stable fiber
I have tried for about an hour to link my voice assistants to my Sonos speaker as an output , I have the media player entity there and it plays when using text to speech tests
I have checked under voice assistants and see no way to set what speaker to use
I have atom-echo-voice-assistant
by m5stack that listens to the wake word and can confirm it hears it
Does anyone know the method to set an output device ? I would love to see this in a drop down under voice assistants
fallen wren
I have tried for about an hour to link my voice assistants to my Sonos speaker a...
this is not officially supported.
I have a fork of the voice assistant firmware with a branch that allows for this. it has some issues and I haven't touched it in months so I cant guarantee it works on latest ESPHome version but its probably fine.
you can give it a go if you want - https://gist.github.com/MichaelMKKelly/79b6f5fcb85f424cb510dc4e3f841aff
fallen wren
yeah its got a speaker
its not very loud obviously cos its tiny.
the M5AE is great for testing/developing but its speaker/mic are realistically not good enough for normal voice assistant usage
for general VA usage then I suggest the Voice-PE it also has a 3.5mm line out so you can set it up with external speakers if you want.
stable fiber
I am taking it this cant be piped through Sonos media player either
fallen wren
I am taking it this cant be piped through Sonos media player either
no but if your sonos speaker has a 3.5mm input then you can plug it in directly
stable fiber
I dont hink it has that, it has ethernet though lol
fallen wren
i think some of them have a usb port too that allows for an adaptor to add 3.5mm in
stable fiber
fallen wren
yeah doesn't look like it has any input ports
stable fiber
Hmmm seems like a oversight all TTS can use this speaker
fallen wren
the voice assistant component that runs the pipeline runs in esphome on the voice device itself and calls the tools.
stable fiber
I mean I guess its all focused at the PE device
OHHH so thats why the custom FW mod is needed
fallen wren
yeah, what the firmware mod does is when the response audio stream is sent to the device the device gets muted and also sends the link to your specified media player to play too.
the VPE's speaker is pretty decent anyway if your only looking at voice. its not good for music though. although there's nothing stopping you using the VPE to control Music Assistant and play music to whatever media player you want.
stable fiber
I am just trying to move away from goofle
I would love to see a next hub like thing for HA
fallen wren
one of my voice pe's has a extra speaker so its a bit louder and better quality for music.
notes here: https://gist.github.com/MichaelMKKelly/5033dec56c5ab6ee6b7db52f690b84e0
audio/voice stuff is still in super active development. so will have to see what the future holds for stuff
hollow breach
the VPE's speaker is pretty decent anyway if your only looking at voice. its not...
Unless you're using jack. 48kHz 16bit PCM should be semi-decent for non-audiophiles. 🙂
hushed flame
@stable fiber I'm piping all the tts through my sonos speakers, including multiple rooms, each with satellite and sonos pairs. I put together a pretty custom firmware, but it may be overwhelming at first lol. Take a look and see if it helps. There are a lot more features beyond tts over sonos, but tts over sonos was the first thing I added. I have the automation and firmware shared. FYI, the README mainly talks about continue conversation and ask_question while using the sonos speakers in the assist pipeline. However, it does show how to do the basic tts over sonos. I do it by hooking into the voice assist pipeline and and an automation.
https://github.com/bofisher/atom_cont_conv
