So, because I don't have a ReSpeaker and am just using a webcam microphone for input, I've had to alter a bit of code. I'm not sure if it's worth integrating into the code for others, but thought I would at least put my fix here. Maybe there is a better place for it?
in the glados.py file, simply specifying the device it needs to be listening to: audio_device='plughw:2,0'speech = LiveSpeech(audio_device='plughw:2,0', lm=False, keyphrase=glados_settings.settings["assistant"]["trigger_word"], kws_threshold=1e-20)
I wasn't sure how to specify which device with pulse, and so had to alter the sphinxbase init.py file to load alsa instead:
elif sys.platform.startswith('linux'):
*** REMOVED TRY EXCEPT BLOCK WITH .ad_pulse ***
from .ad_alsa import *
My only problem now is that it doesn't sense the keyword very well....or at all.... maybe recording volume problems. I'll have to play with pocketsphinx to see if I can get it to work better. I guess it's time I invest in a ReSpeaker!
Hmm... Nevermind... I obviously missed the option in the settings/yaml file :$
I also had disabled pulseaudio and needed to re-enable it. Everything is back to working just fine :)
So, because I don't have a ReSpeaker and am just using a webcam microphone for input, I've had to alter a bit of code. I'm not sure if it's worth integrating into the code for others, but thought I would at least put my fix here. Maybe there is a better place for it?
in the glados.py file, simply specifying the device it needs to be listening to: audio_device='plughw:2,0'
speech = LiveSpeech(audio_device='plughw:2,0', lm=False, keyphrase=glados_settings.settings["assistant"]["trigger_word"], kws_threshold=1e-20)
I wasn't sure how to specify which device with pulse, and so had to alter the sphinxbase init.py file to load alsa instead:
My only problem now is that it doesn't sense the keyword very well....or at all.... maybe recording volume problems. I'll have to play with pocketsphinx to see if I can get it to work better. I guess it's time I invest in a ReSpeaker!