modelling a larynx…

Mar 5, 2007 at 12:10pm

modelling a larynx…

Hi everyone

I’m currently on a project that’s working with expressive speech synthesis, and I’m having to do some stuff from first principles, currently experimenting with applying formants to different oscillators.

Does anyone know what the most convincing way of synthesising an (unfiltered) larynx sound in max? If anyone’s got any experience of this in terms of speech or singing synthesis, It would be great to hear from you.

Cheers!

Andrew

#30633
Mar 5, 2007 at 1:58pm

You might want to look at Perry Cook’s work on SPASM (his
dissertation) — it’s somewhere on his princeton.cs home page.
Also, we use a generic buzz waveform (all harmonics to nyquist at
full amp) for the raw larynx in our RTcmix LPC instrument.

Ok, I took 5 seconds and found Perry’s page:

http://www.cs.princeton.edu/~prc/SingingSynth.html

and here’s the RTcmix docs on LPC:

http://music.columbia.edu/cmc/RTcmix/docs/instruments/LPCPLAY.html

brad

http://music.columbia.edu/~brad

Quoting “Andrew A. Cook” :

>
> Hi everyone
>
> I’m currently on a project that’s working with expressive speech
> synthesis, and I’m having to do some stuff from first principles,
> currently experimenting with applying formants to different
> oscillators.
>
> Does anyone know what the most convincing way of synthesising an
> (unfiltered) larynx sound in max? If anyone’s got any experience
> of this in terms of speech or singing synthesis, It would be
> great to hear from you.
>
> Cheers!
>
> Andrew
>

#98188
Mar 5, 2007 at 11:02pm

Quote: samoyed wrote on Mon, 05 March 2007 04:10
—————————————————-
> Hi everyone
>
> I’m currently on a project that’s working with expressive speech synthesis, and I’m having to do some stuff from first principles, currently experimenting with applying formants to different oscillators.
>
> Does anyone know what the most convincing way of synthesising an (unfiltered) larynx sound in max? If anyone’s got any experience of this in terms of speech or singing synthesis, It would be great to hear from you.
>
> Cheers!
>
> Andrew
—————————————————-

Have a look at the “singing-voice” example in the CNMAT Spectral Tutorials:

http://www.cnmat.berkeley.edu/MAX/downloads/

mz

#98189
Mar 5, 2007 at 11:17pm

On Mar 5, 2007, at 4:10 AM, Andrew A. Cook wrote:

> I’m currently on a project that’s working with expressive speech
> synthesis, and I’m having to do some stuff from first principles,
> currently experimenting with applying formants to different
> oscillators.
>
> Does anyone know what the most convincing way of synthesising an
> (unfiltered) larynx sound in max? If anyone’s got any experience of
> this in terms of speech or singing synthesis, It would be great to
> hear from you.

Matt Gilbert has some patches on his site he uses with some custom
ululation hardware he made: http://www.mattgilbert.net/projects/
vowelSynth/

Vlad

Vlad Spears
Urbi et orbi

http://www.daevlmakr.com

http://www.2secondfuse.com

#98190

You must be logged in to reply to this topic.