Getting Vocal in my old age....

[patchell]

jbv wrote:

>
> Mmmmh... Most tables of formant frequencies I've seen
> included central frequency as well as bandwith and relative
> amplitude for each formant... I doubt that giving each formant
> the same bw & Q wil produce acceptable "vocal" sounds...
>
> Another parameter to consider is the spectral content of the
> excitation source (just like in a vocoder).
>
> Actually, 18 months ago I spent some time brainstorming
> on a similar project of a VC filter bank based around the
> VCF CEM chip (forgot ref #), and did some DSP simulation,
> and things appeared much more complex than I first thought...
>
> As for morphing between vowels, I've always wondered
> if the transition between 2 vocal sounds was a simple
> interpolation between formants parameters (F, bw & Q),
> or if it was some more complex crossfades between
> spectral contents...
>
> At least a bank of formant filters could allow to make
> choirs from lush synth chords (or from orchestral sounds).
> But regarding speech simulation, some more complex technics
> are involved, especially regarding transitions between sounds...
> AFAIR, some speech synth systems in the early 80's were
> using transient between vocal sounds, instead of formant
> tables...
>
> jbv

    All things to consider.  I will have to see how it sounds when it is all
done.  Did a lot of wheel spinning yesterday just trying to figure out the voltage
mapping functions....my brain isn't as quick as it used to be.  Hope to have some
results by the end of the day..

--
 -Jim
------------------------------------------------
* Visit:http://www.silcom.com/~patchell/
*-----------------------------------------------
*If you didn't buy a home in Santa Barbara,
* You didn't pay enough!
------------------------------------------------