what could be causing interrupted playback ( silence ) of a few seconds from RPi4 using usb to AKM4493S based dac/amp?

[Roy_D]

Kind of stumped as to what is causing this. The RPI4 has an overspecced PSU, is cool due to heatsink case, not running anything except OSMC and playing back FLAC and MP3 and OPUS files which when inspected are all without flaws in their data BUT at seemingly random times, not often mind you, there is a break in playback - total silence for a second or two - and then playback continues. Rewinding the song that had the pause ( occurs anywhere in a given song of any filetype ) yields non interrupted playback to where before there was a pause in playback..

The RPI4 is set to output over USB in digital sp/dif mode with a limit of 192kHz. I have also tried USB Analog ( which resamples anything to 48kHz )
but no difference. The occurances are as frequent.

Stumped and feeling stupid =/

 

[phofman]

Anything in OSMC logs around the time of playback break?

 

[audio_tony]

Which software are you using for playback? If MPD then you can enable logging (if not already enabled) and that will report any issues.

Also check /var/log/kern.log which will report any USB issues.

 

[phofman]

IIUC OSMC uses kodi/xbmc internally.

 

[Roy_D]

Enabled logging just now, added specific component audio logging.. Lets see

 

[Joachim Herbert]

Network issues?

 

[Vincent Kars]

The RPI4 is set to output over USB in digital sp/dif mode with a limit of 192kHz. I have also tried USB Analog

Sounds a bit weird as S/PDIF is a protocol of its own. You can't do S/PDIF over USB.
USB audio is either UAC1 or UAC2.

Anyway, try to narrow down the problem. Use an other source e.g. a mobile with a USB-C out or a PC. If the problem persist there might be an issue with the USB receiver of the DAC.

 

[nerdstrike]

Assuming the music stops, then picks up where it left off: The observation could be the audio buffer emptying and not refilling in time. That leads me to think of higher level interrupts holding the request for the disk interface, e.g. storage was put to sleep, or a write operation is stalling and blocking everything. Without wishing to doomsay, a defective storage device could also cause that sort of thing. That one is easiest to check if you have an alternate media source like a USB stick laying around that you can play from to see if you still get the problems. Depends how frequently it happens!

If the music cuts out and returns as if it had kept playing, then I'd focus on the USB side.

 

[Roy_D]

Sounds a bit weird as S/PDIF is a protocol of its own. You can't do S/PDIF over USB.
USB audio is either UAC1 or UAC2.

Anyway, try to narrow down the problem. Use an other source e.g. a mobile with a USB-C out or a PC. If the problem persist there might be an issue with the USB receiver of the DAC.

On the RPI/KODI/OSMC the USB output has the option 'analog' which is 48Khz fixed, or 'digital sp/dif' which then allows you to set various other parameters and it is advised by devs to set max to 192kHz

 

[Roy_D]

Assuming the music stops, then picks up where it left off: The observation could be the audio buffer emptying and not refilling in time. That leads me to think of higher level interrupts holding the request for the disk interface, e.g. storage was put to sleep, or a write operation is stalling and blocking everything. Without wishing to doomsay, a defective storage device could also cause that sort of thing. That one is easiest to check if you have an alternate media source like a USB stick laying around that you can play from to see if you still get the problems. Depends how frequently it happens!

If the music cuts out and returns as if it had kept playing, then I'd focus on the USB side.

Thank you - forgot to mention first thing I tried was a different USB stick - no change =/

 

[Roy_D]

Network issues?

Playing from USB devices ( stick, hdd )

And then it does not occur for hours now as far as I can tell.. Lets see what the logs say in a few days cause my new amp is making me play só much music

 

[Roy_D]

Nothing in the logs, but thought to unplug the 2.4mHz dongle for the raspberry pi mini keyboard and when I did one of those glitches occurred. Keeping unplugged and on with the music - see if we found the culprit

...nope just glitch again. Sigh.

 

[phofman]

How do you get your source file? Locally, or over network/wifi?

 

[Roy_D]

How do you get your source file? Locally, or over network/wifi?

Locally. Off usb hdd or stick. Those check fine. File integrity is good. Temp of Pi is low, Changed all usb cables twice now, changed usb port used, updates on rpi4 all done.

This is going to have me grinding my teeth in my sleep

 

[phofman]

Nothing in dmesg either? Everything goes through your USB, IIUC, both input, output, control...

 

[audio_tony]

@Roy_D Try the following commands:

tail -f /proc/asound/card0/pcm0p/sub0/status

tail -f /proc/asound/card0/pcm0p/info

tail -f /proc/asound/card0/stream0

Your card name might be different - you can determine which cards you have in your system with: cat /proc/asound/cards

Leave a console open with tail as above, and when it glitches, note if any errors occur. I can't remember which is the key command so just try all three.

Also, are you using Pulse or PipeWire? Or just plain Alsa?

I prefer to use plain Alsa for headless systems, as I find that Pulse (should have been named PITA) isn't very good. I've no experince of PipeWire as I have no need for it (all my Linux boxes are headless). You only need Pulse or PipeWire if you need to share audio devices, e.g. if you have 'X' installed and need the browser to share the audio with other applications.

 

[Roy_D]

I will try this. Running with just plain alsa and just usb output as I found trying Pulse on a previous installation it introduced more issues than the added functionality (roomeq plugin) was worth..
OSMC now also officially advises not to use Pulse.

 

[phofman]

IMO it's quite unlikely that PA/PW would cause 1 second dropouts. IMO it looks more like issues with delivery of the source stream/file. Nothing in dmesg, no USB issues?

 

[Roy_D]

Nothing. No errors reported. It is entirely possible this issue is not on the side of the RPI4 and my contact with manufacturer of the dac/amp has hinted as much.
I am doing my due diligence in trying to eliminate issues on the source side and so far that has yielded nothing actionable.

 

[Roy_D]

I thought for some reason to time incidences of the pauzes; oddly it seems to be more or less exactly 30 minutes between instances.....