, of course).

Which is a perfect representation of the <= 10kHz component of each wave.

>The peaks from the
>sawtooth wave are now rounded.

Of course. The high-frequency information is missing.

>Now let's assume a 40khz sample rate with the same 10Khz signals above.
>Each waveform looks quite a bit closer to its original.

They will now be perfect representations of the <= 20kHz component of 
each wave.

>Therefore, a
>higher sample rate == higher detail at the same original input frequency.

Nope. The representation of the <= 10kHz components is identical. 
It's just not the only thing you're representing any more.

>If you double the sample rate, you double the significant samples within a
>waveform, making it closer to the original. Hopefully this clears it up
>100%.

If you want to use "detail" to mean additional high-frequency 
components, you can, I guess, although I find it counter-intuitive. 
But that's the only way this statement is correct.

Usually when people make this argument, they use "detail" to imply 
that the sampled waveforms are somehow "jagged" or "low-res." I'm not 
quite sure if that's what you're saying, but if it is, 'tain't so.

Is this boring the crap out of everyone but me and Neil?
-- 


-----------------------------------------------------------------
Tim Walters : The Doubtful Palace : http://www.doubtfulpalace.com

>How about taking this to:
>digital.stuff.rec.boring
>ZZZZzzzzz....

I guess that answers my question. I'll stop now.
-- 


-----------------------------------------------------------------
Tim Walters : The Doubtful Palace : http://www.doubtfulpalace.com

On Tue, 29 Oct 2002, Tim Walters wrote:

> Gamelans and crash cymbals go up into the MHz, IIRC.

I guess I blotted that memory out.  I had dog ears when I was younger, and
wouldn't go anywhere NEAR a drummer.  I absolutely hate crash cymbals. :)

JDM

How about taking this to:
digital.stuff.rec.boring
ZZZZzzzzz....

----- Original Message -----
From: Neil Bradley <nb@...>
Cc: <motm@yahoogroups.com>
Sent: Tuesday, October 29, 2002 11:53 PM
Subject: Re: [motm] Re: OT: Tales from an Audiophiles Crypt


> >>  represent higher frequencies (and possibly to design a better
> >>  real-world filter).
> >And represent the original waveshape better provided it's not a sine
wave.
> Not in any way except by representing higher frequencies.
> Weren't we just here?

Yes, and what you're stating is incorrect. ;-)

If I have a 20khz sample rate, and I have the following waveforms being
sampled (assuming PERFECT alignment of the sample point and the peaks of
each cycle of each waveform):

10Khz Sine wave
10Khz Square wave
10Khz Sawtooth wave
10Khz Pulse wave

When played back at the same 20khz sample rate, they are *ALL* going to be
sine waves (assuming an ideal filter, of course). The peaks from the
sawtooth wave are now rounded.

Now let's assume a 40khz sample rate with the same 10Khz signals above.
Each waveform looks quite a bit closer to its original. Therefore, a
higher sample rate == higher detail at the same original input frequency.

If you double the sample rate, you double the significant samples within a
waveform, making it closer to the original. Hopefully this clears it up
100%.

-->Neil

----------------------------------------------------------------------------
---
Neil Bradley            In the land of the blind, the one eyed man is not
Synthcom Systems, Inc.  king - he's a prisoner.
ICQ #29402898





Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/

----- Original Message -----

From: "Les Mizzell" <lesmizz@...>
> In Julian's test, auto jumper cables from K-Mart outperformed every single
> other cable tested, including those "golden ears only" $450 per meter
> jobbies...

hello all,
of course you can also use wet string to transmit audio - i think Canfords
did this at their stall at a recent AES in Amsterdam

this and other daft techniques have yet to be tested - though i did try the
setting-fire-to-a-speaker-for-a-low-pass-filter-effect

cheers
paul b

I'm sorry, but I have to chime in here.

FILTERING NOTWITHSTANDING:

The digital data for a sine wave at exactly 1/2 of the sample rate (a 
10k sine wave for 20k sampling) looks like this:

-_-_-_-_-_-_-_

The DATA will be:
-32768  ...  32767  ... -32768  ...  32767  ... etc.

it's EXACTLY THE SAME for a sine wave.

Now, as to if this makes a difference, if people can hear the 
difference between a 20k sine and a 20k square, i couldn't say.

Imagine THIS, though.  The digital data for a 7.5k anything (square, 
sine, whatever) at 20k:

-_--__-_--__-_--__

at 20k, you CAN'T record a 7.5k sound, you can wiggle between 10k and 
5k, though.  Sure, my details on what it looks like might be wrong, 
there are other ways it could be represented, but either way, it's 
ugly.

The filter makes all this "OK" by reconstructing what the sound 
should have been, by lopping off aliasing frequencies that 
are 'outside human hearing.'  This is also ugly.

Just try to imagine making something like this on graph paper, what 
it would look like to try to represent something between the grid 
lines.

So, really, I think you need at lesat 192k.  Can you hear the 
difference?  Hell if I know, but at least the data is THERE.  
Filtering off at 20k is just a hack.  It's lame.  I'd like to have a 
system with NO filtering.  Not likely, sure, but imagine a sampling 
rate so high that at 20k, even drawing everything SQUARE your sound 
would be so high rez that you couldn't tell.  THAT's what i'm talking 
about, yeah, yeah.

::nods off::


--- In motm@y..., Tim Walters <walters@d...> wrote:
> >That's already a given, and you missed my point anyway. If you 
sample a
> >sine wave @ 20Khz and a square wave @ 20khz, you will only get a 
10khz
> >square wave when you go D to A. The sine wave will lose detail.
> 
> No, it won't. That's the whole point of the Nyquist theorem. 
> Everything below the Nyquist frequency is reproduced *exactly* 
(given 
> ideal filters etc.). A 20kHz sine wave is just as detailed when 
> sampled at 44.1kHz as when sampled at 96kHz; either way, it 
contains 
> all the information of the original wave.
> 
> The only thing increasing the sample rate does is allow you to 
> represent higher frequencies (and possibly to design a better 
> real-world filter).
> 
> >I don't know what "statement" you're referring to, other than the 
quality

> >of the waveform has zip to do with the Nyquist Theorem.
> 
> That would be it.
> -- 
> 
> 
> -----------------------------------------------------------------
> Tim Walters : The Doubtful Palace : http://www.doubtfulpalace.com

Sorry, I thought I could stay out of this thread but I guess I was wrong :)

As you know, 20K is a very low sampling rate.  There is a huge difference
between 20 and 44.1  There is an audible improvement depending on the
source with 88.2.  However, it's a case of diminishing returns.  The same
applies to bit depth.  There is much less of a difference between 24 and 20
bit, than there is between 20 and 16 bit.

Things to consider:

1) There is a difference between digital recording and digital mixing.
There isn't any evidence to support the idea that 192 recording sounds
better than 96, but there is some evidence that 192 mixing is better.  This
is one reason why I will continue doing my final mixes in analogue.

2) There are 16/44.1 converters that sound way better than 24/96
converters.  Most of the sound depends on the power supply, shielding,
analogue components used, etc.

3) It's mostly marketing.  Manufactures want to come up with higher numbers
in order to pressure studios into buying the latest gear to impress their
clients.  So you say "I think you need at least 192K".  What happens after
they come up with 384??  I remember when the Digidesign Pro Master 20 was a
big deal.  That lasted about six months.

Quality analogue is still a good investment, but digital goes up in power
and down in price at an alarming rate.  So unless you have unlimited money
to spend, the way to have the best digital sound you can afford is to buy
trailing edge technology.

>So, really, I think you need at lesat 192k.  Can you hear the
>difference?  Hell if I know, but at least the data is THERE.
>Filtering off at 20k is just a hack.  It's lame.  I'd like to have a
>system with NO filtering.  Not likely, sure, but imagine a sampling
>rate so high that at 20k, even drawing everything SQUARE your sound
>would be so high rez that you couldn't tell.  THAT's what i'm talking
>about, yeah, yeah.

The example with the 7.5k sine wave is a good one but the analysis 
is faulty.  It doesn't matter what the waveform looks like after 
conversion to digital.  What matters is that the harmonics which are 
present in the digital representation of the waveform which are not 
present in the original sine wave are all higher than one half the 
sample rate and will therefore (in theory) be removed by the filter 
in the D to A converter.  In practice, as has been stated here is 
that the filters are not perfect and neither is the clock.  The lack 
of perfection in the filter and clock are what causes the harshness 
in digital sound.  Filters and clocks are much better than they used 
to be and in high quality pro gear can be outstanding.  Many lower 
quality and lower cost converters sound bad becasue of the poor 
quality of the filter and clock.

As far as the problem being in the mixing not in the conversion:  
This can be true, it depends on the internal bit depth of the 
software/hardware combination that makes up your mixer.  If you have 
only 24 bits internally and first reduce the level of your signal by 
48 db and then reamplify it by the same amount you are left with a 
16 bit signal rather than a 24 bit one.  The system I use (ProTools) 
has an internal bit depth of 56 bits which is more than sufficient 
for extensive processing without fidelity loss.  I assume most pro 
grade systems are similar in bit depth.  The quality of the 
processing algorithms are very important as well.

The Nyquist criterion is provable mathematically assuming perfect 
filters and clocks.