UCLA Phonetics Lab

SPEECH SYNTHESIS

Parameter (formant) synthesis: The lab uses Synthworks from Scicon, a version of a Klatt hybrid synthesizer with a lab-designed spreadsheet interface which works on Macs and PCs. Some information about using it a few years ago is provided by this handout for a Ling. 104 assignment (thanks to Melissa Epstein); this will be updated in Fall 2007.

For resynthesis, mainly to vary or control F0 and/or duration, we use Praat.

Text-to-Speech (TTS) synthesis: Our Macintosh computers have Apple's PlainTalk. Some information about using it is provided here and here. A page with useful information about OS X and Win XP is here. In the past some folks have played around with MBROLA for text to speech of understudied languages.

Synthesis on a bigger scale: Ying Lin's list of information about articulatory-acoustic models available through the web. However, this list of links is not maintained and may be out of date.

Of historical interest: a link to a page about Lloyd Rice's CompuTalker (Lloyd a former lab member)

Back to facilities
Back to the UCLA Phonetics Lab