Snips satellite - desire for mic on satellite, speaker on base

Hi snips team,

first fo all très grand merci to the whole Team for making this awesome project possible.

I have a Question: I want to run a snips satellite on a PI Zero with a respeaker 4-mic array. Neither PI-Zero nor 4-mic respeaker have an audio out, so i want to output to the base snips, being located at the television set, just 3 meters away. Problem is the mic won’t work nicely on the base unit if the TV is switched on, so i had to shout hey snips, therefore the idea of using a satellite. Besides that: I also don’t want a huge clunky satellite, so the speaker should be within the base unit, hidden behind the TV. (I hope you can guess the scenario)

I tried with that scenario and the mic of the satellite works nicely and i can see hotword detection using sam watch on the base unit, so traffic is properly posted from satellite to base using the MQTT broker of the base unit. Still the speaker seems to be expected on the satellite as well, which is not what i want to have. Is there a way to configure e.g. snips.toml to do the Playback on the base?

Many Thanks in Advance

ps: snips base is on a Raspberry pi model 3.

I’m not sure you can, but using pulse audio you can create sinks where to play the sound, including on the network. That could be a way for you to achieve what you want

Thanks @Psycho for the quick reply and the idea, using sinks. I have to admit i try to stick as close to the standard configuration as possible. My experience is, that with every new extra package and non standard way, the stability of the whole system suffers a bit and i end up debugging :grin:
Maybe i will order another respeaker 2-mic array, it has an audio out.

Just to understand: I assume the whole sound output also goes via MQTT to the snips audio server, so i would guess that it is just a matter of configuration of that audio server on the base unit, to play what it sees on MQTT anyway. Am i wrong?

Hello loriot,

The hotword detection configuration accepts multiple mqtt items for the audio option, like this:

[snips-hotword] section, uncomment audio = [+@mqtt] and change it for audio = ["YOUR_BASE_SITE_ID@mqtt", "YOUR_SAT_SITE_ID@mqtt"] .

If your satellite siteId is kitchen and your base siteId is living-room change the audio value for audio = ["living-room@mqtt", "kitchen@mqtt"] .

Maybe this will work for the audio-server bind option as well. I haven’t tried my self.