I have this one in order:
Curious to see if I can get my streamer working on it to work as satellite, it looks like a nice small device
1 Like
Nice finding. Thank you for sharing.
Would be great if you can also share details of your testing.
Thank you
cnxsoft did a review a couple of months ago.
1 Like
Yesterday I slapped some code in to see if I could get it going. I was checking on some examples on how the mic should work.
I am rewriting the streamer as a statemachine, so that is also part of this experiment.
The amount of code needed was very small and this is the result so far:
My plan for these devices is not to use as a satellite, but as AudioSwitches.
Each device should have a small set of specific wakephrases, maybe 1 or 2, which are directly linked to commands.
For instance, the phrases “turn light off” and “turn light on” will be two wakewords, which turn the light off or on.
A device can be installed in a room, lets so the bathroom, so when you come into the bath room and say “turn light on”(or maybe even shorter like “lights on”) the device turns the light on.
The phases should be implemented as ondevice wakewords, following this idea:
So, not overhead voice assistant, but just a simple audioswitch which does just 1 or a small amount of actions. There is no intent needed, because the wakewords or directly linked to a command.
I do not know what your opinions are on this idea, but I’'d like to hear them.
It came for the idea that I did not want an VoiceAssistant in every room, just a device reacting to voice commands to that specific room is enough in a lot of usecases. Just like a lightswitch 
2 Likes
Nice! I really like the idea of an “audio switch”.
Interesting to hear that you’re rewriting your code as as a state machine. Just yesterday I implemented an ESP32 project as a state machine for the first time and I really liked the conceptual simplicity of the approach and the clean code it resulted in.
1 Like
I am learning as I go, probably there will be design flaws, but it’s all in the game.
I have the M5 Atom Echo working as satellite but the TTS sound is really ‘Tinny’. Is this the best what can come out of it or can this be improved by changing volume, frequency, hermes settings etc?
Does anybody know if the sound quality of the m5 ATOM Speaker Kit is a bit better?
– EDIT
Changing the volume is not implemented for this device so this is not an option to improve the quality.