JarbasAl
commented
1 year ago do you have a VAD plugin enabled? might be the same as https://github.com/OpenVoiceOS/ovos-listener/issues/17 (undiagnosed)
Closed goldyfruit closed 11 months ago
JarbasAl
commented
1 year ago do you have a VAD plugin enabled? might be the same as https://github.com/OpenVoiceOS/ovos-listener/issues/17 (undiagnosed)
goldyfruit
commented
1 year ago do you have a VAD plugin enabled? might be the same as OpenVoiceOS/ovos-listener#17 (undiagnosed)
Yes I do, the webrtcvad plugin but the plugin is correctly loaded.
JarbasAl
commented
1 year ago did you have wakeword/utterance upload enabled?
could this be coming from https://github.com/OpenVoiceOS/ovos-backend-client/issues/36
goldyfruit
commented
1 year ago │ listener │ │
│ sample_rate │ 16000 │
│ mic_meter_ipc │ False │
│ record_wake_words │ True │
│ save_utterances │ True │
│ wake_word_upload │ │
│ disable │ True │
│ url │ │
│ VAD │ │
│ silence_method │ vad_and_ratio │
│ speech_seconds │ 0.1 │
│ silence_seconds │ 0.5 │
│ before_seconds │ 0.5 │
│ min_seconds │ 1 │
│ max_current_ratio_threshold │ 2 │
│ initial_energy_threshold │ 1000.0 │
│ module │ ovos-vad-plugin-silero │
│ ovos-vad-plugin-silero │ │
│ threshold │ 0.2 │
│ ovos-vad-plugin-webrtcvad │ │
│ vad_mode │ 3 │
│ ovos-vad-plugin-precise │ │
│ debug │ True │
│ retry_mic_init │ True │
│ mute_during_output │ False │
│ duck_while_listening │ 0.3 │
│ phoneme_duration │ 120 │
│ multiplier │ 1.0 │
│ energy_ratio │ 1.5 │
│ wake_word │ hey_mycroft │
│ stand_up_word │ wake_up │
│ recording_timeout │ 10.0 │
│ recording_timeout_with_silence │ 0.8 │
│ instant_listen │ True │
│ continuous_listen │ False │
│ hybrid_listen │ False │
│ listen_timeout │ 45 
j1nx
commented
1 year ago Can you reproduce the issue with a restart of PulseAudio?
goldyfruit
commented
1 year ago Can you reproduce the issue with a restart of PulseAudio?
Yes, this is the same behaviour as when PulseAudio or PipeWire are restarted except that when it happens, I got a specific message in the logs:
Expression 'alsa_snd_pcm_poll_descriptors_revents( self->pcm, pfds, self->nfds, &revents )' failed in 'src/hostapi/alsa/pa_linux_alsa.c', line: 3665
Expression 'PaAlsaStreamComponent_EndPolling( &self->capture, capturePfds, &pollCapture, &xrun )' failed in 'src/hostapi/alsa/pa_linux_alsa.c', line: 3887
Expression 'PaAlsaStream_WaitForFrames( stream, &framesAvail, &xrun )' failed in 'src/hostapi/alsa/pa_linux_alsa.c', line: 4285So basically, an error occurs within the audio system making the voice service not reconnecting anymore.
In this case, the sound system most likely quickly recovers itself and just carry on, however when the voice service looses connection, it never re-connect again.
It is an old problem that surfaces.
This can be solved by some sort of microphone watchdog that quickly reconnects and/or restarts when it looses connection to the microphone. This does not solve the underlying reason of the sound system error, but sound issues/errors can happen often for different reasons without any clear cause. However the voice service should be able to identify that it lost microphone connection and recover from it.
goldyfruit
commented
1 year ago It is an old problem that surfaces.
I guess it would be valuable as well for the audio service.
Very randomly, the
listenerservice get stuck, consuming CPU resources. I was not able to really pinpoint the issue but it seems that the service is looping waiting for something.Here are the logs when it started to not been responsive anymore.
And here is a
stracePythonovos_listenerservice.I got this forever in
stracelog.For some reason, the service seems to look for this
/home/ovos/.mycroft/mycroft.conffile which doesn't not exist but I don't think this is the issue.