Audio setup quirks #6
Comments
|
Looks like there is something in my home directory that is not playing well with the audio setup. I created a dummy user "Test" and logged in as that account. Audio and mic worked. I'll keep this open until I find what is interfering. |
|
Unfortunately I couldn't find what was causing it. I created a new user, copied all my files over bit by bit and audio still works after doing that. I did find out that it might be something application-specific. Spotify loads without cutting out audio, but when I load other applications like QuiteRSS, Fluent Reader, Akregator, even Google Chrome the audio seems to break and will remain broken until I reboot. Firefox seems to be fine though. |
|
Okay, some interesting findings. If I swap pipewire and pulseaudio back (i.e. The audio works and applications don't break it, but then I encounter the bug you raised in https://gitlab.freedesktop.org/pipewire/pipewire/-/issues/1452 |
|
If you swap audio from pipewire to pulseaudio or vice versa be sure to open the audio mixer (pavucontrol) and choose the internal speakers for playback. I think when I move to pulseaudio it swaps to an HDMI output which doesn't do anything. Might be something silly like that and if it is let me know and I'll add a note in the audio section. |
|
You should only have to do this once (unless you want to intentionally change sources). It will persist through reboots, etc. |
|
Yes, the switch to pipewire was on the guide, which I did, and that's why I thought maybe that was related. By switching it back (e.g.doing the same command but switching the pipewire/pulseaudio options around) allowed the sound to work even when I loaded the apps that were causing it to cut out, but the mic has the known issue you have already raised. |
|
Here's a video: https://photos.app.goo.gl/qf6U5agFVUngQuQj8 I'm playing Spotify then try to load Google Chrome and the audio cuts out |
|
I can reproduce this easily with Chromium. All I have to do is start a youtube video in Firefox, then open Chromium and audio goes dead. |
|
Yes, for some reason Firefox is fine, but when you load Chrome, Chromium or other apps, bang audio cuts out. Spotify works OK, so you can also do the same test with Spotify open while playing audio (like I did in the video), then open Chrome/Chromium. I didn't find anything obvious in logs. |
|
Chromium and probably whatever other apps are breaking sound are launching the pipewire service, probably via the pipewire socket and that's what's breaking sound. You can do it just as well by launching a video in firefox then running I was trying to disable the socket with |
|
If you run If you ever need/want to undo this I don't think this is a Pixelbook issue so much as a Pipewire one... |
|
Thanks for the heads up, I will try that when I can get a chance. I've also subscribed to the bug on redhat. Thanks for raising that wit the details. Lets see what happens. |
|
Alright, got a response on the bug. Unmask the service, Then run the commands: This works for me. |
|
Perfect, I will give that a try later this week and let you know. Thanks :D
…On Tue, 7 Dec 2021 at 13:08, Jason Montleon ***@***.***> wrote:
Alright, got a response on the bug. Unmask the service, systemctl --user
unmask pipewire.socket
Then run the commands:
sudo dnf swap --allowerasing pipewire-pulseaudio pulseaudio
sudo dnf swap wireplumber pipewire-media-session
sudo dnf swap pipewire-jack-audio-connection-kit jack-audio-connection-kit
sudo dnf remove pipewire-alsa
This works for me.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#6 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAVA5O4LF4UI5J6KPTESTI3UPYBOBANCNFSM5JECI62Q>
.
Triage notifications on the go with GitHub Mobile for iOS
<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>
or Android
<https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.
|
|
Please feel free to reopen this or a new issue if it doesn't fix the problem. I've tested it and it's working for me. I've also updated the instructions in the README so future users won't hit the same problem. |
|
Thanks very much I will test that. I'm also working on an Ansible playbook that will run your instructions if that is of interest? |
|
Yes. A playbook would be great. Happy to help with it too. Feel free to submit a PR. |
|
Will do when I have tested, thanks.
…On Fri, 10 Dec 2021 at 21:02, Jason Montleon ***@***.***> wrote:
Yes. A playbook would be great. Happy to help with it too. Feel free to
submit a PR.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#6 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAVA5O7PQX2BCKR35ZRBNILUQJTHTANCNFSM5JECI62Q>
.
Triage notifications on the go with GitHub Mobile for iOS
<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>
or Android
<https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.
|
|
I just ran the new instructions manually (with the swapping of the connection-kit and wireplumber) and have noticed a few things:
This was run on a fresh Fedora install using defaults, with no new apps installed.
BTW, did you change the disk layout in any way when installing Fedora? |
|
Also, I don't appear to be able to reopen this issue |
|
If you only see headphones the pixelbook-alsa-ucm package isn't installed or something is otherwise wrong with alsa-ucm. |
|
Bluetooth sees devices for me. I can see a bunch of devices including my TV using blueberry and I have no issue with reboots hanging. For disk layout I typically just do /boot, /boot/efi, and root and swap on lvm, vfat for /boot/efi, and xfs for all else. Just my preference though. |
|
I'm going to wipe and try again. The packages, including alsa-ucm installed without errors, but I'll try again, both with the default btrfs and with the ext4, single partition I had in the original request which worked |
|
Did some tests using a fresh install (that uses the BTRFS filesystem)
Will try now using a standard partition setup (no btrfs, no lvm( |
|
Running the steps on a clean install on ext4 also had the same problem, and using the recovery image you used in your guide. |
|
I just noticed, in #5 the updated kernel you issued to fix the "It looks like a problem occurred" was 5.15.6-202.pixelbook.fc35.x86_64, yet the version that is now being installed is 5.15.7-201.pixelbook.fc35.x86_64. Could the new kernel version be doing something funny? I'm restoring the installation which we had the mic and audio working and then see if it offers an upgrade to the package (and then upgrade and see if it breaks) Broken: Working: |
|
Restored the working installation, and yes, it does offer the upgrade And yes, after the upgrade completed, and rebooted, the microphone disappeared and I mysteriously had my internal speaker under "Headphone" |
|
It's not the kernel. alsa-ucm was updated a few days ago and moved the configs into a conf.d directory. I'll fix the package and instructions. New package should be available in ~30 minutes. |
|
Awesome, will test with another clean installation when you're done. Let me know the new alsa-ucm version number so I can make sure its picked up during my testing, thanks. |
|
It just hit stable 3 days ago. Since you were reinstalling and updating you were probably first to see it. I see it this morning too; possibly also yesterday, but wasn't staring at it and my systems been powered on for a couple days, so likely wasn't affected until a reboot. https://bodhi.fedoraproject.org/updates/FEDORA-2021-90ca1a9fed |
|
The way this works is if you have audio at all the firmware and kernel are working. It just can't figure out the layout without alsa ucm. kbl-r5514-5663-.conf and HiFi.conf help with that. These files were taken from the ChromeOS recovery image as well and lightly modified for ucm2 and to fix a problem. If you run Builds running, I'll post back in a minute when done. Just run |
|
Builds done, just installed the new package from the repo and should be ready to go. |
|
pixelbook-alsa-ucm version 1.0.0-4-fc35 is installed, but I still see the Headphones option. You mention copying configs. Do I need to do that on the local machine, or was that inside the pixelbook-alsa-ucm package you referenced |
|
You should just have to reboot after installing it for the other options to appear As long as those files are in You shouldn't see any errors if you run |
|
That now works. I think I've also got the playbook working too. At the moment, just for audio (hence all these comments). I'll fork, add and PR the changes |
|
PR has been raised, #8 |
|
Thanks for submitting this. It definitely makes it easier copying over the firmware files. What about something like this: Added a small README with requirements. Also added a note in the main README
Probably some other things, but those were the ones I could think of. |
Typo: autoamte -> automate 😄 Also, I believe you need to add the hosts via the Ansible should also not be run as root (I did put that into the script which will refuse to run ansible if it detects the current user is
Hah, I spent ages trying to find out the right way of doing that and ended up using the lynx hack 😆
Do you only provide packages for fc35? If so, you might also want to check for the fedora version too.
That would make sense, it was hanging on processing
But the files were local, not remote. 🤷
Awesome, didn't realise that was an option, didn't know copr was a module itself, I was hunting in the
Yes, I believe the lack of
Generally should not run as root. Anything that requires root should use the Using root while running ansible also changes certain things inside ansible while running. For example, if you try to interact with If you run the ansible process as root, then you don't really need
They are great additions, I'll need to incorporate those into my app setup playbook as well, some tricks I was not aware of. |
|
Also, I had to remove my |
|
Some things I've found so far using your tweaked playbook:
This is fixed by replacing as follows: Finally, I have also discovered that, since I've now started installing my apps, including some snaps, that kpartx no longer mounts to loop0p3 because it can't, due to some of my snaps taking those slots: So the easy way to handle this is the add to the README the steps must be done ideally before any apps are installed, especially snaps, or record the output from |
|
Thanks. I changed the _files to _results to get rid of the _files.files stutter and forgot to fix the cleanup. Interesting with the snaps. I'll see if I can use the output from the commad to do something. Fun. Maybe I can do something to keep the archive too. If the playbook is pretty reliable hopefully people will only have to run it once. I'm not sure on the need of a hosts file; might be the difference of There might be an easy way to fix the split issue too; maybe just and |
|
You can refer to the version I submitted. If the archive already exists, it doesn't redownload it (it doesn't use a temporary folder though). For extracting the zip file, it checks if the extracted file exists (via I guess you could do something similar with the get_url module. i.e. check if the zip file exists and if it does, assume you don't need to download it again. Due to the way Ansible downloads, the file won't appear in the destination until the download is successful, so we can assume if the file is there, it's downloaded successfully. |
|
Alright, it will keep the zip file now; it will also guard against unpredictable loop devices I believe, add Let me know if you still see anything outstanding. My /etc/ansible/ansible.cfg is entirely commented out except one line; I don't see any user configuration and I don't have to specify hosts will -i I don't believe. Might be the only thing outstanding to see if I have something hanging around that's making it work for me. |
|
Thanks, I will test tomorrow when I am off call (need the pixelbook today in case I get a callout 😄) |
|
Seems to be fine. I think we can close this ticket off. As for the PR, you can probably close that as you have incorporated the changes in the repo. |
|
PR and improvements are merged. Thanks for your help with this! |
Trying to get the audio working using the instructions on this repo.
I have downloaded the recovery image (I used the exact same one as the guide, just in case), unzipped and copied the files into the relevant locations as described on the instructions
I get multiple output devices and input devices showing:
Speaker - Built-in Audio
Headphones - Built-in Audio
HDMI 1 - Built-in Audio
HDMI 2 - Built in Audio
Internal Microphone - Built-in Audio
Headset Microphone - Built-in Audio
None of the audio devices work, and neither do the microphone options.
If I connect my earbuds via Bluetooth however, I do get the audio coming through the earbuds, so the sound subsystem seems to work. Repeated reboots did not seem to help.
I tried doing the "Power + Refresh" option, but doing as described in https://support.google.com/pixelbook/answer/9134769?hl=en-GB (Turn off machine, leave for 1min, then Power on while holding refresh). This unfortunately also didn't help.
Any ideas if I'm missing something?
Side Note: I did copy my home directory over from Ubuntu so perhaps something in there might be interfering somehow?
The text was updated successfully, but these errors were encountered: