The hearing of sampled sounds, and FFT-looped samples as such
willreveal, especially though reflections and audio effects *how* the
samples repeat,
the loops have been welded and the FFt and interpolation have changed the
character of the harmonics, unless the various jobs have been done well,
with musical (or other) results in mind.
It isn't possible to satisfy all parts of a "behind a curtain" listening
panel when playing back results of this kind and compare it for instance
with a real piano, and of course mathematically speaking: every
simplification
and every uninvertab le (non-bijective) operation means some loss of audio
freedom, and probably some people who want to suck onto launching and
operating
such a boat of processing facilities.
The spectrea of sounds with a loop of course will audible repeat, especially
for mature musicians, even when looped without hearing the splice point or
method, and often the loop will at least be a multiple of the 1/Fsample or
worse: a multiple of FFT-bin size with some averaging, either of which
becaomes
very audible when playing chords. Mathematically the playing with small
integer
variations of harmonics reperition intervals is provably dull in no time,
and that shows as listening fatique easily. And I agree wavetables are great
fun, but of course it's hardly a *new* science...
So I am much for the spirit of the Open Source to keep saying what is what,
unless of course commercial patents get in the way...
Theo Verelst
(Who was checking the astronomical numbers of multidimennsional FFTs of some
NVidia Cuda and Fermi(double precision) graphics cards the other day, that
whould be fun to play with for constructing sound possibilities, any takers
?)
--
dupswapdrop -- the music-dsp mailing list and website:
subscription info, FAQ, source code archive, list archive, book reviews, dsp
links
http://music.columbia.edu/cmc/music-dsp
http://music.columbia.edu/mailman/listinfo/music-dsp
