I'm curious about people's experience with speech recognition development. For myself, I'd like to be able to use an external service of some sort (similar to google's API), supply it a grammar (i.e. list of names from a voicemail directory), and have it be supported by a company who has a good knowledge of speech recognition (i.e. their recognition engine is robust and gives reliable results).
I would so love to use google's speech API, but it's not really production-ready, as google doesn't support it for commercial use (there's a quota and you can't supply custom grammar so far as I've been able to tell). Still, it's an excellent engine in so many ways. There's always been Sphinx, but I have never had the impression this has gone far beyond the academic, and from what I've seen it's not something to be taken on lightly. Then there are the commercial products, but I really can't figure out how to choose amongst them. Anybody had any good experiences they'd be willing to share? Jim
