Linux TTS Voices

Fred Roller fred at
Fri Feb 26 11:47:30 UTC 2010

Kyle wrote:
> eSpeak is currently the best free/open source voice synthesizer 
> available for Linux, and it's really quite good, especially considering 
> its size and memory usage. I use it on my system every day and it 
> definitely sounds better than flite, better than the old Microsoft 
> voices such as Sam, Mary and mike and much better than the proprietary 
> and unmaintained IBMTTS or ViaVoice or TTSynth or Voxin or whatever 
> they're calling it now. That being said, some people are used to hearing 
> the outdated voices, and there are some other proprietary voices that do 
> sound good. Since it's all about user preference and giving people 
> choices, an interface to speech-dispatcher may be the best thing for 
> your application, unless you are writing it for personal use only. 
> Something simple could basically pipe through spd-say or have its own 
> simple client. Something more complex can interface directly with the 
> speech-dispatcher API for much more user control and/or functionality.
> Hope this helps,
> Kyle
Also, you could look into "mbrola":

This works /with/ espeak and helps humanize the voices significantly 
over espeak regular (in my experience)

Here is an example I was working on a while back:

granted it's a bit clippy but that would be because I had the speed set 
to 200 wpm.



"Life is like linux, simple.  If you are fighting it you are doing something wrong."

More information about the ubuntu-users mailing list