TTS voices project

Henrik Nilsen Omma henrik at
Wed Dec 6 10:27:36 GMT 2006

Luke Yelavich wrote:
> On Tue, Dec 05, 2006 at 02:16:26AM EST, Jonathan Duddington wrote:
>> Then it's a process of gradual improvement, both of the phoneme sounds
>> and the letter-to-phoneme translation rules, and also various prosody
>> aspects such as stress, rythem and cadence, intonation, etc., depending
>> on how much work someone wants to do. 
> Heh. You know, I am really pondering creating an Australian english 
> voice. :)

Doit! I'm going to have a go a the Norwegian one, in part so I can learn 
the process of how voice tweaking works. I want to get a fairly detailed 
description of that up on the wiki page: so that we can 
make it REALLY EASY for people to contribute.

Anyone who has experience with working on eSpeak voices please help 
flesh out the getting started section of that page. I'd like to make the 
instructions really clear before turning to other groups for help.

For example, we have mentioned before that there are some technical and 
some less technical ways to contribute. Perhaps we can list those and 
try to describe each in some detail.

 * Write test case text files in a given language for use in testing and 
development (non-technical, native-speaking)
 * Create initial first-guess language files (highly technical, 
 * Making .wav files of different languages/voices for others to listen 
to (non-technical, non-native-speaking)
 * Listen to voice output and suggest improvements (non-technical, 
 * Adjust language files based on listener feedback (semi-technical, 
native-speaking helps)
 * Record sounds and extract patterns (semi-technical, native-speaking)
 * Modify eSpeak code to accommodate new language features (highly 
technical, non-native-speaking)

Anything else?


More information about the Ubuntu-accessibility mailing list