On 21/05/2010 11:01, bharat joshi wrote:
Accuracy is much better as we use SPHINX-4. How easy would it be to use VEDICS with another speech recognizer? On a related note, in my opinion the speech interface should be pretty much independent from the actual speech recognizer used. Has there been any attempts to abstract the "speech interface" in any way? I.e. provide a standard that can be used in conjunction with any speech recognizer that is able to recognize e.g. phonemes. Sounds like the "dynamics" of VEDICS is a step in this direction as it helps decouple the application interface from the speech interface. Helge |