I'm asking if, in current version of Canonical (and in CsoundAV) the 
analysis stage of LPC is broken. In fact, I can render good speech 
using old LPC analyisis files, but, if I attempt to analyze a new file 
with current Csound (even using default parameters) all what I obtain is 
some modulated noises that seem to have no relation with the original 
.wav files containing speech. On the contrary, if I re-synthesize the 
old analysis files with current csound version I can get the correct 
result. Unfortunately, I've lost the sources of the old Csound version 
that allowed me to make good LPC analysis.

If someone is able to make good analysis/re-synthesis with current 
Csound release (that is canonical csound v4.23f11gbs, isn't it?) or with 
CsoundAV (that implements the same LPC analysis code) please tell me 
what parameters is using.

thanks in advance

-- 
Gabriel Maldonado
http://csounds.com/maldonado
--