Good afternoon. I have a problem with the connection speed of Home Assistant Voice. The initial connection speed is around 135 Mbps, but after starting the player (radio playback), the speed drops to 6 Mbps, and the audio begins to stutter, making it impossible to listen.
The Home Assistant Voice device is located about 4 meters from the router, with a concrete wall in between. If playback is stopped, the speed slightly increases and may recover after some time.
Is there any way to solve this problem?
This sounds very similar to what I am experiencing now. Unfortunately I don’t have a solution for it at the moment.
Update
I was able to mitigate this issue, but I wasn’t able to identify a root cause. In my case this seemed to be a combination of some audio processing issues and electrical issues. Here are some details about my setup and the steps I took.
Update: Here’s a bit more about my setup.
- I am using Music Assistant to stream podcasts from Audiobookshelf to my HA Voice Previews.
- I confirmed that using Music Assistant to stream music from Spotify to the HA Voice Previews works fine (so Music Assistant alone seemingly isn’t the issue).
- I confirmed that I can stream Audiobookshelf to my laptop fine (so Audiobook shelf alone seemingly isn’t the issue).
- I confirmed that this happened regardless of the location of the HA Voice Previews. One was a few feet away from my network Access Point, and one was two rooms away.
- I am using HA Voice Previews connected to Anker Soundcore 2’s via aux.
Update: I am not sure exactly what the root cause was, but I was able to resolve this. Here are a few things that helped:
- (Specific to Music Assistant) Disable “Enable volume normalization” and “Enable limiting to prevent clipping” under the “Queue playback” section of (Music Assistant) Settings > Players > Player Settings. This resolved some stuttering that is closer to @MartinG-RU’s description.
- Instead of having separate USB-C power inputs to the Anker Soundcore 2’s and the HA Voice Previews, I used a dual USB-C power input. This resolved some “popping” which follows silence between portions of audio.
