Is your feature request related to a problem? Please describe.
I have pulled in https://github.com/tetele/onju-voice-satellite/pull/77 to my local branch, and now am able to use full duplex audio. I in fact can play music from my Onju Voice, and still get responses from Assist running on the speaker. Unfortunately I have to either shout or put my face right up to the thing, because there is no echo cancellation right now.
Describe the solution you'd like
Ideally AEC could be enabled to cancel out the playback from the speaker and allow better recording while the media player is in use. This would probably require support in https://github.com/gnumpi/esphome_audio, and I can't find any evidence of discussion either here or there about adding this.
Is it possible to support this with the hardware in the Onju Voice? The original repo suggests it is possible.
Describe alternatives you've considered
Speaking loudly, pausing the speaker manually when trying to say the wake word
Checklist
Is your feature request related to a problem? Please describe.
I have pulled in https://github.com/tetele/onju-voice-satellite/pull/77 to my local branch, and now am able to use full duplex audio. I in fact can play music from my Onju Voice, and still get responses from Assist running on the speaker. Unfortunately I have to either shout or put my face right up to the thing, because there is no echo cancellation right now.
Describe the solution you'd like
Ideally AEC could be enabled to cancel out the playback from the speaker and allow better recording while the media player is in use. This would probably require support in https://github.com/gnumpi/esphome_audio, and I can't find any evidence of discussion either here or there about adding this.
Is it possible to support this with the hardware in the Onju Voice? The original repo suggests it is possible.
Describe alternatives you've considered
Speaking loudly, pausing the speaker manually when trying to say the wake word
Additional context
https://github.com/justLV/onju-voice?tab=readme-ov-file#limitations-of-this-release suggests ESP-ADF and/or ESP-Skainet is needed for this. I think it's referring to https://docs.espressif.com/projects/esp-sr/en/latest/esp32/audio_front_end/README.html which has docs on using AEC, but it's unclear to me whether this is compatible with the way esphome_audio is implemented.