How much data are the oscillators in pvsosc based upon, & if i wanted to
shoot some .pvx analysis files of oscillator shapes of my own (that were say
simple cycles of 1024 or 512 or 2048 samples in length) what would be the
simplest & best way to do that - can i capture all the info theoretically in
"a single frame" & save it to .pvx?

percieved advantages:

* no nyquist issues on oscillator based synthesis (as the GEN30 "workaround"
is based on an FFT analysis anyway...)
* further pvx processing possibilities / merger of synthesis & sample
resynthesis
* effectively makes concurrent fsig oscillator signal streams "phase
syncronous" if being reproduced under the same parameter conditions
(overlap, bin count etc...) ??? (Victor? Anyone?) 

percieved disadvantages:

may result in a less "pure" or representative synthesis on large
transposition values of the signal upward or downward (which would have to
use pvsscale assumedly...)

has anyone gone down this route at all using pvx to achieve phase syncrous
concurrent signals for crossfadeing ?

as far as FM goes, i could always do it & then create an fsig stream based
on it using pvsanal 

the whole reason for the desire to pre-analyse the others oscillators is to
avoid nyquist flippage on (hopefully phase synceable, & hence
crossfadeable..) square-like wave shapes... 

but then assumedly any nyquist defying square shape will also pollute any
subsequent fft analysis? 

but then how does gen30 "solve the problem"?

please help me understand!

many thanks...




-- 
View this message in context: http://www.nabble.com/pvs-oscillator-collection--%28potentially-%22hello-Victor%22-again..%29-tp15058819p15058819.html
Sent from the Csound - General mailing list archive at Nabble.com.