Is it the time for "KSpeach"?
djurban at pld-dc.org
Tue Sep 14 21:05:26 BST 2004
Jarosław Staniek (Tuesday 14 of September 2004 21:05):
> What do you think?
Anyway Polish speech synthesizers suck bigtime and there is no opensource
organisation/movement in poland that could get the funds or that would want
to fund to create a good set of voices.
We have one set of voices for mbrola/festival (women only, it was someone's
thesis), but festival as a speech framework sux big time (no support for
latin2 chars, problems with punctuation and date reading).
Also we have a 'powiedz' (localized version of 'say', both a fork of rsynth)
synthesizer but it is based on Klatt algorithms - which makes it a great tool
for reading a single sentence (supports reading timestamps, emotional
punctuation signs, creates accents on questions properly), it sucks with
multiverse texts, since it does not recognize linebreaks and when
synthesizing constantly the text tends to be ununderstandable.
Anyway, I havent used kttsd since 9 June 2004 and it didnt work then, it seems
there was a major rewrite though.
djurban at pld-linux.org
More information about the kde-core-devel