Lately, I’ve been playing with the i²s audio stuff A LOT in ESPHome. I’ve found some details that I posted here: Original yaml of the 13$ atom echo voice assist? - #37 by SpikeyGG a couple of months ago. Basically, the Atom Echo won’t say anything in response if it’s configured as a media_player instead of a speaker. Lately, I’ve been building my own microphone + speakers combo and I’m finding the same thing!
I think this is a bug in ESPHome. When I program the device as a speaker the voice assistant will respond verbally after executing the command I issued with my voice. When I program the device as a media_player everything seems to work until the response is supposed to be issued – then the actual speakers don’t play any voice.
Also, I can go into the developer options in HA and manually send TTS voice for the device to play and it works very well so I know that the equipment works for that, it’s just at the end of the voice assistant response where it breaks.
