I finished evaluating VoiceIn v.4.1, an ASR (automatic speech recognition) product of Fonix company. The results are very promising. Libraries are natively written in C, but a .NET wrapper is provided, allowing for using them in Visual Studio .NET with C# or VB.NET to program a Windows Mobile application.
The interface is though a bit complicated. Some methods need to be declared unsafe to allow using pointers that are required in some interface calls. The interface is documented though very well and I was able to create a small sample application that reacts on voice commands in a very short period of time. The ASR engine is speaker independent and supports many languages, like US and UK English, French, German, Italian, Japanese, Korean and Spanish. Specific language vocabulary can be extended with no limit (but considering OS limits of course).
Fonix also provides a product called DECTalk, a TTS engine for Windows Mobile with a very small footprint. I have not tested it, but I might consider use it too to make a voice confirmation after speech recognition.
No comments:
Post a Comment