This requires an updated XMOS firmware that outputs fully cleaned up and partially cleaned up mic audio
This PR adds a new nabu_microphone component that has two child microphones, one for each channel. The yaml file shows how to set this up to pass the AEC audio to mWW and the fully cleaned up audio stream to the voice assistant. There is an optional parameter to amplify a microphone to match the built in factor of 4 in the current ESPHome code (since the mWW audio channel doesn't have gain control automatically applied). I've also rearranged the logic around the wake up noise so that the voice assistant pipeline starts only after the wakeup noise has played.
This requires an updated XMOS firmware that outputs fully cleaned up and partially cleaned up mic audio
This PR adds a new
nabu_microphone
component that has two child microphones, one for each channel. The yaml file shows how to set this up to pass the AEC audio to mWW and the fully cleaned up audio stream to the voice assistant. There is an optional parameter to amplify a microphone to match the built in factor of 4 in the current ESPHome code (since the mWW audio channel doesn't have gain control automatically applied). I've also rearranged the logic around the wake up noise so that the voice assistant pipeline starts only after the wakeup noise has played.