Behringer UMC404HD stuttering / audio gaps

Talk about your MIDI interfaces, microphones, keyboards...

Moderators: MattKingUSA, khz

luke08
Established Member
Posts: 19
Joined: Mon May 02, 2022 11:09 am
Has thanked: 4 times
Been thanked: 11 times

Re: Behringer UMC404HD stuttering / audio gaps

Post by luke08 »

Thanks. I tried that on mine. With one speaker attached. Been playing for a few seconds now. I hear no stuttering in the white noise.
Cool, if your device is bcdDevice=1.12 I think that will work fine. No one lamented problems until 1.35.
Anyway, when you test with speaker-test please remain listening more that few seconds. Allow at least 20 seconds since the stuttering
on some platform has a much more relaxed frequency ( my Dell XPS i7 box stops every 8 seconds whereas Raspberry Pi
stops every 2.5 sec. ). It is also influenced by sampling freq: on 96khz stops are two times more frequent than at 48khz.
How can I check that myself?
I am afraid that you can't since your device is working properly :)
I agree. Do you think it's the hardware or the firmware change from 1.12 to 1.35?
It is surely the firmware upgrade. All posts from people with a bcdDevice of 1.12 confirms that the device behaves correctly and viceversa.
freshman
Posts: 2
Joined: Wed May 04, 2022 1:32 pm

Re: Behringer UMC404HD stuttering / audio gaps

Post by freshman »

Hey, just bought a brand new UMC204HD (indeed, not the same model) with bcdDevice=1.35, same error in dmesg output "clock source 41 is not valid, cannot use" and similar problems: pretty much exactly every 8 seconds, there is a gap in the audio. Running kernel 5.15 and ALSA version 1.2.6. Turning on implicit_fb=1 turns off atleast playback audio (didn't bother trying to record).

It seems like all the new devices come with an updated firmware that somehow breaks ALSA compatibility... Read some reports online of people swapping their U-Phoria devices for other models, but I think they just got lucky and got one from the old batch. Tried the device earlier on a windows machine, worked perfectly out of the box with no gaps in audio.
User avatar
Linuxmusician01
Established Member
Posts: 1523
Joined: Mon Feb 23, 2015 2:38 pm
Location: Holland
Has thanked: 756 times
Been thanked: 134 times

Re: Behringer UMC404HD stuttering / audio gaps

Post by Linuxmusician01 »

Are y'all sure its a firmware issue and not a hardware revision? I've tried to Google on the term "bcdDevice" but I'm not sure if that indicates firmware or hardware. Behringer might have changed one little (hardware) component that regulates a clock hence producing the dmesg output "clock source 41 is not valid, cannot use".
luke08
Established Member
Posts: 19
Joined: Mon May 02, 2022 11:09 am
Has thanked: 4 times
Been thanked: 11 times

Re: Behringer UMC404HD stuttering / audio gaps

Post by luke08 »

Linuxmusician01 wrote:
Are y'all sure its a firmware issue and not a hardware revision?
I think could be either of them. It is a product revision number. Considering that behringer devices are not firmware-upgradeable by users
they could increase just for firmware upgrades. Having said that, it is also likely that some hardware components internally could be changed.
But even in this case, the device is still marked as USB Audio Class 2 device and as such it should work without drivers.
I have tested in Mac Os X and works like a charm. So even if something changed in hardware, I still believe that is a bug in the snd_usb_audio implementation.

Linuxmusician01 wrote:
Behringer might have changed one little (hardware) component that regulates a clock hence producing the dmesg output "clock source 41 is not valid, cannot use".
Actually, when I rebuilt the module adding the quirks QUIRK_FLAG_CTL_MSG_DELAY and QUIRK_FLAG_IFACE_DELAY (that are required by many other devices producing the same error "clock source 41 not valid") , the error messages disappears from dmesg.
For instance: https://bugzilla.kernel.org/show_bug.cgi?id=108961
However the stuttering remains. Maybe there are there two different issues here?

freshman wrote:
Hey, just bought a brand new UMC204HD (indeed, not the same model) with bcdDevice=1.35, [...].
Running kernel 5.15 and ALSA version 1.2.6. Turning on implicit_fb=1 turns off atleast playback audio (didn't bother trying to record).
Yep this is consistent with another user experience in another post I read: implicit_fb=1 is a workaround for UMC204HD....
but not for UMC404HD !
On UMC404HD , loading the module with implicit_fb=1 makes playback completely silent... :(
freshman
Posts: 2
Joined: Wed May 04, 2022 1:32 pm

Re: Behringer UMC404HD stuttering / audio gaps

Post by freshman »

Linuxmusician01 wrote:
Yep this is consistent with another user experience in another post I read: implicit_fb=1 is a workaround for UMC204HD....
but not for UMC404HD !
On UMC404HD , loading the module with implicit_fb=1 makes playback completely silent... :(
Sorry if I was unclear; loading snd_usb_audio with implicit_fb=1 makes everything silent for my device as well...

Now I wonder what version of the device they were running, odd that it would work for their device but not mine!
User avatar
bluzee
Established Member
Posts: 338
Joined: Mon Nov 30, 2020 11:43 pm
Has thanked: 18 times
Been thanked: 88 times

Re: Behringer UMC404HD stuttering / audio gaps

Post by bluzee »

If it's working fine in Mac as well as Windows then I guess it's time to bug an ALSA devel for help. Obviously Behringer changed something, but if it still works as Class Compliant in Mac then unfortunately it must be the Linux driver that needs help.
luke08
Established Member
Posts: 19
Joined: Mon May 02, 2022 11:09 am
Has thanked: 4 times
Been thanked: 11 times

Re: Behringer UMC404HD stuttering / audio gaps

Post by luke08 »

Sorry if I was unclear; loading snd_usb_audio with implicit_fb=1 makes everything silent for my device as well...
Now I wonder what version of the device they were running, odd that it would work for their device but not mine!
Sorry, my fault, it was the UMC202HD not the UMC204HD, as reported in this blog:
https://nandakumar.org/blog/2022/02/umc202hd-linux.html

Of course I cannot confirm.

Anyway the situation is as bluzee wrote:
If it's working fine in Mac as well as Windows then I guess it's time to bug an ALSA devel for help. Obviously Behringer changed something, but if it still works as Class Compliant in Mac then unfortunately it must be the Linux driver that needs help.
Since it is a Usb Audio Class 2 compliant device, it should work. Considering that on other OS it works well, it is more likely a bug in alsa driver.
luke08
Established Member
Posts: 19
Joined: Mon May 02, 2022 11:09 am
Has thanked: 4 times
Been thanked: 11 times

Re: Behringer UMC404HD stuttering / audio gaps

Post by luke08 »

I have filed a bug for this at:
https://bugzilla.kernel.org/show_bug.cgi?id=215934

feel free to register and add to the CC list. If more users are concerned maybe gets higher priority.
Moreover add comments to report additional user experiences with the device can be useful too.
User avatar
Linuxmusician01
Established Member
Posts: 1523
Joined: Mon Feb 23, 2015 2:38 pm
Location: Holland
Has thanked: 756 times
Been thanked: 134 times

Re: Behringer UMC404HD stuttering / audio gaps

Post by Linuxmusician01 »

freshman wrote: Wed May 04, 2022 2:02 pm Hey, just bought a brand new UMC204HD (indeed, not the same model) with bcdDevice=1.35, same error in dmesg output "clock source 41 is not valid, cannot use" and similar problems: pretty much exactly every 8 seconds, there is a gap in the audio. Running kernel 5.15 and ALSA version 1.2.6. Turning on implicit_fb=1 turns off at least playback audio (didn't bother trying to record). [...]
[edit] I removed my reply. Saw that the 202 and the 204 are different devices and that the workaround for one does not work for the other. The 202 and the 204 look exactly the same on the front, however they are different on the back:

The Behringer U-Phoria UMC202HD:
Image

The Behringer U-Phoria UMC204HD:
Image


BTW, Notice that the bcdDevice number is the same for both the 204 (2 in- and outputs) and the 404 (4 in- and outputs). Both devices look the same on the outside, share the same specs and they probably are the same on the inside too. :idea:
User avatar
Linuxmusician01
Established Member
Posts: 1523
Joined: Mon Feb 23, 2015 2:38 pm
Location: Holland
Has thanked: 756 times
Been thanked: 134 times

Re: Behringer UMC404HD stuttering / audio gaps

Post by Linuxmusician01 »

luke08 wrote: Wed May 04, 2022 3:56 pm [...]
But even in this case, the device is still marked as USB Audio Class 2 device and as such it should work without drivers.
I have tested in Mac Os X and works like a charm. So even if something changed in hardware, I still believe that is a bug in the snd_usb_audio implementation.
Good point. I'd try to contact the ALSA developers first then, which I see you already did. That's kind of you towards the Linux community. I joined the CC list like you asked.
luke08 wrote: Wed May 04, 2022 3:56 pm Linuxmusician01 wrote:
Behringer might have changed one little (hardware) component that regulates a clock hence producing the dmesg output "clock source 41 is not valid, cannot use".
Actually, when I rebuilt the module adding the quirks QUIRK_FLAG_CTL_MSG_DELAY and QUIRK_FLAG_IFACE_DELAY (that are required by many other devices producing the same error "clock source 41 not valid") , the error messages disappears from dmesg.
For instance: https://bugzilla.kernel.org/show_bug.cgi?id=108961
However the stuttering remains. Maybe there are there two different issues here?
Hmmmm. Again, good point. However, could be one and the same problem. What I mean is: if one suppresses a valid error message then the actual (hardware) problem remains. This might be why you got rid of the error message whilst the audio stutter remained.
luke08
Established Member
Posts: 19
Joined: Mon May 02, 2022 11:09 am
Has thanked: 4 times
Been thanked: 11 times

Re: Behringer UMC404HD stuttering / audio gaps

Post by luke08 »

However, could be one and the same problem. What I mean is: if one suppresses a valid error message then the actual (hardware) problem remains. This might be why you got rid of the error message whilst the audio stutter remained.
Yes, I it could be either way. The delay addition does not suppress the message, the fact that it disappears should mean that the driver is able to drive the clock interface properly.

Indeed, if you can, try to load the module snd_usb_audio with implicit_fb = 1 on firmware version 1.12 to see if the playback goes silent there too
modprobe -r snd_usb_audio
modprobe snd_usb_audio implicit_fb=1
speaker-test -D plughw:U192k -r 96000
User avatar
Linuxmusician01
Established Member
Posts: 1523
Joined: Mon Feb 23, 2015 2:38 pm
Location: Holland
Has thanked: 756 times
Been thanked: 134 times

Re: Behringer UMC404HD stuttering / audio gaps

Post by Linuxmusician01 »

luke08 wrote: Thu May 05, 2022 7:10 pm
However, could be one and the same problem. What I mean is: if one suppresses a valid error message then the actual (hardware) problem remains. This might be why you got rid of the error message whilst the audio stutter remained.
Yes, I it could be either way. The delay addition does not suppress the message, the fact that it disappears should mean that the driver is able to drive the clock interface properly.
Thanks for clearing that up.
luke08 wrote: Thu May 05, 2022 7:10 pm Indeed, if you can, try to load the module snd_usb_audio with implicit_fb = 1 on firmware version 1.12 to see if the playback goes silent there too
modprobe -r snd_usb_audio
modprobe snd_usb_audio implicit_fb=1
speaker-test -D plughw:U192k -r 96000
Did just that. Am listening to white noise as I'm writing this. Weird, isn't it?
luke08
Established Member
Posts: 19
Joined: Mon May 02, 2022 11:09 am
Has thanked: 4 times
Been thanked: 11 times

Re: Behringer UMC404HD stuttering / audio gaps

Post by luke08 »

Did just that. Am listening to white noise as I'm writing this. Weird, isn't it?
Ohh, I suspected!
So there is something that is changed in the device behavior that makes it still a standard-compilant usb audio class 2 device
but does not play nice with snd_usb_audio driver in both standard and implicit_fb mode (both working in bcdDevice=1.12).

Thanks for the test, this is very interesting.

I also have found this:
https://bugzilla.kernel.org/show_bug.cgi?id=204631

that is related to a behringer device UFX1604 ( producing clicks during playback ) and seems correlated.
The patch extends the quirk already present for another behringer device (the UFX 1204) also to UFX1604 and clicks in playback disappear.
The interesting thing is that the patch for UFX1204 is in the kernel for since long time (5 years) starting from the version 4.15 in the 4.x trunk
but disappears in the 5.x trunk.
Since the 5.x version of pcm.c is quite different, I'm thinking to rebuild the 4.18 kernel extending the patch for UFX1204 to UMC404HD.
It's a just an attempt since the two interfaces could be radically different, but it is worth trying as a quick lucky fix.

If it fails, when I will have time I will recompile the kernel with dyndbg support start investigating if and why there are buffer overrun every k seconds.
Maybe I can produce useful data for ALSA folks.

I'll report here and in the filed bug as always.
User avatar
bluzee
Established Member
Posts: 338
Joined: Mon Nov 30, 2020 11:43 pm
Has thanked: 18 times
Been thanked: 88 times

Re: Behringer UMC404HD stuttering / audio gaps

Post by bluzee »

Have you heard anything back from Behringer?
luke08
Established Member
Posts: 19
Joined: Mon May 02, 2022 11:09 am
Has thanked: 4 times
Been thanked: 11 times

Re: Behringer UMC404HD stuttering / audio gaps

Post by luke08 »

Have you heard anything back from Behringer?
No, because Music Tribe gives me a headache.... it is slow and bloated, honestly I gave up during the registration for exhaustion.
If you have a working account maybe you can report what we are discussing there. Maybe they can shed light on something useful.
Post Reply