so I create a combined sink from two audio devices, my HDMI (DP actually) monitor with OK speakers and the analog stereo audio out, which got a subwoofer attached, without satellites. pactl load-module module-combine-sink, this works swimmingly, sound is awesome, I can set exact audio levels for each output, no notes.

problem is, I'm listening to an internet audio stream via Rhythmbox (tried mp3s off a network share, same deal) and when there's heavy CPU activity, the sound slows down, goes oooOOooOooOOoo like if you remember audio tapes from the olden days that got stretched/deformed. this only happens with the combined sink, if it's just HDMI audio, no issues.

used to happen infrequently, only when launching games or android studio doing compiling but nowadays with every other site having some cloudflare/anubis-like CPU torture, it's getting really tired, really fast.

tried looking at htop, pw-top, following the journal but nothing of substance is observable; possibly because this lasts for a second or so and those tools ain't exactly real-time.

maybe the issue is the devices being 48 KHz whilst the stream is 44.1 and there's some resampling happening? tried looking at pipewire/wireplumber documentation but that stuff is way above my head; "quantums", wtf?!

so is there some priority I can elevate or something so that this audio stack doesn't get compromised? I don't wanna give up the setup nor Rhythmbox.

Fedora, Plasma, Wayland, i5-4670, RX 580.

I'm using Jellyfin Desktop in TV mode. it simplifies the UI and reacts to cursors, enter, and esc for navigation. I got a bluetooth controller for android boxes I paired with my desktop and it emits those keys via InputRemapper. There's also a separate key used to start JD in TV fullscreen mode, with --scale-factor 2, got a 4K screen and the UI is tiny otherwise.

I always thought JD should be able to play youtube videos, but so far, no change on that front. So up until recently, I also ran macast as a DLNA sink so I can send it youtube videos and it would play them in fullscreen via mpv. Grabs focus, mpv's input.conf makes it react to the same keys I use for JD, exits when it's done.

But because I couldn't figure out how to pause JD if it's playing and resume after, because of the insane amount of log spam from macast, and because I needed a separate sending app on my android device (allcast) I tried to simplify the setup and came up with the following.

On my desktop I got this bash script:

#!/usr/bin/env bash

JELLYFIN_WAS_PLAYING=

case "$1" in

  *youtube*|*youtu.be* )
    case "$1" in
      *channel* ) : ;;            
      *)
        while IFS= read -r player; do
          [[ $player == JellyfinDesktop* ]] || continue

          if [[ $(playerctl -p "$player" status 2>/dev/null) == Playing ]]; then
            playerctl -p "$player" pause
            JELLYFIN_WAS_PLAYING=1
            export JELLYFIN_WAS_PLAYING
            break
          fi
        done < <(playerctl -l)

        mpv --fs --no-terminal "$1"
        mpv_status=$?

        if [[ $JELLYFIN_WAS_PLAYING == 1 ]]; then
          playerctl -p "$player" play
          unset JELLYFIN_WAS_PLAYING
        fi

        exit "$mpv_status"
        ;;
    esac
    ;;


  * )
    firefox --new-tab "$1" >/dev/null 2>&1 &

    ;;

esac

and I set it as default browser in Plasma.

So what I do now is just send the link via KDEConnect. Plasma sends it to my script, it checks if the URL is from youtube and plays the thing. If JD was playing at the time, it pauses it and resumes after mpv exits. If it's a normal URL, open it in firefox.

I'd prefer to use freetube to play the video, as I would like to save the URL to history, but I don't know how to make it play in fullscreen. So I figure, next best thing is to work out out how to add the URL to freetube's history.

Got any comments, suggestions, ways to improve it? Edge cases I haven't considered?

edit: typo in code

so I tried for the longest time to game on my 12-16 year old system with varying results. the constant issue was with Far Cry 5 - whatever I tried, it was always laggy, choppy, sometimes it's 30 fps, sometimes it's 50, altering resolutions, details, nothing worked or at least, no consistent results. my specs are above the recommended ones (i5-4670K, 16 GB, RX 580 8 GB, Fedora 44, Lutris, Wine 11.x) yet it always ran like ass.

so I read somewhere that the new NTSYNC is da bomb for CPU heavy games, which this one is, and in lieu of any succinct tutorial, I stumbled around and did the following:

1. sudo dnf install ntsync-autoload
2. sudo modprobe ntsync
3. turned off Esync and Fsync in Lutris

ran it with MANGOHUD_CONFIG=full to verify, and verily, there it is, says NTSYNC.

the game - forgetaboutit, butter smooth, constant 60 fps, no issues, no stutters, no frame gen spikes, this is like a new machine!

so it's possible this isn't the proper way to do it, but if you got issues, doesn't hurt to try.

anyone else? rolling back to previous flatpak solves it.

$ flatpak remote-info --log flathub org.jellyfin.JellyfinDesktop

Jellyfin Desktop - Jellyfin desktop client

        ID: org.jellyfin.JellyfinDesktop
       Ref: app/org.jellyfin.JellyfinDesktop/x86_64/stable
      Arch: x86_64
    Branch: stable
   Version: 2.0.0
   License: GPL-2.0
Collection: org.flathub.Stable

Download Size: 139.7 MB Installed Size: 323.2 MB Runtime: org.kde.Platform/x86_64/6.11 Sdk: org.kde.Sdk/x86_64/6.11

    Commit: a6d784a0c6db00f3d260e08c026a30072c2ec2ad7e902e8805033809296ddcb0
    Parent: 101468e6a77cf4cc4343f3069933025a86ebaf765eb662644dc6536003831c1d
   Subject: Merge pull request #18 from Justinzobel/runtime611 (03154a53a1a8)
      Date: 2026-10-30 22:50:12 +0000
   History: 

    Commit: 101468e6a77cf4cc4343f3069933025a86ebaf765eb662644dc6536003831c1d
   Subject: Add org.jellyfin.JellyfinDesktop (23eebe530cc6)
      Date: 2025-12-27 13:09:06 +0000

    Commit: a00879190e5e977f1185537e67426773db3944bddffc01ff9c0aaa9e7ab7e741
   Subject: Add org.jellyfin.JellyfinDesktop (23eebe530cc6)
      Date: 2025-12-27 13:09:06 +0000

the Oct 30 one doesn't play them, the 27 one does. also whole lotta fucked up UI glitches, doesn't respect the tv and desktop launch flags, some major changes afoot.

check out modded warfare's video. worked for me (PS4 Pro 13.04) first try; supposedly flaky on regular PS4?

forgot which sites are good, used to get my stuff from romsfun, but now it just gives me sass for every game I tried... if you have some rec's please share. fmhy.net has got you covered.

didn't see it mentioned here, so a brief YSK.

stefano marinelli is working on #littleFedi, a self-hosted, single-binary activitypub/fediverse peer that hosts a mastodon-like instance with, in my mind, three killer features.

one, the thing can run on your premises completely, no need for paying for a VPS, domain name, nothing, bypasses NAT/CGNAT, encrypted comms to other instances by way of lighthouses that do the initial intros.

dos, the thing runs on something as light as a Pi zero 2W, with a miniscule memory footprint and immense CPU optimizations due to running on the weakest hardware there is (also running off SD storage).

finally, a blog-like view of your long-format tweets/toots/posts, naturally with replies appearing as comments.

original post introing the thing: https://littleone.littlefedi.social/@stefano/4252ae6f-b949-42f8-a021-fd2b698f8b37

no code available atm (coming soon) same goes for binaries. the first instance is at https://littleone.littlefedi.social/ if you want an invite code to check things out (the settings page's got a lot of cool stuff), hit up stefano.

I am beyond stoked, this clears the two biggest hurdles for selfhosting your shit - the VPS/domain expense and hassle, and the hardware requirements for running it.

edit: missed that the website is up, explaining everything way better: https://littlefedi.org/

in its first iteration, it strips all AI shit, doesn't just disable it; the hero went through the source tree and patched it by hand!

it differs from LibreWolf and friends as there's an intention to hard-fork the thing (LW should be based on basebrowser) and let the abomination that is the mozilla shell-game-corp of today behind, thus returning to the original idea behind creating mozilla, best described by the dude who gave us the thing:

In my humble but correct opinion, Mozilla should be doing two things and two things only:

1. Building THE reference implementation web browser, and
2. Being a jugular-snapping attack dog on standards committees.
3. There is no 3.

mastodon thread: https://mastodon.social/@sarahjamielewis/116994970688711387

the patch repo: https://codeberg.org/basebrowserproject/organization

firefox repo: https://github.com/mozilla-firefox/firefox

build instructions: https://firefox-source-docs.mozilla.org/contributing/contribution_quickref.html

the only thing missing's how to apply the patch. it's as simple as going to the firefox source folder and applying the patch file with:

patch -p1 < the.patch

and then proceed with ./mach build.

this was my first time building firefox from source and it wasn't half as bad; about 30ish minutes on an i7-870 in the background without any excessive load that woulda made the system unusable.

In case you were one of the lucky ones able to snag a free account in the olden days, the terms have changed. Instead of 4 CPU and 24 RAM that you were free to whack up amongst 1-4 VPS, now you got 2 CPU and 12 GB RAM, so either one dual-core VPS or two single-core. If you keep using the old setup beyond a certain date (can't be bothered to look up when that is) you'll get billed; at reasonable rates, I'm sure.

If I'm reading this correctly, the block volume storage allotment hasn't changed, but their whole setup is hella confusing, so might be off on that.

Was really a matter of days when they was gonna axe the thing, they fired half a planet's worth of workers, I'm amazed they left even this much.

edit: I'm behind the times, this was reported more than a month ago: https://linuxiac.com/oracle-quietly-cuts-free-tier-ampere-a1-resources-in-half/

was away from django development for close to a decade, decided to jump in for a smallish project. don't wanna use my preferred IDE from way back when (pycharm) and am trying to use vscodium.

with the standard python and django extensions, the autocompletion/suggestion is... uh, lacking. what else do I need? a search in the marketplace nets tons of suggestions, don't know which are superfluous, overlapping, deprecated...

hated on the thing since I first spotted it and dismissed it, never to think about it again - doing everything wrong and then adding the shittiest of solutions, who's gonna fall for that, right? but that thing is gaining and this post and recent threads just make me irrationally angry.

oh, the thing is also vibecoded. if you had any reservations about letting this idiocy loose on your data, there's a relief, huh? head bozo detects ever-so-subtle AGI whilst slop-fondling and easily "verifies" 30K LoC that was shat out. stomach-churning shit.

not to mention, said bozo is doubling down on his fash idiocies. I wish I was making half of this up.

looking for non-systemd solutions out there and thought I'd share this, didn't see it posted.

if you haven't seen any of tony's videos, you're in for a treat. enjoy!

what's going on here? I don't remember hiding any community and I'm def not finding any option to un-hide it

I'm using t.programming.dev, if it's important