Would it be possible for any of the programmers on the list to write an
implementation of the 'Spectral Rate Extractor' (used in Soundhack and
Max-DSP) for Csound?
In the Soundhack implementation,  a soundfile is analysed (FFT), and the
speed of frequency deviation is measured. Two soundfiles are returned, one
with the stable frequences, and one with the transients. The criteria for
deciding wether  the frequencies are stable/transient, are defined by the
user in terms of deviation (Hz) per analysis frame.
Personally I find this util exceptionally useful.
Perhaps a potential CSound opcode could take an FFT file from PVanal (..
would this work?), and return a signal which only contains partials which
deviate less than kdev per analysis frame. ie.

a1 spextract kdev,ifilcod[,iskip]

where kdev is the maximum freq deviation in a1 etc.
Might it be possible to add this function as an option to Pvoc, without
writing a new opcode?

yours hopefully,

JamieB