[sdiy] dx, chorus and Spock

[Don Tillman]

   > Date: Fri, 16 Aug 2002 19:25:53 +0930
   > From: "Batz Goodfortune" <batzman at all-electric.com>
   > 
   > >Don Tillman <don at till.com> wrote:
   > > >
   > > >I have a theory about our sense of hearing: I claim that we
   > > >can't hear waveforms or harmonics very accurately at all, but
   > > >we are remarkably sensitive to the subtleties of the *process*
   > > >that's making the waveform.
   > 
   > There is probably something in all of this but if you look at the
   > way the human hearing mechanism works, it's really the ultimate
   > in real-time fourier transformation. It is therefore almost
   > entirely harmonically based.

I think we're all aware of how the ear mechanism works, the diagrams
are in any encyclopedia, and the hairs in the cochea certainly suggest a
Fourier operation...

Yet we're not especially good at hearing Fourier spectrums.  If I'm
playing a Stratocaster, and the seventh harmonic happens to be
missing, you're not going to stand up and shout, "Hey man, you're
missing the seventh harmonic!".  It's not going to happen.  So while
there's some Fourier processing of some sort going on at a micro
level, at a macro level, where the neural bundles between the ear and
the brain get involved, there's much more abstracted processesing
going on which renders the micro Fourier information unaccessable and
unimportant.

So my theory is that the point of all the work in the neural bundles
is to recognize the processes that cause the sounds, both because
recognizing processes (such as the "footstep" process) is vitally
important for survival, and millions of years of evolution will hone
such skills so as to be able recognize processes from the subtlest of
clues.

The effects of my theory are that:

   We should be easily able to recognize sounds from the same process
   even if their Fourier spectrums or waveforms are wildly different.

and:

   We should be easiliy able to differentate sounds from different
   processes even if their Fourier spectrums or waveforms are nearly
   identical.

And sure enough, there are numerous examples.  For instance, anybody
can tell the difference between the sound of a real Hammond organ and
a midi keyboard playing Hammond samples.  Even in a worst-case
listening environment, say over a voice-grade phone line.  This is
because our sense of hearing can differentiate between the tonewheel
process and the digital sample playback process.

   > I contend that because of the workings of the human ear and it's
   > associated neurology, Synthesis is all about harmonics. Nothing
   > else really matters too much.

Right; that would be the exact opposite of my theory.  :-)

  -- Don

-- 
Don Tillman
Palo Alto, California, USA
don at till.com
http://www.till.com