Thank you for your donation!


Cloudsmith graciously provides open-source package management and distribution for our project.


Difficulty with Linux Mint using Moode as an A2DP sink
#1
Good Afternoon,

I am attempting to output audio from my laptop, a Linux Mint 18 64-bit system with a Plugable USB-BT4LE bluetooth dongle. I can connect to my moode (4.2) audio player's bluetooth renderer as an A2DP sink (and moode's "Playback" screen reports the bluetooth connection from my laptop), but no audio is actually output by the moode in this configuration. I've tried test sounds, as well as selecting bluetooth as the audio output device in VLC on the Linux Mint laptop and playing a sound file.

I have, however, tried connecting my iPhone SE to the moode's bluetooth renderer in the same manner, successfully playing audio from various apps on the iPhone.

I have also been successful at playing audio to a Windows 7 laptop as an A2DP sink with the same Linux Mint laptop that has thus far been unsuccessful outputting audio to the moode.

Incidentally, although it's not a configuration I'll need, I was also unable to use the Linux Mint laptop as an A2DP sink for the moode.

So, to sum up:

Linux Mint ---> Moode: No sound
Moode ---> Linux Mint: No sound
iPhone ---> Moode: Sound!
Linux Mint ---> Windows 7: Sound!

I am using what I believe to be the latest bluetooth dongle firmware from Plugable (https://plugable.com/2014/06/23/plugable...-on-linux/)

Any ideas?

Regards,
William
Reply
#2
What version of Bluetooth software stack is Linux mint using?

-Tim
Enjoy the Music!
moodeaudio.org | Mastodon Feed | GitHub
Reply
#3
Linux Mint is using BlueZ version 5.37.

Regards,
William
Reply
#4
(07-28-2018, 09:30 PM)willyside Wrote: Good Afternoon,

I am attempting to output audio from my laptop, a Linux Mint 18 64-bit system with a Plugable USB-BT4LE bluetooth dongle. I can connect to my moode (4.2) audio player's bluetooth renderer as an A2DP sink (and moode's "Playback" screen reports the bluetooth connection from my laptop), but no audio is actually output by the moode in this configuration. I've tried test sounds, as well as selecting bluetooth as the audio output device in VLC on the Linux Mint laptop and playing a sound file.

@willyside

FWIW, I just plugged an ASUS BT400 USB-Bluetooth adapter into my Linux Mint 18.3 laptop. (I had to go find and download the appropriate Broadcom firmware before the adapter would load properly.) I see the same result as you.

The LM18.3 bluetooth daemon reports it is version 5.35 5.37, well behind the version 5.49 in moOdeOS.

As a result of my attempts to play test sounds through the connection, the moOde syslog has a ton of repeated bluealsa error messages
Code:
pi@moodeLR:~ $ tail /var/log/syslog
Jul 28 22:18:07 moodeLR bluealsa[1062]: /usr/bin/bluealsa: Unsupported RTP payload type: 1
...


I see that error discussed in https://github.com/Arkq/bluez-alsa/issues/13

Looks like the first order of business would be to compile the latest bluetooth stack in LM, as Tim hints. 

Regards,
Kent
Reply
#5
Thank you for pointing me to this, @TheOldPresbyope.

I've tried using a later version of bluez (5.48), but the problem remains. Reading a bit further in the post you referenced, it appears the root cause of the problem is with PulseAudio anyway, not with the version of Bluez running on the A2DP source (i.e., PulseAudio should be specifying a payload type of between 96 and 127, but instead it's specifying a 1).

That being said, it also appears that the problem can be circumvented on the RPi if bluealsa is configured with "--disable-payloadcheck" at compile time; see https://github.com/Arkq/bluez-alsa/

It's not clear to me whether bluealsa is being compiled during the moOde build process, or if it's installed from binaries at some point. @Tim Curtis, can you clarify? If it is being compiled, would it be possible (i.e., would you be willing) to utilize this option (--disable-payloadcheck) so that any of us trying to send audio from a computer running PulseAudio to the Bluetooth renderer will have a better chance of being successful? Let me know what you think.

Regards,
William
Reply
#6
Hi William,

The Bluetooth stack is compiled during the Build. Download the Build Recipe and refer to STEP 4, number 2.
http://moodeaudio.org/downloads/prod/bui...e_v2.5.zip

I'll run it by the Test Team and see what the Bluetooth guys think. If its an option that won't cause side effects or instability then I'll go ahead and add it.

-Tim
Enjoy the Music!
moodeaudio.org | Mastodon Feed | GitHub
Reply
#7
(07-30-2018, 09:33 AM)willyside Wrote: I've tried using a later version of bluez (5.48), but the problem remains. Reading a bit further in the post you referenced, it appears the root cause of the problem is with PulseAudio anyway, not with the version of Bluez running on the A2DP source (i.e., PulseAudio should be specifying a payload type of between 96 and 127, but instead it's specifying a 1).
...
Ah. Sorry for putting you through the unnecessary effort. And thanks for persevering.
Regards,
Kent
Reply
#8
(07-30-2018, 09:33 AM)willyside Wrote: Thank you for pointing me to this, @TheOldPresbyope.

...

That being said, it also appears that the problem can be circumvented on the RPi if bluealsa is configured with "--disable-payloadcheck" at compile time; see https://github.com/Arkq/bluez-alsa/

...
@willyside
FYI, I've reported back to the Test Team that I've successfully used my Linux Mint laptop to output music via BT to a moOde 4.2 player after rebuilding bluez-alsa with the payloadcheck disabled. 
My Android devices and my sole iOS device continue to work with moOde as expected, but you know the saying "Trust but verify". 
Still trying to understand from the Bluetooth specs what dragons may be lurking behind the disabled check.
Regards,
Kent
Reply
#9
Maybe @Koda59 or @badbat75 could have a look :-)

I'm getting close to releasing 4.3 along with a 4.2 --> 4.3 update and so this would be a perfect opportunity to slipstream this fix into the Builder as long as its low risk for creating regressions.

-Tim
Enjoy the Music!
moodeaudio.org | Mastodon Feed | GitHub
Reply
#10
(07-30-2018, 06:38 PM)TheOldPresbyope Wrote:
(07-30-2018, 09:33 AM)willyside Wrote: Thank you for pointing me to this, @TheOldPresbyope.

...

That being said, it also appears that the problem can be circumvented on the RPi if bluealsa is configured with "--disable-payloadcheck" at compile time; see https://github.com/Arkq/bluez-alsa/

...
@willyside
FYI, I've reported back to the Test Team that I've successfully used my Linux Mint laptop to output music via BT to a moOde 4.2 player after rebuilding bluez-alsa with the payloadcheck disabled. 
My Android devices and my sole iOS device continue to work with moOde as expected, but you know the saying "Trust but verify". 
Still trying to understand from the Bluetooth specs what dragons may be lurking behind the disabled check.
Regards,
Kent

PS - In all cases, the sender and receiver negotiate the use of the default SBC (subband codec) which is the only one this version of bluez-alsa knows. The resulting audio quality will be "average". 

@badbat75 has been working on including the AAC codec which means, in principle, we could up our game with Apple devices.

Regards,
Kent
Reply


Forum Jump: