Neurocranium

Have you enabled a location module in the microG settings? I had exactly the problem you described on my recently installed LineageOS (`18.1-20211222-microG-sake`) and only after activating the Nominatim module...

> The initial voice detection process is pretty straightforward - > * Every time through the event loop, when waiting for voice audio, read a sample from the microphone (512...

Oh, good point. My setup is container based as well. Then I even could just volume mount my customized `voice_assistant.cpp` onto the right place in the container and would be...

Injected this little piece of debug code ```cpp case State::WAITING_FOR_VAD: { size_t bytes_read = this->read_microphone_(); ESP_LOGD(TAG, "Read %d from mic", bytes_read); if (bytes_read > 0) { vad_state_t vad_state = vad_process(this->vad_instance_,...

My logs and observations are pretty much consistent with what @rccoleman found. In the hard of hearing state, the Atom is constantly looping over reading from mic and getting a...

I ran into the very likely the same bug with a slightly different hardware configuration. ### NVIDIA Open GPU Kernel Modules Version 575.57.08 ### Operating System and Version Archlinux ###...