Just getting started doing some basic keyword recognition on a live audio stream using an RPi 3... currently running the latest pocketsphinx/sphinxbase from github. Working pretty well, but I have one issue... noise (e.g. my mouse moving on the desk) while not speeking is being recognised as keywords. Is there any way to control the volume threshold for recognition within pocketsphinx config, or am I limited to reducing the gain on my mic?
Cheers
Gary
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Just getting started doing some basic keyword recognition on a live audio stream using an RPi 3... currently running the latest pocketsphinx/sphinxbase from github. Working pretty well, but I have one issue... noise (e.g. my mouse moving on the desk) while not speeking is being recognised as keywords. Is there any way to control the volume threshold for recognition within pocketsphinx config, or am I limited to reducing the gain on my mic?
Cheers
Gary
You need to adjust keyword threshold, not noise threshold.