For the most part the voice pe is working great, but more often than I like it seems to stop listening.
I decided to connect the voice pe to my computer and connect the serial port. When the issue happens I see "Wake word model predicts 'Hey Mycroft', but VAD model doesn't".
Following is from the console:
[D][speaker_media_player.pipeline:114]: Reading FLAC file type
[D][media_player:074]: 'Media Player' - Setting
[D][media_player:081]: Media URL: http://192.168.15.199:8123/api/esphome/ffmpeg_proxy/27928b9c36a1afbca840cc6e65c2a2f8/gie4Sf9r7F4LOw95Fw9RDg.flac
[D][media_player:087]: Announcement: yes
[D][micro_wake_word:355]: Detected 'Hey Mycroft' with sliding average probability is 0.96 and max probability is 1.00
[D][media_player:074]: 'Media Player' - Setting
[D][media_player:078]: Command: STOP
[D][media_player:087]: Announcement: yes
[D][micro_wake_word:352]: Wake word model predicts 'Hey Mycroft', but VAD model doesn't.
[D][micro_wake_word:352]: Wake word model predicts 'Hey Mycroft', but VAD model doesn't.```
The announcement played and then several minutes later I tried to issue a command and got the VAD model doesn't message.
Tried some searches and did not find any hits?