What do you think guys of the quality of this synthesizer, it is claimed to be expressive.
I tweaked also speed and volume to add some fine tune touches.
I am sure it is not an alternative to good voice acting, but I have a feeling it is better than bad voice acting.
This sounds pretty good,where is it from?
It is IBM Watson Text to speech service.
http://www.ibm.com/smarterplanet/us/en/ibmwatson/developercloud/doc/text-to-speech/using.shtml#ssml
Weird that they never mention a license for the voice output. There are much much better text-to-speech services (some even with cool accents) but the licensing on reusing the produced audio can be a bit restrictive.
I really like this one: https://www.cereproc.com/
Lots of cool voices:
https://www.cereproc.com/products/voices
…actually… now I can’t find a license for theirs either at quick glance.
Edit: this was another good one:
https://www.ivona.com/
pspeed thanks for sharing I will check them as well they support many voices hope if they support SSML as well
Unfortuantly watson is very limited in voice but it could be handled by audacity
For watson the generated tone will different when using the markup for my text with (good news, bad news, apology), i also tuned the text to slow down at the end with lower volume using ssml
For license it is paid cloud service, (but it does not cost alot honestly) I will check if there are restrictions on redisrtibution of generated wave file
For what it’s worth, the IVONA FAQ says you are free to distribute the produced audio if you have a license. I didn’t look into what it would take to get a license.
I might point to marytts, wich is pretty good and has open licensed voices avaible.
I just noticed something, the expressive speech for Watson is only working with one Female voice, all other voices are normal speech synthesizer and ignore the expressions , it was really nice and works pretty good, to have speech synthesized emotion and sounds, but it is very limited.