-
-
Notifications
You must be signed in to change notification settings - Fork 270
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
weird quirks / issues that occurs randomly #2002
Comments
Is there a way to manually attach to the running instance of EasyEffect and capture some of running state and useful informations? similar to backtraces used by abrt and kde? Because without a reliable way to reproduce or a reason it's happening i doubt it would get fixed. |
As far as I can see there is nothing wrong with the new gate. But as it has different parameters it probably has to be configured in a different way.
In my opinion this is just the signal getting lower than the gate threshold. And as a result the gate attenuates it even more. Once you manually adjusts the volume up the signal gets above the gate threshold and the attenuation that was being applied is removed.
No. And based on the king of problem you are describing it would be pointless. If it is just the gate threshold not being high enough for the typi5 signal level you feed to it nothing unusual would be seen in debug or similar output. |
If that is the case then it has nothing to do with the gate plugin or even easyeffects. When the stream is disabled it is bypassed from the effects pipeline. |
This would assume it's pipewire / wireplumber bug triggered by EasyEffect? because when I force stopped EasyEffect the audio turns back to normal. If that's the case I think I need to report to upstream. (where to report though? pipewire-pulse or wireplumber?) |
It depends. When you bypass the signal does the volume level go back to what it is supposed to be while bypass is enabled? If yes then it is a gate configuration problem. But if the volume is not ok even while bypassing all effects then I don't think the problem is on our side. |
Ok, got it. I guess i'll wait for the next time it occurs and try it. |
This occurs just now:
Triggered by playing video in LibreWolf (Firefox fork). The symptom is complicated:
And, just as I'm experimenting, the EasyEffect now looks completely stuck. Anything going over
Edit: Oh, toggling Global Bypass makes the sound works again, without force stopping EasyEffect (And this time toggling Dlobal Bypass does make some changes unlike the beginning of this comment, the effect is similar to toggling |
This is normal. The bypass removes only the plugins from the pipeline. Audio will still pass through our virtual sink and the spectrum plugin.
The probability this is being caused by the gate plugin seems very low. I would expect a little bit of silence but not changes in frame rate. This seems to be related to latency (audio buffer size) or sampling rate changes.
Usually when Pavucontrol is opened PipeWire switches to a lower latency. So PipeWire's automatic latency switching seem to be related somehow... |
Is there anything suspicious in the system logs or maybe in PipeWire's logs? Some problem in lower level libraries may be happening on your side. At least on my computers I do not see this problem(I use Arch Linux). |
Searching through These are the only things that seem relevant: Nov 22 03:21:51 fedora pipewire[2465]: mod.protocol-native: 0x5566e5a4bd10: connection_data: client 0x5566e5c029f0 error -71 (Protocol error)
Nov 22 03:21:51 fedora easyeffects[3537]: pipe_manager.cpp:1375 Remote error res: Broken pipe
Nov 22 03:21:51 fedora easyeffects[3537]: pipe_manager.cpp:1376 Remote error message: connection error
Nov 22 03:35:52 fedora pipewire[2465]: pw.node: 0x5566e5d68f40: scheduling non-active node
Nov 22 03:35:52 fedora systemd[1176]: dbus-:[email protected]: Main process exited, code=exited, status=143/n/a
Nov 22 03:35:52 fedora systemd[1176]: dbus-:[email protected]: Failed with result 'exit-code'.
Nov 22 03:35:52 fedora systemd[1176]: app-flatpak-com.github.wwmm.easyeffects-130480.scope: Consumed 9.290s CPU time. (I was making a guess based on the order of the logs but looking at my previous comment time it's completely wrong) 3:16 AM UTC is the time I've sent this comment #2002 (comment) 3:20 AM UTC is the time I've toggled Global Bypass to bring back the audio 3:24 AM UTC is the time EasyEffect froze 3:27 AM UTC is the time I've force stopped EasyEffect And no coredump at all. > coredumpctl list --since=today
TIME PID UID GID SIG COREFILE EXE SIZE
Tue 2022-11-22 03:41:19 UTC 132701 1000 1000 SIGSEGV present /usr/bin/python3.10 16.5M (Not related to pipewire and it's long time after EasyEffect was force stopped and restarted) |
I posted some additional info in #1856 |
This is what it sounds like (be sure to unmute the video) : github.mp4 |
@arch-btw run |
@wwmm thank you for your help. I was able to capture with pw-top when it happened: |
github2.mp4 |
A latency of Which plugins did you enable? |
Sorry for the late response. I have:
In that order and in Convolver I have an impulse file loaded. |
The only situation where our convolver is known to generate noises is when PipeWire switches latency or sampling rate. But when this happen what is heard is not that continuous strong noise in your video. Unless PipeWire is crazily switching latency without stopping. But I doubt something like this could happen. |
I have just seen this problem happening on my computer and I could observe the following:
This is really weird because when the global bypass is enabled the pipeline becomes
And when the
Our null-sink is just an ordinary PipeWire module. In other words it is not easyeffects code. The only thing I can think of is that when "special conditions are met" something goes really wrong inside PipeWire when filters are loaded and once they are removed whatever it was causing the problem goes away. I am not sure about how to even ask PipeWire developers for guidance... We may be just a victim of some kind of PipeWire bug. |
@wwmm Wow great job!! I wish I had more to add but my knowledge on this is very limited. It does seems like it happens when I pause one application with an audio stream and then start the other, but it's difficult to confirm the actual "pausing" is the problem. It might just be that some applications are simply affected and that others aren't. Some of them appear to be immune to it. |
Quick update: killing the easyeffects process and then starting it again is a temporary fix for this. Up until now I kept resetting settings and re-importing my profile and while that works too, it's more work. I also noticed a helpful comment about restarting the service here, that might be even easier: #1355 (comment) |
Is this any relation to a bug I've been having lately? Using 7.0.7, distribution package, on Arch, with Pipewire 0.3.77-2, also distribution package. The issue seems to come up any time something starts or stops playing audio with the pipewire-pulse daemon, and doesn't seem to affect clients using pipewire directly. It manifests as all outputs briefly bypassing the output effects chain, and since my current output effect (convolver) does attenuate audio, this results in the audio briefly playing loudly. |
It depends. The bug in this issue is of the kind that comes with this type of noise #2002 (comment). I haven't seen it since the time this issue was opened. So maybe PipeWire has fixed it. As you are using the convolver and it does not handle PipeWire's dynamic latency switching well you may be able to workaround your problem by settings a fixed PipeWire latency. |
I found an alternative that seems to work, and I should suggest it to the autoeq.app maintainer(s) instead of Convolver. Basically, exporting the headphone preset as EqualizerAPO Parametric EQ txt file, then importing it into the EasyEffects Equalizer filter, using the APO button. The GraphicEQ presets use way too many band ranges and not as effective settings, and therefore can't really be imported into the Equalizer plugin. |
EasyEffects Version
7.0.0
What package are you using?
Flatpak (Flathub)
Distribution
Fedora 36
Describe the bug
Note: I can't manually reproduce it. It seems completely random to me.
The original comment is here: #1994 (comment), Digitalone1/EasyEffects-Presets#11 (comment)
I've encountered issues that are very similar to descriptions in #1994 (comment).
Audio distortion and volume change occurs randomly, disabling / excluding the application from EasyEffect doesn't help. I've not yet tried the Global Bypass (because it happens randomly, i can only try one thing at a time).
The reliable way to fix all the quirks is to kill EasyEffect in System Monitor, and the audio instantly turns back to normal. But then it may occur again randomly. It's not frequent but it does happen once in a while.
The most recent one occurs today (21/11/2022), the volume turns very low as i'm playing the music with Lollypop. This time i didn't force stop EasyEffect but instead try to change the volume level for a bit, then the volume turns back to normal, generally matching the description from original issue (except it's quiter for me):
Everything occured in the past that I could remember:
Enable
andExclude
are checked), fixed by uncheckingEnabled
I've used this loudness equalization preset, the same as the original issue. but the preset author state that it's more of a general EasyEffect issue than preset-specific issue.
Sadly because it occurs complete randomly and it doesn't occur frequently, I can't just open EasyEffect with debug output all the time.
Expected Behavior
No response
Debug Log
Debug Log
Additional Information
This issue wasn't present before. I can't say the exact timeframe but it's clearly before the changes in
Gate
that broke legacy version of EasyEffects Presets, as I never had the issue when i was using the old version.But this might be not relevant, it could also be a pipewire / wireplumber update that breaks my audio.
The text was updated successfully, but these errors were encountered: