[net.audio] The Digital Approach to CD

shauns@vice.UUCP (06/25/84)

I have an open question for Mr. Pearson, subject of recent flames.

You speak of the right way for digital being absolutely perfect square waves.
This implies a set of basis waveforms that are square waves, not sinusoidal.

For all of us out here that are still stuck in Fourierland, the basic idea of
Fourier waveform approximation is that sinusoids of differing
frequencies form an orthogonal set of basis vectors in "music space", if you
will, exactly the same as the x,y, and z axes do in 3-space.  Hence the
position of any point can be represented by appropriate scalings of each basis
vector.  If we can find another set of orthogonal basis vectors, it will do
just as well as sinusoids for waveform approximation as the number of vectors
in the set gets large.

It seems to me I remember a series of waveforms, I think they were called
Bloch or Bragg functions, that are square waves.  Hence, you only need one
component, or a very small number anyway, to adequately reproduce an impulse.
Unfortunately, you need an infinite number to properly reproduce a sine wave.

SO, the decision comes down to:
what does your source signal look like?  Does it look like an impulse? Then
don't use sine waves.  Does it look like a sine wave?   Then use sine waves.
I'm not even going to discuss the problems of generating in a causal fashion
perfect square waves and what one has to do to play back the signal on
conventional audio systems.

Mr. Pearson, might this be some of what you are alluding to?

the wandering squash,
-- 
				Shaun Simpkins

uucp:	{ucbvax,decvax,chico,pur-ee,cbosg,ihnss}!teklabs!tekcad!vice!shauns
CSnet:	shauns@tek
ARPAnet:shauns.tek@rand-relay

tony@asgb.UUCP (06/28/84)

In reference to the mention of non-sinusoidal basis functions, I believe
you're referring to "Walsh" functions.

And on that note, I believe I'll hit the 'U' key and rid myself of faulty
arguments put forth by people who obviously have no knowledge of signal
theory. ark@rabbit's letter was quite correct, and to refute incorrect
arguments any further would require starting from the ground up, and I
don't feel like typing in chapters of good ol' McGillem & Cooper.

Tony Andrews
BSEE '81