Speech Recognition on Raspberry Pi is very unresponsive

mattordev / coda

C.O.D.A is a smart assistant that I've began to rewrite. The new implementation makes use of ElevenLabs' voice synthesis and has many more features planned like GPT3.5/4 integration and home assistant capabilities

GNU Affero General Public License v3.0

1 stars 0 forks source link

Speech Recognition on Raspberry Pi is very unresponsive #2

Open mattordev opened 1 year ago

mattordev commented 1 year ago

Through internal testing, it seems when CODA is running on the rasp-pi the voice input is more miss than hit. It often misses the wake word detection, which in turn, means that no other audio is processed.

mattordev commented 1 year ago

This might be to do with the mic I'm using to input my voice. Will have to test on another microphone when I can.

mattordev commented 1 year ago

after coming back to this project it's more than likely that the Pi I'm using is underperforming - I suspect this to be the reason but don't have any addition hardware to test it on. It would make sense as it works a lot better on my PC and development machine but that's the only correlation I can find currently.