0

I have a request for a software contract that calls for recognizing spoken commands from a very limited set, such as "up", "down", "close", "open" etc.

I don't seem to need full-blown speech recognition, i. e. I don't have to do wave-to-text transformation. Is there a specific term I could google?

I simply don't know enough about the subject to be able to code that stuff myself. If I want to tackl ethis at all, I would need a fairly mature library for this subproblem, preferrably open source.

The target platform would be some kind of industrial PC running Linux and I would use either C or Python for the application.

Hanno Fietz
  • 30,799
  • 47
  • 148
  • 234
  • Actually, you *do* need a full-blown speech recognition system. What you're probably thinking of is a dictation recognition system, which is much harder than a "command and control" recognition system (which only recognizes utterances within a fixed, but possibly complex, grammar). CMU Sphinx, as far as I know, is a command and control recognizer, but I haven't used it for any real projects. – Eric Brown Mar 17 '11 at 20:21

2 Answers2

1

This looks like a suitable candidate: CMU Sphinx.

Haven't tried it, though.

Hanno Fietz
  • 30,799
  • 47
  • 148
  • 234
0

You could check out pyspeech, but I think it is Windows only. See Speech processing library in Python for speech to text

Community
  • 1
  • 1
Michael Levy
  • 13,097
  • 15
  • 66
  • 100