synthesizers (was: Re: eSpeak and Festival)

Hynek Hanke hanke at
Wed Aug 6 06:19:59 EDT 2008

> On 13 Jul, Hynek Hanke <hanke at> wrote
>> It would be great if somebody who thinks that Festival
>> is actually worse than eSpeak in quality of speech
>> could try to elaborate more about the reasons.

Hello all,

thank you for all your feedback. Now it is much more
clear to me why many users prefer this or that TTS
system which I would think is not as natural as the other
ones. This is a valuable input for the future developement
of Speech Dispatcher.

One of the outputs is an idea that we could offer in
Speech Dispatcher a capability to run-time switch between
synthesizers like espeak and festival based on the nature
of the text being read and the speed of pronounciation.
So if the user switched on this capability, for example,
espeak would be used for high speed reading and ordinary
work while a synthesizer like festival with potentially
more natural sounding voices could be used for reading
of long texts, books and in other occasions where natural
speech is prefered. Of course this would have to be

Now in Speech Dispatcher 0.6.7 which was released
this week and will soon be available in the distributions,
there is another possibility on the side of TTS systems
trying to sound naturally, which is espeak with MBROLA
voices. It is free but not Free Software, so it contains significant
limitations, although on the other side, it offers quite a broad
set of voices. This can be used with the new espeak-mbrola-generic
output module.

With regards,
Hynek Hanke

More information about the Speakup mailing list