[linux-audio-dev] idea: wavelets of variable frequencies

[Tom Weber]

Hi!
I just joined this list. I have an idea for a sound format which is
similar to spectrum analysis (but better?): represent the sound as a
synthesis of sinus waves where the frequencies are free and can change
with time.
Instead of doing a discrete fourier transform when reading a small
frame of the sound, do a dense transform (every 0.1 Hz?) and pick out
the peaks. Then assume that a similar enough frequency in the next
frame comes from the same source, keep joining those and your sound
will be represented by a lot of oscillators (wavelets) with amplitude
and frequency curves. I guess this would overcome the weaknesses in
having fixed frequencies.

This idea shouldn't be new, where can I read about it? Has it been
implemented anywhere?

Thanks in advance,
Tom Weber