· I hear distortion or noise when I'm playing music
What type of Zone is affected by this problem?
· *All of my Zones* are affected.
Does the issue affect all file formats?
· The issue affects *multiple/all* file formats.
Does the issue still occur with all Digital Signal Processing (DSP) disabled in MUSE?
· Removing DSP has no effect. I'm still having the same problem.
Does the issue happen with local library music, streaming service music, or both?
· *Only local library *music is affected.
Where is your local content stored?
· On the same machine where RoonServer is running, internal drive
Please try playing content of a lower sample rate (44.1kHz or 48kHz), does this work as expected?
· No, lower sample rates are still affected
Do you have an approximate timestamp of when the issue last occurred?
· I haven't been able to specifically reproduce as music is constantly playing and it occurs throughout the day.
What are the make and model of the affected audio device(s) and the connection type?
· Noted in previous answer. 3 C399, 1 M10, and one Intel NUC running Rock as endpoints. Roon server is debian-based.
Describe the issue
Every few tracks (but not usually every track) the first few seconds of play back is distorted (maybe up until ~5-10 sounds). It sounds out of tune or like it might be slowed until it readjusts. This is running on five grouped endpoints (Three Blusound C399, one M10, and one from an Intel NUC-based ROCK device) networked via ethernet with a local debian-based Roon server with local file storage. No external streaming services. Formats vary everywhere from low quality mp3 to 24-bit/196 flac.
Describe your network setup
Local 10g network based on a Cisco SX350X-08. There are a couple trendnet switches in place. All endpoints are hardwired with at least Cat 6.
Lately I also have experienced distorted playback on my Lyngdorf TDAI-3400, hardwired ethernet connection. It seems random and not repeatable. First distorted, later fading and disappearing music while playback visibly moves on. In the past three months, this happened twice.
Hi @rkjh31lkhekq,
Thanks for reaching out to us about this issue. To troubleshoot further we need to take a deeper look at your logs. Can you please use the directions found here and send over a set of logs to our File Uploader?
Thanks for sending those over! A few follow up troubleshooting steps for you to try:
Simplify your network chain - get a direct ethernet connection from your server to your router, bypass any switches or additional network gear in between.
If Exclusive Mode is enabled for the audio devices, try disabling it temporarily to see if it resolves the issue. Exclusive mode gives Roon exclusive control over the audio device, and sometimes this can cause conflicts with other system processes.
In Roon, try lowering the output resolution (e.g., from 192kHz to 44.1kHz or 48kHz) and see if the distortion persists. Some devices might struggle with high-resolution audio, and reducing the resolution could improve playback.
Ensure that the audio drivers on the PC or Mac and any firmware updates on the endpoints are fully up to date.
Test out disabling any DSP settings, including crossfading or volume leveling.
Test out using different cables, as well as ports, if possible.
Thank you for your patience so far. We’ve reviewed the logs you submitted to capture the specific track transition points you’ve mentioned above. Here’s what we can see happening:
When sample rate switches from 16/44 FLAC to 24/44 FLAC/MQA:
The five-endpoint group you referenced above performs a sample rate calibration to the new sample rate. Clock syncing occurs and the audio stream initiates. The NAD, however, sends a teardown request instead of initiating the audio stream; RoonServer instructs the device to prepare for playback and the entire grouped Zone resyncs once again. During this period, the ROCK-managed RAATServer imposes a 0.05 second resync delay.
Playback then continues without having stopped. There are no logged errors (within RAATServer and RoonServer).
You mentioned the digital distortion manifests as a “slowdown” effect, almost as if content is playing back at the wrong sample rate. Does this audible effect occur on all the grouped endpoints, or just some?
What happens if you remove the NAD C399 from this group, one at time? Based on the pattern we’ve observed, there is a single NAD C399 that isn’t behaving in coordination with the others.
Please provide a basic overview of where the managed switches fit in the topology between RoonServer, the three C399s, and the ROCK.
We don’t have any indication that the M10 is involved from logs, so I recommend changing none of the settings or network pathways relative to that unit at this time.
We’ll watch for your response - at your convenience, please upload another logset from the Debian RoonServer instance after the troubleshooting above.
Hey Connor, thanks for following up.
The distortion manifests across all endpoints.
It seems like it presents pretty consistently within the ~0:03~0:10 range when changing from 16 to 24-bit.
It seems to happen sporadically moving between 24-bit tracks as well.
I tried removing the C399s one by one and the issue persisted.
However, removing all three appears to resolve the issue (though this was very limited <10 minute testing).
Reintroducing any of the C399s re-creates the issue.
For network topology, the ROCK server and two of C399s are directly connected on the same switch. The other devices (a C399s and the M10) have a dumb switch in between.
It’s impractical to move the secondary-switched C399 at the moment, but the matter persisted after removing it from the chain entirely.
Thanks for your patience and for such a thorough and diligent response.
Here’s what we can ascertain from the most recent logset:
Sample chunks of consistent size are being re-requested by two of the C399s during the last live test with all three. This does not break playback, but it does point generally to either packet scheduling issues (common with Roon and managed networks), buffer overflows on a switch, or network congestion. Since two of the C399s are on the same switch serving the ROCK, let’s focus on those and ignore the C399 on the unmanaged switch.
For the managed switch serving the ROCK / two C399s, verify that any settings related to multicast forwarding and IGMP snooping are engaged. If there are QoS rules in place, try disabling them temporarily.
The M10 is reporting intermittent sample loss - curiously, it’s actually this unit that causes the unnatural dropouts that stop playback entirely. What happens if you remove this device from the group, leaving only the Intel NUC endpoint and the 3 C399s?
Our requests for diagnostic logging continue to fail despite the RoonServer reporting itself as online - it’s possible that stateful network security is interfering. We’ll have to request a manual set again at your convenience.
Thank you @rkjh31lkhekq and we’ll watch for your response!
The dynamics of this are interesting for sure.
A transparent squid caching proxy (http/80 only) was part of the problem with updating a Rock NUC and has been removed, though I’m unsure if that would be an issue here.
Also thank you for identifying the cause of the hard stops. I’ve noted that creep up as this issue has developed and had some concern where that was coming from.
BluOS pushed an update a week or two ago and I’m unsure if that’s had any impact though it seems more prevalent and lasting longer now.
I’ve moved from a Debian Linux server to an Intel NUC, and the issue seems somewhat more apparent.
Other changes are enabled IGMP and multicast on the primary switch as those weren’t enabled.
I intend to move the other two devices in the next day or two so all devices can be networked from the primary switch instead of the M10 and C399 being on an unmanaged switch.
I don’t have any particular rules in place to block Roon logging–are there any domains/IP/ports I can whitelist to allow for remote logging?
Since a few days have passed, I wanted to check in and see if you’ve run into any issues?
Unmanaged switches are actually much better to use compared to managed switches, which often cause issues during playback - I’d steer away from anything around managed switches.
If you don’t have any rules in place, you should be fine otherwise.
The issues persist, though they seem to be less apparent when the hubbed C399 and M10 aren’t grouped.
They’re somewhat inaccessible at the moment and I’d planned to move them this weekend, but family matters came up.
Hoping to move them this week so they can all be on the same switch and see if that improves things.
Are my logs accessible remotely now, or do I need to send another set?
Hi Ben, touching base.
Was able to move the M10 and furthest C399.
There was an RJ45 wall jack that I suspect was causing some issues.
They’re now hubbed directly into a switch off the main router without any wall jacks in between.
I haven’t had any cut outs/drop offs since moving them.
The audio distortions persist, though do seem reduced.
I plan on running direct lines in the next week or so to see if this helps.
Hi Ben, following up.
All devices have been directly networked to the main switch.
The intermittent stopping/disconnection problems seem to have resolved.
The distortion seems to have lessened quite a bit, though still does occur.
I noticed two incidences this morning at 11:58am ET playing “Am I Dreaming,” and then again at 11:59am ET playing “Calling”.
Thank you for your response. The latest diagnostics didn’t show the hallmark logged errors of digital distortion related to transcoding failures or under-processing during playback of those tracks.
RoonServer is downloading a 96Khz FLAC file to distribute in a Grouped Zone that includes both 7.1 and stereo Zones. Some of these are configured for 96Khz playback. RoonServer is downsampling to 48Khz for at least one Zone.
Logs show the clock lead adjusting between +500 and -200 samples to accommodate for the other Zones.
Can you elaborate on the precise nature of the distortion you’re encountering at this point, and which Zone (make/model of the endpoint) displays the distortion? Do you hear tracks speeding up (chipmunked) or slowing down? Are any errors displayed in Roon, or do you see the processing speed dipping below 1.0 in Signal Path?
Hi Connor!
The issue seemed less pronounced for a while, though seems to have started up again in the past couple days with random stops in playback reoccurring as well.
One endpoint is configured for 16-bit/48Khz as it’s a streaming box and for light sync only; however, these issues predate the addition of that box, and it’s otherwise direct connected the same as the other endpoints.
The distortion is a few second distortion typically at the 5~10 second mark, and depending on the track may not be painfully obvious. I’m not even sure it’s a slowdown so much as a change in pitch? Almost like the track is being shifted down an octave or two until it corrects. This distortion is audible across all endpoints as they play concurrently in different locations throughout the house.
All endpoints are configured with a 5000ms resync delay and 1000 zone grouping delay.
The endpoints are three NAD C399s and an NAD M10 all hardwired to the same switch as the ROCK. There is also a Marantz Cinema 30 and Hue Sync Box connected to the ROCK itself via its two HDMIs outs. There is a Bluesound Node connected via ethernet switch, though I’ve recently disabled that again to see if it’s contributory.
The issue is brief and inconsistent enough that I’m not able to monitor the signal path when it occurs just because I can’t anticipate it.
