Similarly to what the undermentioned states:
I’ve been having problems deciding a specific video too, via both org.mozilla.fenix and app.revanced.android.youtube.
Specifically, it plays smoothly for ≈ 1s, then becomes noticeably stuttery. It’s significantly worse in FF Nightly than in YT.
Rebooting and force-stopping doesn’t remediate it.
Build number: FP5.UT2E.B.078.20241105
Incredibly, disabling the “Animator duration scale” in Developer Settings to off [1] has remediated it.
Oh, so I’m not the only one experiencing something similar. I always thought it’s a quirk with ReVanced.
For me, it’s very rare and random when and how this happens. The only lead I could go on, is that it’s almost always a 4K video. Maybe even 4k60. Played back at maximum quality.
But turning the quality down to 1080p always fixes it. And since the colors of videos get oddly desaturated on FP5 when you play back 4k videos on youtube, I’ve just defaulted to 1080p all the time.
I’ve had all the animation scales set to 0.5 for a snappier feeling phone, ever since I have it. In case that changes anything.
@Nabalazs, I expected that to be the cause. However, it appeared to occur irrespective of quality. I tested 1080p and 1440p.
If only I could switch whatever dreadful video decodement library they use [1] for something like org.freedesktop.Platform.ffmpeg-full.
So you had it at 1.0 and now turned it off? Or was it at a different level before? While it may lessened the problem, I don’t think it’s the real culprit here. Or at least I suspect some sort of accomplice.
Just wandering… what’s the experience with other video platforms and have you tried different YouTube players?
Either way happy to hear you found a solid and simple workaround…and shared with us.
You use a FP5? This is a FP4 topic?
@JohnnyD, 1.0 (the default) to off.
It only took me about 6 minutes to get from problem to remediation, so I didn’t get a change to see. If I have the time to revert the preferences and check, I’ll do so and report here what I see. However, it was already rather unique in that it solely affected that video, so it would probably take some time to reproduce.
The best way to check would be to view it on a PC without proprietary encoding libraries installed, but I can’t be bothered.
Setting “Animator duration scale” to off doesn’t help. It isn’t event specific to a resolution.
This videos plays well without stuttering (2160p HDR 60fps): https://www.youtube.com/watch?v=LXb3EKWsInQ
this doesn’t for 1440p/2160p 60fps. 1080p 60fps is ok: https://www.youtube.com/watch?v=Mw1MWfilz_k
I have this issue often as well, so far every time a video has lagged it was 1440p. Setting to 1080 fixed it. But it’s still annoying to have to do that. Especially with the new quality menu. Happens in Shorts and regular videos.
Animation scale has always been on 0/off with the toggle to disable animations in accessibility settings enabled also.
Funnily enough, on my FP5 the first video stutters at 4K but plays fine at 1440, the second video already stutters at 1440.
@kuch3n, those are great examples. Playing Mw1MWfilz_k at 2160p via Fenix Nightly 136.0a1, I solely see key frames every ≈5s.
Does anyone here know how to evaluate which decodement library is being utilized? I suspect that FP has bundled OpenH264 or similar, in order to bypass royalty payments.
However, I’m no AOSP developer, and would have presumed that GMS would come with basic MPEG-4 codecs if this performance wasn’t so unacceptable.
Stats for Nerds gives VP09 (315) as used codec
@kuch3n, good point. Same for me:
I’ve obviously spent too much time wrangling with codecs on Fedora and OSTW. Apologies.
I have had the same issues with stuttering when changing specific resolution of videos only on yt, annoying it may be has there been any input from fairphone…
On that note, has anyone tested whether this applies to any 3rd-party AOSP distributions, like CalyxOS? (I reference that one because without GMS, its stack should be usefully different.) I’ll try to remember to report this to them directly via ZenDesk regardless.
If it does reproduce there, we can report it to their bug trackers to get some actual manpower on the issue. However, if it does, perhaps it’s a consequence of the SoC choice, since there’s usually hardware acceleration for standard codecs.
…On that note, does anyone have an example video that is only available in MPEG-4? Some videos on YouTube, presumably due to their source format, aren’t re-encoded during upload to VP9 (Opus); solely H264 and/or H265 MPEG-4. Testing whether those are better or worse would help, because it might just be the YouTube client choosing the wrong codec…?
@Aleksandr_Avezbakiye, I’d just like to compliment you on your exquisite music taste. You have shown me a new world.
