Microsoft Sam, SAPI alternatives

Question

We have a application that we were planing to use Microsoft speech API for. Now we tested it on Windows XP using Microsoft Sam voice and frankly it sound terrible ... It's almost impossible to hear what the voice is trying to say.

Are there other, better voice. Are there any updates or newer versions out there that are better. Are there other product, open source projects etc that can work as an alternative?

Just to clarify - It needs to have some sort of API so I actually can program against it.

Answer 1

On Windows about the best I have found was using the speech API and voices from AT&T Natural Voices: http://www2.research.att.com/~ttsweb/tts/demo.php

They are however VERY expensive if available at all. I have run into projects where the usage/business model was so far from what AT&T was thinking of that they wouldn't even sell a license.

There is a free software alternative, Festival: http://www.cstr.ed.ac.uk/projects/festival/ , the quality though is horrible. It is about 10 years behind the current sound quality of commercial systems. It is however free.

A third alternative which has worked well for me was to shift the voice synthesis part of a few projects to OS X. OS X has a decent set of tools and speech APIS and a fairly decent set of stock voices. The downside of course is that prorams written for these APIs run only under OS X which runs only on Apple hardware.

Answer 2

AT&T Natural Voices engine produces great speech but its not free
there is also NeoSpeech which are also good - Not free as well

Answer 3

You don't describe your licensing needs, so I don't know if any of these will be suitable in that regard, but all of the following are sources of SAPI 5 compatible voices:

Ivona (http://www.ivona.com/) - I'm using their Kendra voice on a SAPI project.

AT&T Natural Voices (http://www2.research.att.com/~ttsweb/tts/)

Loquendo (http://www.loquendo.com/)

Acapela (http://www.acapela-group.com/products/products.asp)

Cepstral (http://www.cepstral.com/)

fonix (http://www.fonixspeech.com/tts.php) - only if you loved the original Speak & Spell.

Nuance RealSpeak (I'm not sure about this one...)

Answer 4

You can use free and open source Festival. The default Festival voice sounds a little like Stephen Hawking but you can use some other much better HTS voices. For example try selecting Peter HTS 2011 voice on this demo page: http://www.cstr.ed.ac.uk/projects/festival/morevoices.html. Most of HTS voices for Festival that I've seen are not allowed for commercial use however this one seems to be free: http://homepages.inf.ed.ac.uk/jyamagis/software/page54/page54.html You can check this youtube tutorial: http://www.youtube.com/watch?v=MmcLFJQpv2o