search

arkq / bluez-alsa

Bluetooth Audio ALSA Backend
MIT License
864 stars 189 forks source link

connected airpods pro2, the music is played without sound #635

Closed luoyuyi111 closed 1 year ago

luoyuyi111 commented 1 year ago

Hello, I have a problem. I try to ask if you have the same problem. (bluez5.54 + bluealsa3.1.0)

The device is connected to airpods pro2 and plays music, but there is no sound. Strangely, only this airpods pro2 appeared. I got the air log and hci log, the audio is normal.

arkq commented 1 year ago

Try, latest release or better yet current master. Due to lack of resources, old releases are not maintained on per-release bases.

Also, which codec you are using with your airpods2?

bluealsa-cli list-pcms
bluealsa-cli info <PCM-path>
luoyuyi111 commented 1 year ago

Try, latest release or better yet current master. Due to lack of resources, old releases are not maintained on per-release bases.

Also, which codec you are using with your airpods2?

bluealsa-cli list-pcms
bluealsa-cli info <PCM-path>

I'm used SBC codec. I tried to update to current master code and got a sound. So ,this problem is the reason why the bluealsa version is too low.

I found a discrepancy to be more likely, this commit: https://github.com/arkq/bluez-alsa/commit/f105eb58dcc2cab56a6416b1bad115d3f45d08fd

arkq commented 1 year ago

I found a discrepancy to be more likely, this commit: https://github.com/arkq/bluez-alsa/commit/f105eb58dcc2cab56a6416b1bad115d3f45d08fd

I'm not sure I get you correctly. Does it mean that everything works as expected AFTER that commit or BEFORE? Because, as far as SBC codec is concerned, the https://github.com/arkq/bluez-alsa/commit/f105eb58dcc2cab56a6416b1bad115d3f45d08fd commit does not change anything (I think)...

luoyuyi111 commented 1 year ago

this commit mainly changes the algorithm for the timestamp variable。 The old calculation（V3.1.0） : 

timestamp += pcm_frames * 10000 / samplerate;

This violates the Bluetooth spec, Now change it to:

timestamp += pcm_frames;

Validate pass Thanks

arkq commented 1 year ago

this commit mainly changes the algorithm for the timestamp variable

Yes, yes. I've made that fix some time ago and I thought (I couldn't recollect exactly which codecs were affected) that the SBC was OK before that change. But apparently it was not :D

So, I'm closing this issue as resolved.