[CLAM] synthesizing SMS models in Pd

[Rich E]

Okay, sorry for getting into too much pd stuff... I basically just
need to know what needs to be done to the 1STF data before I can
synthesize it.  Is it ready to go, or is windowing still necessary?
Should the frames be overlapped?

I'll move the other questions over to the pd list.

regards,
rich

On 12/14/07, Xavier Amatriain <xavier at amatriain.net> wrote:
> Rich, I think you are writing to the wrong list :-) We can probably help
> you with the SMS stuff if you
> formulate it in general terms but not so much with pd internals ;-)
>
> Rich E wrote:
> > Hi list,
> >
> > I've been working on getting a real-time spectral-model synthesizer
> > going, so far it handles 1TRC SDIF files quite well!  Now I am onto
> > synthesizing the noise residual, but want to ask a couple questions
> > first:
> >
> > - Is there any other processing that needs to be done to the 1STF
> > data, or is it ready to hand to [ifft~] as is?  I assume that I need
> > to apply a window first, but maybe this is already done.  Also, would
> > pd's [rifft~] suffice (It adds the second half of the fft by symmetry)?
> >
> > - With the partial data, I interpolate the frequencies for every
> > sample because I am handed a frame for roughly every 10ms.  However, a
> > 1STF frame should cover 512 samples, which is roughly the same time.
> > Should I just set pd's blocksize to the fft size and send it a new
> > 1STF frame every time I synthesize a new 1TRC frame, or will I need to
> > interpolate the 1STF data to get smooth transitions still?
> >
> > If you can help clarify these issues for me, it is much appreciated.
> > Sorry if it is listed somewhere in Serra's publications, or in the
> > code.  I looked, couldn't find what I needed to know.
> >
> > RIch
> > ------------------------------------------------------------------------
> >
> > _______________________________________________
> > CLAM mailing list
> > CLAM at iua.upf.es
> > http://www.iua.upf.es/mtg/clam
>
>