3 ms·

How good is the recognition?

• catalogia 7y ago
  It seems it uses pocketsphinx or kaldi. I've never tried kaldi but I've tried pocketsphinx before and didn't find it accurate enough to be useful.
  • woodson 7y ago
    Kaldi is much much better, but when used for low latency recognition on the device, the accuracy will be lower than it could be because one would use much smaller models adapted to the constrained proccessing power of the device. It would still be much better than pocketsphinx.
• synesthesiam 7y ago
  The Kaldi models come from Zamia (https://github.com/gooofy/zamia-speech https://github.com/gooofy/zamia-speech) and their accuracy is quite high. They're a bit slow on a Pi, though there are some performance improvements coming soon. Rhasspy is designed to recognize user-specified voice commands, so the accuracy will highly depend on the complexity of your commands. If needed, you can also try doing open transcription: https://rhasspy.readthedocs.io/en/latest/speech-to-text/#open-transcription https://rhasspy.readthedocs.io/en/latest/speech-to-text/#ope...