Oh yes, it is not only possible, it has been done, but by Trevor Wishart, for CDP
(end of commercial plug), using pvoc analysis files. So prolific has he been that I
have never had cause to look at the pvoc support opcodes in Csound, but it seems
that the new opcodes by Richard Karpen should do quite a lot of what you want. The
averaging you describe might need a new opcode, however.


Richard Dobson

QATJD wrote:

>         Hi. I may be dreaming, and I don't know how it would be actualized, but
> couldn't this be done by doing an analysis (pvoc, hetro, lpanal, etc -
> probably pvoc) of both files and averaging them sample by sample by
> averaging _each frame_  of spectral data with its counterpart in the other
> file? Then you would be 'blending' the two spectra at a pvoc level of
> resolution. You could then make it a weighted average with the weight
> shifting from one file to the other over the 'note' time.
>         No?
>                                                                         Q