Mental Exercise

Eli Brandt eli at gs160.sp.cs.cmu.edu
Fri Mar 14 04:22:51 CET 1997


FlavorMaus at aol.com wrote:
> sorry, a machine is incapable of true randomness.

There are any number of ways to define "random", and none is the
"true" one.  For musical purposes, the appropriate definition is
psychoacoustic: does the signal have perceptible artifacts?  
(If, say, you cut corners on your shift register, it very well may.)
Gold-plated Platonic-ideal random numbers are not at issue here.

-- 
   Eli Brandt
   eli+ at cs.cmu.edu

[pedantic note for purposes of procrastination:]
A more stringent definition, and the one I find most elegant, is taken
from algorithmic information theory: loosely, can the data be
generated by any program shorter than it?  A computer can generate data
that meets even this strong criterion, given an insanely long time.



More information about the Synth-diy mailing list