[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: normalizing pitch in human speech



We do that pretty well in SuperVP/AudioSculpt:
- do pitch analysis
- transpose accordingly to any fixed (or not) reference
thats's it

Regards


Mathias Oechslin a Ãcrit :
Dear List,

We like to produce speech stimuli which are characterized by normalized pitch hight - so called flattened speech.
That means, finally the speech signal does not contain pitch changes anymore,Ã but remains easy understandable.


I imagine that the allorythm has to do FFT, definition of the normalized pitch, afterwards resynthesis.

Does anyone has any experience with this?

thanx for all ideas,

Mathias

--

Ã


**************************************************

Mathias Oechslin, lic. phil.


Doktorand

Lehrstuhl fÃÂr Neuropsychologie

Psychologisches Institut

UniversitÃÂt ZÃÂrich


Telefon: +41 44 635 74 07

Faxà à : +41 44 635 74 09


m.oechslin@xxxxxxxxxxxxxxxxxxxx

http://www.psychologie.unizh.ch/neuropsy


**************************************************



--
Xavier Rodet -- rod@xxxxxxxx --  33 (0)1 44 78 48 68, Fax: 33 (0)1 44 78 15 40
Responsable de l'Equipe Analyse/Synthese - Head of the Analysis/Synthesis Team
IRCAM 1 place Stravinsky 75004 Paris FRANCE --  W3 site : http://www.ircam.fr/
W3 site of Analysis/Synth. Team: http://www.ircam.fr/equipes/analyse-synthese/