Listener doesn't always start the first time?

[MysterC]

I’m playing with mycroft on my AIY voicekit v1 (RPi3B + voicehat) and it worked out of the box with the Picroft image, but since I can’t leave well enough alone, I wanted to get it working on 64-bit OS (in theory some of the more math intensive stuff could run faster).

After running the git clone and install, I didn’t have any sound then I realized my pi user wasn’t in the audio group, so I added that, rebooted and sure enough I had access to sound devices (duh!)

From there, however, was another problem. Both with and without the AIY specific asound.conf, I had no control of the volume and damn is it loud at 100%.

I tried a number of things but the only thing that would give me back volume control was uninstalling pulseaudio

After I did that I had another issue, the first time mycroft-start was run, the audio listener would never start and even typed queries wouldn’t result in spoken replies (but I would get text responses in the debug dialog)

After running mycroft-stop and then mycroft-start a second time, everything seems to come together.

Has anyone else tried this combo (64-bit OS + AIY voicehat)?

Thanks!

[Lcstyle]

Seems related

A guide covering how I built my homebrew Mycroft device

Congratulations, several months ago I tried doing exactly this, installing picroft and core on a rpi4 several times and failed, kind of. I do have a similar setup to you now, except if I try to run the Mycroft gui on the lcd, there’s a conflict between the user running the alsa audio listener and Mycroft gui. Sometimes the audio input doesn’t work. Other times, the audio output doesn’t work. Either way it leads to a device that can listen but not speak, or vice versa. In the end, like you, I d…

[MysterC]

Thanks for this.

I did some more tinkering over the weekend and I think I understand what’s happening now.

First, Mycroft expects pulseaudio and doesn’t always play nice with direct Alsa devices, so I re-enabled PA.

Second, for some reason PA thinks the AIY Voicehat (not v2 voicebonnet, I don’t have that to test) has a hardware volume control, but in reality is does not.

It looks like I can tell PA to ignore the hardware mixer but I need to sus out some device names and configs since PA is new(ish) to me and has always “just worked” in the past.

Thanks!