Poor sound quality & suspect bluetooth.service

[seamusdemora]

After resolving my "build" issues, I'm trying to get some sound.
starting mpg123:

nohup mpg123 -a bluealsa --loop -1 ~/rainstorm.mp3 &
[1] 1731

FWIW: I've edited this post a few times. In the initial post, I had NO SOUND at all. I eventually realized from reading man mpg123 that there were output options I had never had to use with pipewire! After adding the -a bluealsa option to mpg123, I got low quality sound (an improvement!!). I suspect I'll need to do something similar for mpd/mpc. I'm now wondering if there are examples in the bluez-alsa documentation showing how to use various players? If not - maybe it's worth adding that?

and connecting w/ bluetoothctl:

[bluetooth]# connect B8:F6:53:9B:1A:97
Attempting to connect to B8:F6:53:9B:1A:97
[CHG] Device B8:F6:53:9B:1A:97 Connected: yes
[NEW] Endpoint /org/bluez/hci0/dev_B8_F6_53_9B_1A_97/sep1
[NEW] Endpoint /org/bluez/hci0/dev_B8_F6_53_9B_1A_97/sep2
[NEW] Transport /org/bluez/hci0/dev_B8_F6_53_9B_1A_97/sep1/fd3
Connection successful

I hear sound - but it's pretty awful... starting and stopping every second... I check bluetooth.service; it seems to have issues :

$ systemctl status bluetooth
● bluetooth.service - Bluetooth service
     Loaded: loaded (/lib/systemd/system/bluetooth.service; enabled; preset: enabled)
     Active: active (running) since Wed 2025-05-14 17:52:01 UTC; 10h ago
       Docs: man:bluetoothd(8)
   Main PID: 413 (bluetoothd)
     Status: "Running"
      Tasks: 1 (limit: 416)
        CPU: 735ms
     CGroup: /system.slice/bluetooth.service
             └─413 /usr/libexec/bluetooth/bluetoothd

May 14 17:52:01 rpi2w bluetoothd[413]: Bluetooth management interface 1.23 initialized
May 14 17:52:01 rpi2w bluetoothd[413]: profiles/sap/server.c:sap_server_register() Sap driver initialization failed.
May 14 17:52:01 rpi2w bluetoothd[413]: sap-server: Operation not permitted (1)
May 14 17:54:43 rpi2w bluetoothd[413]: Endpoint registered: sender=:1.18 path=/org/bluez/hci0/A2DP/SBC/sink/2
May 14 17:54:43 rpi2w bluetoothd[413]: Endpoint registered: sender=:1.18 path=/org/bluez/hci0/A2DP/SBC/source/1
May 14 17:54:43 rpi2w bluetoothd[413]: Endpoint registered: sender=:1.18 path=/org/bluez/hci0/A2DP/SBC/source/2
May 14 17:54:43 rpi2w bluetoothd[413]: Endpoint registered: sender=:1.18 path=/org/bluez/hci0/A2DP/SBC/sink/1
May 14 23:13:28 rpi2w bluetoothd[413]: profiles/audio/a2dp.c:load_remote_sep() Unable to load LastUsed: lseid 9 not found
May 14 23:14:47 rpi2w bluetoothd[413]: /org/bluez/hci0/dev_B8_F6_53_9B_1A_97/sep1/fd2: fd(27) ready
May 15 04:42:19 rpi2w bluetoothd[413]: /org/bluez/hci0/dev_B8_F6_53_9B_1A_97/sep1/fd4: fd(27) ready

What do I need to do to cure the "halting sound"???

[borine]

The logs you posted are fine and normal, there are no errors there and no indication that any problem occurred.

I don't have a pi zero 2 W to test with, but a quick web search turned up this issue on the RPI-Distro repo:
RPi-Distro/firmware-nonfree#33 (comment)

The file that is being edited there is:

/lib/firmware/brcm/brcmfmac43436-sdio.raspberrypi,model-zero-2-w.txt

It is not clear which rpios release is being used in the comment I linked above; the original issue was raised several years ago with the bullseye release, so may not be relevant for bookworm.

I have no way of testing this, or of knowing how safe it is. If you decide to try it, make sure you take a backup of the file first, and be prepared to completely re-build the sd-card from the original image file if it all goes wrong.

If firmware config, or other low-level OS config, is not the problem, then you will have to seek help from RPi experts. This does not appear to be a bluealsa (or pipewire), issue.

BTW

examples in the bluez-alsa documentation showing how to use various players?

No, at least not in the way you mean. If you want to know how to select a specific output device for an application then you must read the manual page for that application. It has nothing to do with bluealsa. If an application has a specific problem with bluealsa that it does not have with other output devices, then perhaps that would be worth documenting in the wiki. But it would need a volunteer to write it and maintain it; such volunteers are a very rare species indeed.

[seamusdemora]

then you will have to seek help from RPi experts.

Not much chance of that :) remember that these are the guys who designed WiFi and Bluetooth to use the same antenna :-/

The RPi2W is a "different cat"... it uses some chips that are not made by Broadcom. I noticed yesterday that a "new release" was made on May 13th 2025, so I'll re-install & start over... will let you know how that goes.

And thank you for your help!

[seamusdemora]

I'm creating a "step-by-step" here as I re-install & re-configure:

1. I checked the file mentioned in "the issue", and found that the "# Improved Bluetooth coexistence parameters" were already included just as Phil Elwell proposed them.

2. I learned earlier today that there has been a patch made to bluez for another BT issue (BT plays through only one speaker) I've had on multiple RPi models. I may be too old to hear before this patch is picked up in Debian/RPi, so I'll start by building bluez: ... uh-oh ...

Arrgh! Failed during make due to missing systemd obex file... also learned after submitting a PR that the GitHub repo for bluez is a sham... the actual repo is (somewhere) in the linux kernel repos. This is just chaos.