How to tune meSpeak.js for better sound?

Question

I am new to meSpeak.js .

I cant go with the default voice they have in their demo. I tried playing around with the options that they have, but it didnt make much difference.

I would like to know how do I tune it so as to get the voice close to the level of Siri, not that robotic as in their demo. I am focusing on the sound quality.

Answer 1

the quite robotic voice is actually a feature of eSpeak (the embedded *NIX-application, http://espeak.sourceforge.net ). I don't know, if this might be actually overcome and might be made to sound like Siri. But, at least, some of the voice variants are sounding a bit better than others. You might want to search any boards and discussions related to eSpeak and voices (there are some threads on this subject to be found on the internet). Any tricks or definitions that work with eSpeak should work with meSpeak, too. (You may setup a voice-file with your own definitions and load this into meSpeak.) -- Sorry, but there's no definitive guide I would know of. It probably involves a deeper insight in how the wave forms for the individual phonems are generated. Maybe we have to accept some limitations in the voice quality. On the other hand, eSpeak is still the standard tool for this on Linux ...