Re: [ARSCLIST] Sampling Theory (was Fred Layn's post on the Studer list re: Quantegy)

[George Brock-Nannestad <pattac@xxxxxxxx>]

From: Patent Tactics, George Brock-Nannestad

In my view, most of the lack of definition and distortion we perceive when we
are listening to old CD-style digital audio is caused by the severe band
limitation and its influence on time delay between audible signal components.

Non-linear distortion does not improve matters.

However, I have come to think of modulation as another source of time delay
(phase) problems, apart from the filtering in the incredibly narrow range of
20 kHz to 22.05 kHz. Any amplitude modulation of an audible signal of a
suitable high frequency occurring in nature is accompanied by upper and lower
sidebands. If you remove one sideband (in this case the upper, because it may
be lost in the anti-aliasing filter), if I remember the rotating vector
diagrams of 35 years back, you will end up with a time delay distortion. The
situation is aggravated when Frequency Modulation is active, because the
sidebands go much further away from the carrier. The more there is an
imbalance between the mirrored signals in the sidebands, the greater this
distortion is.

Still 22 years after its appearance, a paper by Peter Fellgett, then at the
University of Reading in the U.K., has a great importance for us, and I would
urge anybody seriously discussing Sampling Theory, Nyquist Frequency (hotly
debated in this list a couple of months ago), and distortion to get hold of
this article. The bibliographic information is the following:

Fellgett, P.B., 'Some comparisons of digital and analogue audio recording',
The Radio and Electronic Engineer, Vol. 53, No. 2, pp. 55-62, February 1983.

Personally I think that the reason that digital was so acclaimed when it
arrived in its primitive form the 44.1 kHz, 16 bit CD, was that it enabled a
clean bass at a level as high as you wanted it, without acoustic feedback.
However, do not permit yourself to believe that you have 24 bit or even 20
bit linearity at the relatively low prices you have to pay these days. For
precision you have to have an incredible temperature stability and when you
look at the precision expressed in percent or promille, you will realize that
20 bit is no joke, electronics-wise.

Best regards,


George

The above in response to Eric Jacobs response:

> For kicks, I just ran the Cardas Test record, Tracks 2b and 2c, 1 to 30 kHz
> sweep, and was effectively flat at 30 kHz.  The second harmonic at 60 kHz was
> still clearly visible at about -30 dB, with the noise floor at -60 dB. The
> response wasn't perfectly flat throughout the sweep, but the point is that
> good analog gear (vinyl or magnetic tape) is perfectly capable of reproducing
> a usable signal well beyond 20 kHz.
>
> Digitizing and playing back a sine wave is easy.  Try a square wave and
> see what happens.  Try a 10 kHz square wave sampled at 44.1 kHz - you
> will find that the DAC will return a 10 kHz sine wave, not the original
> square wave.  It starts getting close at 192 kHz - the square wave starts
> looking more square with a good DAC.  What does sample rate mean to impulses
> and other non-sinusoidal waveforms as produced by brass, percussion and
> vocals?
>
> The above comments in no way speak to what people can hear or not hear,
> nor does it make an argument that higher sample rates are always better.
> It does make the point that the ADC-DAC chain is not yet a perfect
> reproduction of the original analog signal.  Whether what is lost in the
> ADC-DAC translation is audible is something that many of us are still
> trying to assess.
>
> Eric Jacobs
> The Audio Archive
>
>
> -----Original Message-----
> From: Association for Recorded Sound Discussion List
> [mailto:ARSCLIST@xxxxxxx]On Behalf Of Dave Bradley
> Sent: Monday, January 17, 2005 12:04 PM
> To: ARSCLIST@xxxxxxxxxxxx
> Subject: Re: [ARSCLIST] Sampling Theory (was Fred Layn's post on the
> Studer list re: Quantegy)
>
>
> I find it interesting that people point to subjective studies about what
> can be sensed above 20 KHz, or above 15 KHz for that matter, and say that
> because digital doesn't go there it's inferior. Take note that the cartridge
> on your turntable isn't going there either, and with good reason, the vinyl
> that the stylus is being dragged through can't reproduce up there either.
> There may be harmonics up there, but analog isn't delivering it. So if the
> problem is that digital isn't giving you something which you can't hear, but
> can sense, keep in mind that analog isn't giving that to you either.
>
> The fact of the matter is that a pure sinewave, when digitized properly,
> and then played back through a proper DAC will give you that same pure
> sinewave again. It doesn't give you stepped response. It doesn't give you a
> triangle or square wave. It gives you that pure sinewave.  Just because
> zooming in on the waveform in your favorite DAW software shows you individual
> sample values and a less than smooth wave form doesn't mean that the playback
> will be that way.  Keep in mind that in the end, what you are hearing is
> analog generated from a set of digital instructions or samples.  You are still
> hearing analog.