2nd: Comments on the Text to Speech "algorithm"
jonsd at talktalk.net
Fri Mar 5 15:49:04 CET 2010
On 05 Mar, marc <sintsixtus at gmail.com> wrote:
> Now I've looked into vinux, and I must say: quite fast voice, nearly
> unreadable to me.
> That's why I remain with the mp3/oog idea,
Try this experiment.
1. Write a few sentences.
2. Make a list of all the different words in those sentences.
3. Sort the words into alphabetic order (so that they are not in the
same order as in the sentences).
4. Make a recording of each of the words (in the new order), each in a
separate sound file, using your voice.
5. Use a sound-editor program (eg. audacity) to construct a sound-file
for each of the sentences, by joining together the individual words.
6. Play the combined sound-files for the sentences.
Then re-consider whether it sounds like a good idea.
Also, how many words are there in the English language (including place
names, personal names, etc)? How long would it take to record them all
(all with the same tone of voice so that they are compatible with each
> but of course ... the industry might jail you to invent good things
> and not to make it. (do we make free software for ourselves or the
Sorry, but that sounds incoherent.
More information about the Speechd