Remember to use soxr as per Paul's guidance.

audio-swresample-o=resampler=soxr,precision=33

File: mpv-shot0023.jpg (278 KB, 1920x1080)

278 KB JPG

sub-ass-override=force
profile=high-quality
vo=gpu-next
volume=45
glsl-shaders="~~/shaders/Anime4K_Clamp_Highlights.glsl;~~/shaders/Anime4K_Restore_CNN_M.glsl;~~/shaders/Anime4K_Upscale_CNN_x2_M.glsl;~~/shaders/Anime4K_AutoDownscalePre_x2.glsl;~~/shaders/Anime4K_AutoDownscalePre_x4.glsl;~~/shaders/Anime4K_Upscale_CNN_x2_S.glsl"
audio-file-auto=fuzzy
audio-file-paths=.

>>102630975
KILL YOURSELF

File: 1725731589532536.png (297 KB, 1832x1308)

297 KB PNG

mpv won

File: 1696887083841.png (1.21 MB, 1206x1664)

1.21 MB PNG

I hate mpv so much

>>102631289
why the fuck does iina have so many stars

File: file.png (9 KB, 256x256)

9 KB PNG

Let me guess...

>>102631289
What a retarded picrel.

Retards stargazing a repo fork (ffmpeg) lol.

Michael Niedermayer would rather nuke ffmpeg before he learns to use github/gitlab/gitea/... instead of his mailing list scripts.

He considers those scripts more valuable than any development workflow improvements for everyone else, and more valuable than the potential of attracting new developers/contributors.

It's always hilarious to see how he gets panicky and starts to dance around the subject and does his best to put up manufactured bureaucratic blocks on the way every time this subject comes up.

>>102632224
Wasn't FFmpeg supposed to move to github at some point? The mailing list is such a shit show and the bug tracker basically has no eyes on it.

>>102625413
Of all filters, they decided not to add the specific one I need ;)

I'm currently trying to get GPU acceleration for motion interpolation. I'm trying to build fftw with openmpi + rocm for mvtools, if that doesn't work i'll try to build rife-ncnn

>>102633126
just wait for haasterpolation, a gpu accelerated filter for motion interp in libplacebo!

>>102633198
>haasterpolation
no google results for this, are you the one writing it?

>>102632325
specifically github, i don't think so.
but any talk of a move to a similar workflow is immediately shutdown as described in the parent comment.

here is the last example that comes to mind. the man proposed a hilarious world domination 5 year plan where the slowly dying ffmpeg will somehow magically develop multiple products that will somehow manage to compete with multiple multi-billion dollar startups. and AI gets a mention too of course.


https://ffmpeg.org/pipermail/ffmpeg-devel/2024-April/325798.html

but then the small suggestion of switching to a pull request based interface so patches don't keep getting forgotten forever is made. and he starts losing his mind:

https://ffmpeg.org/pipermail/ffmpeg-devel/2024-April/325915.html
https://ffmpeg.org/pipermail/ffmpeg-devel/2024-April/325929.html


>muh too much work
>how about another useless level of indirection using trac?
>muh LKML works

lol

>>102633322
>unironically suggesting that ffmpeg should turn into a discord clone
multimedia is doomed

yup i'm sticking with ffmpeg4

i think mpv's commit quality would improve if it went back to the mplayer2 mailing list

>>102633482
nu-mpv's problem is not the development workflow.
on that subject. i asked the other week if someone had the mplayer2 list archives, and no one responded. shame.

>>102633512
>i asked the other week if someone had the mplayer2 list archives
ask pigoz

>>102633303
Kek, man, I do feel for ya

SIMD queen fixing shader compilation
https://github.com/mpv-player/mpv/pull/14962

pedotroon thread

>>102634651
No need to announce you arrival

>>102633512
>nu-mpv's problem is not the development workflow.
What is the nu-mpv problem?

>>102633322
This is a well known issue. People offered help to host and maintain the gitlab instance, but no real progress is being made. Ultimately the issue is not with the tools they use, but with the attitude towards contributions and patches. No one cares and unless you have push rights already it is almost impossible to contribute even a small patch. The new tools though would expose those issues. At this point I have no hope they will improve, they are gatekeeping ffmpeg to few selected people.

>>102630767
prebiously >>102583960

>>102633010
you can ignore my conspiracy theories and chuck this as a "never attribute to malice ..." moment on nu-mpv developer's part.
you can also grab this opportunity and become a certified nu-mpv dev. just submit a pull request that adds "--enable-gpl" to the end of line 149 in the build-mingw64 file under the ci folder.
actually, every wintard is welcome to try this. whoever manages to do this first shall win the prestigious prize.

profile=high-quality

Rate my config.

>>102638521
this but

profile=high-quality
loop-file=inf

Other than deband is there any way to decrease compression artifacts with mpv?

How do I watch a 1440p youtube live stream on mpv?

>>102639874
set deband-grain to 4096

>The librempeg version of surround filter is more optimized for speed and quality
Anons, do you use librempeg already? Go discuss here https://github.com/librempeg/librempeg/discussions

>>102633322
Neural network filters make sense. Ffmpeg has a lot of filters, some of which are just for getting data from the video such as motion vectors. Why not support neural network filters as well?

>>102640994

mpv [LINK TO LIVESTREAM]

should work fine.

>>102630975
upscaled anime looks so fucking bad
you can just tell it's off and shouldn't be like that, stop running that stupid shit.
>>102631712
it's just mpv but with a focus on mac. Stock mpv on mac sucks, I can't even load the high quality profile or a cfg file in general without it crashing.

>>102642844
It works only up to 1080p

>>102642945
>I can't even load the high quality profile or a cfg file in general without it crashing.
Sounds like some seriously bad bug if true

>>102644394
ye, not sure how it is for other people, but most people that know about mpv use iina on mac.
I was too lazy to report the issue or try troubleshooting it much. The few things I found while searching was just people saying to use iina.
The high quality profile and other mpv commands work fine on here. It also already has thumbnails, a decentish ui, etc. so I didn't need to go digging for luas either.

>>102630767
this humie needs some feedin'

>>102644470
Try bundle from CI, it should work. If not would be nice to report it.

add libjxl to ci

>>102644986
you do it, it is easy

new chinese 0day just dropped
https://github.com/mpv-player/mpv/pull/14973

>>102644470
sounds like precedent and a good showcase of why support for windows should live in a separate fork!

>>102645727
what does windows have to do with macos?

>>102645521
lmao at the submitter not anticipating the NWDC "not windows, don't care" replies from the nu-mpv kindergarchy.
i don't remember what was the discussion or where it took place, but i remember "vfork considered dangerous" and all the debate around it being revived lately. maybe the submitter stumbled upon the same discussion(s) and decided to do something useful with some newly acquired knowledge. should have known better.

>>102645727
what the fuck does macos have to do with windows

>>102645835
not a single person mentioned windows in that PR. what the fuck are you talking about

>>102645796
if a posix but sufficiently different from the open-source stack platform like macos benefits from a dedicated fork. then i think it logically follows that windows would benefit too. mpv proper can then fully focus on core functionality and fully open-source stacks/platforms.

File: Screenshot 2024-10-01 213525.png (14 KB, 1065x70)

14 KB PNG

I want this to use my dedicated gpu but mpv.conf doesn't exist.

>>102645894
you sound like a nocoder winbabby

>>102645978
Create folder named portable_config in the same place with mpv.exe, then put mpv.conf (or any other thing) inside that folder.

>>102645978
just make the file yourself holy shit

>>102646002
really?
how did you figure?

>>102646377
it wasnt subtle

guido love
https://github.com/mpv-player/mpv/issues/14966#issuecomment-2387545042

>>102646396
funniest part is the double-wrongness.
i'm a coder and very much not a wintard. lol.

>>102646459
>im a coder!!
lol... no one calls themselves a coder. first you got exposed for not actually using 0.29.1, then the braphogs,,, now this...

>>102630767
As a mostly normie, would I notice any significant difference between MPV and VLC, or do I need to be autistic to appreciate it?

>>102646522
vlc just works

>>102646522
yea

>>102646521
not a 0.29.1 memer either.
maybe stop watching and schizoing in these threads too much.

>>102642945
>>102644470
nigga, what? where did you get that shit from? never happened anything like that

>t. mpv on mac user

Why are 10 bit videos smaller? Are they encoding at a lower bitrate? The file size seems enticing, but there must be some drawback?

>>102642453
ffmpeg is dead project
It is 99.9% gatekeeping independent patches from independent contributors that do not have financial deals with FFlabs.

>>102644986
>add a zeroday to ci

anyone know a libplacebo based player written in rust?
actually, is there any libplacebo based player out there other than mpv?

>>102648696
plplay, ffplay, vlc

>>102648033
>ffmpeg is dead project
It is interesting how stagnant FFmpeg has become, from external contributions to its infrastructure and build system. They only fix Coverity static analysis warnings, and that's the extent of the improvements we see.
> ./configure
> No CI at all, except for the external FATE builders
> mailing list
> a huge codebase without maintainers responsible for specific parts
> external contributions get memoryholled
> Paul left
It's like we are still in 2010

>>102649719
>ffplay, vlc
do they use libplacebo for vo?

>>102650122
ffplay for vulkan backend
vlc, I don't know, libplcabo is integrated, but to what extent I have no idea. Wait for VLC4

>>102633010
grab a new build from here and test:
https://github.com/mpv-player/mpv/actions/runs/11144043342

>>102650137
>Wait for VLC4
no. thanks.

>>102650225
aware

>>102650215
It werkz again, thanks.

>>102650137
>ffplay for vulkan backend
i didn't know about that. thanks.
i actually build ffmpeg without libplacebo support to avoid the circular dependency ffmpeg -> libplacebo -> ffmpeg.
is haasn going to fix that?
also, is it because it's disabled in my build, or is -vulkan_params not documented at all?
and what's with the random use of underscores in argument names?

>>102650309
cool.
i don't use or condone the use of windows btw. lol.

>>102650333
meds

>>102650137
>Wait for VLC4
I wonder if that will come out before I die

>>102650137
>vlc, I don't know, libplcabo is integrated, but to what extent I have no idea. Wait for VLC4
vlc4 uses it for vulkan and tonemapping

>>102650312
libplacebo only depends on ffmpeg for demos/test. The actual library doesn't use it.

>>102650702
are you telling me i should package libplacebo without plplay? that's a sacrifice i'm... totally welling to make. thanks.

https://github.com/mpv-player/mpv/pull/14977#issuecomment-2388678862
>Rejected on IRC.
who rejected it?
the FF mafia representative?

>>102650517
hey.
even my backup to the backup to the backup github account is an MPV contributor now.
RESPECT!

>>102631712
chinks

>>102630867
>precision=33
Why?

>>102651018
> the FF mafia representative?
pretty much, since the reason was that "only the official ffmpeg is certified"

[21:37:09] <llyyr> but anyway, if we're building ffmpeg forks then there's more useful ones. like rpi or the yt-dlp one that allows hevc in flv
[21:38:43] <kasper93> link?
[21:39:51] <llyyr> it's just a patch https://github.com/yt-dlp/FFmpeg-Builds/blob/master/patches/ffmpeg/master/0001-Nonstandard-HEVC-over-FLV.patch and rpi one has like multiple different trees
[21:40:29] <llyyr> not that I'm seriously suggesting we provide these builds, my point was that by adding that we're just opening ourselves up to more requests for ffmpeg flavors
[21:40:38] <llyyr> if someone wants to use a fork they can build it themselves, it'd be trivial with mpv-build
[21:42:34] <gh-mpv> [mpv] kasper93 closed pull request 14977: [RFC] ci/mingw: add librempeg builds -- https://github.com/mpv-player/mpv/pull/14977

kys jeet

>>102651600
the real reason: llyyr is still holding grudge towards Paul because he was the main ffmpeg maintainer who kept rejecting his low-quality patches

he would never allow librempeg to thrive for political reasons

does having a higher refresh rate display require more resources and power to play videos? let's say I play a 30Hz video on a 360Hz screen. the computer has to interpolate and repeat the same 12 times, so that's 6x more processing than interpolating on a normal 60Hz display. is that correct?

>>102652126
If you use display-resample, yes

>>102650994
okay. i updated my builds and enabled libplacebo in ffmpeg.
with ffplay+vulkan, i get these warning/info log lines:

[AVHWDeviceContext @ 0x566217a49e40] device extension "VK_EXT_hdr_metadata" not found, excluding.
[AVHWDeviceContext @ 0x566217a49e40] device extension "VK_EXT_image_drm_format_modifier" not found, excluding.
[AVHWDeviceContext @ 0x566217a49e40] device extension "VK_KHR_portability_subset" not found, excluding.

is my AMD GPU showing its age, or is support for this in mesa limited?
also, i checked ffplay code. I thought -vulkan_params would provide functionality similar to --libplacebo-opts in mpv, but it's of much less utility if i'm not mistaken.
so putting the plplay demo aside, there doesn't appear to be any players that fully utilize libplacebo.

>>102652252
>so putting the plplay demo aside, there doesn't appear to be any players that fully utilize libplacebo.
other than mpv, i meant to say.

>>102652252
You can vulkaninfo | grep "extension" to double check but it's probably just that your GPU doesn't support them. Technically, mesa could implement support for those extensions on old hardware, but probably not likely to happen. I'm unsure if any kernel patches would be required too.

>>102651018
No need to fear. Paul builds™ soon.
https://github.com/mpv-player/mpv/pull/14977#issuecomment-2389042774

>>102652148
what does displa-resampe do? I thought it actually used less resources than other sync methods. how is it better than the other video-sync methods? Isn't it the best and most high quality sync method there is?

>>102652318
>in my own fork of mpv
FUCK YEAH
this is the good part. who cares about builds for wintards.

>>102651018
how tf it already has so many more features?

>>102652318
>mpv*someting
why do they always hide that they're using 0.29.1?

You guys get angry at CounterPillow this avih dude sounds like an even bigger asshole. What's his deal?

how much does dithering matter? does error diffusion provide noticeably better results than fruit/ordered?

>>102652404
It makes mpv render every single vsync. The difference is that the video is timed directly to your display's vsync interval and the audio is synced to the video. This makes it use more resources not less.

>Isn't it the best and most high quality sync method there is?
In theory but if you are using the default audio sync mode on a display that supports the exact multiple of the display FPS, it's probably no real difference in practice.

>>102653024
>In theory but if you are using the default audio sync mode on a display that supports the exact multiple of the display FPS, it's probably no real difference in practice.
Not true. You get stutterpv issues with default audio sync mode even if it's an exact integer, this is just how the default audio sync mode works. It doesn't know anything about your display refreshrate, and just spews out frames and the hope/expectation is that it lines up with a vsync, but there are various reasons for why a frame might not line up with a vsync, and it causes a stutter. This stutter becomes even more noticeable and frequent the higher your refreshrate is due to the faster vsync intervals. (every 40~ seconds for 120hz, every 20s for 240.. etc etc.)

>>102653016
Try it, in my experience absolutely not aside from having way higher performance. I think I really had to pixel peep to see some areas that are debatably "better", just different in my eyes really.

>>102653129
yeah I didn't notice much of a difference either. same with deband. as long as you use a high quality source, dithering and debanding shouldn't be a problem.

>>102653111
how does diplay-resample compare to the other diaplay- and display-resample- modes?

>>102653169
yeah mpv is optimized for low quality sources. if thats not an issue for you, try this config

scale=sinc
dscale=lanczos
dither=error-diffusion
error-diffusion=stucki

and on vo=gpu, make sure to use fbo-format=rgba32f

>>102653577
>sinc

New PR from akemi's boyfriend
https://github.com/mpv-player/mpv/pull/14979
I still don't know what akemi sees in him, every one of the chink's prs (including this one) have been flawed

>>102655459
try it instead of being a cumguzzling sheep

>>102655479
how can he debloat shit like that?!
this is unacceptable behavior.
i mean, at least implement some wintardations or something first and build up some credit before trying that stuff!

>>102652318
Paul fork soon, no longer no-mpv!

mpc-hc is better albeithoughever

avih to chink
>You have a track record of good code
Lmao.

>>102653577
I have both low and high quality sources

mine are

scale=ewa_lanczossharp (high-quality profile)
dscale=catmull_rom
default dither
video-sync=display-resample
vo=gpu-next

>>102656814
those are... very bad and blurry

i would switch out scale with lanczos and dscale with lanczos. at the very least.

>>102655479
This looks like a pretty good debloating to me. Assuming it works of course.

>>102656882
forgot the rest
scale-antiring=0.8
dscale-antiring=1.0
tscale=oversample
tscale-antiring=0.8
correct-downscaling=no

are these still blurry? how is sinc better than ewa_lanczossharp? how is lanczos better than catmull_rom for downscaling? lanczos is sharper but downscaling is supposed to be softer, plus catrom with 1.0 antiring is sharp too.

is linear-downscaling good or is it better to turn it off? does it make it video quality better?

>>102657174
>scale-antiring=0.8
>dscale-antiring=1.0
mega blur, antiring adds blur
>tscale-antiring=0.8
bro...
>correct-downscaling=no
nani the fuck??

you need to stop copypasting random shit from the internet dude. dscale antiring doesnt even work with catmull rom... and oversample does not ring so why would you apply antiring to a temporal filter which is already clamped... and correct downscaling is the only way to not have aliasing while downscaling. just nuke your config dude, its not helping you out

>>102657218
it makes it worse

>>102657361
I see, so why is it enabled by default? same with correct-downscaling

>>102657718
Because the mpv devs know what actually looks good, anon.
Would you trust their defaults or some random schizo from here?

>>102657753
BAHAHAHAHA

oh, were you serious?

File: mitchell.png (171 KB, 1000x1000)

171 KB PNG

scale=lanczos
dscale=mitchell

>>102658045
If you are much more knowledgeable, why are we not using your media player?

File: mpc-hc madvr lavfilters.png (445 KB, 640x510)

445 KB PNG

>>102660220
many are

npv (nanahi pantsu viewer) when?

If there was some style control over the way people write here and emotionally react and then people would just come and judge it impartially this would be one of the most intelligent generals on the website but alas... from the style alone it seems like it's made up of insufferable angry autists (which aren't wrong but you wouldn't know).

>>102660902
wow cool meta commentary dude, you must be really smart

File: 1711966734121826.png (2.33 MB, 1262x1420)

2.33 MB PNG

>>102653016
I've toyed a lot of with the raw implementations in avisynth long time ago and they all matter and have different impact, especially for compression, any dithering is good if your source is coming down from a 10+ bit source and you scale it down to w/e, it removes banding. The very complex ones remove it gracefully without any noticeable artifacts, but even the fastest and friendliest to encoding one gives a superior visual experience to just raw banding.

On MPV I always have it disabled because I'm watching 8 bit sources on a 8 bit monitor 99.99% of the times.

The best dither is the monitor chip dither, because it accelerates the best software solutions there are beyond what the software could do, it's much faster and it's free in terms of resources, in terms of end-to-end latency to the pipeline it doesn't even add anything meaningful to the point that even gaming e-sport used monitors have it implemented and enabled (like they might still be 6 bit panels with the chip to dither larger sources down to it).

>>102660945
keep up

>>102653024
>it's probably no real difference in practice.
didn't we test it a month ago and there is, we had various test videos and even playing them back 1:1 refresh rate, source fps people saw stuttering and all that even on files lacking audio only display-resample have perfect paced playback on all tests but there's still mystery if gpu-next has some internal logic issues with it, someone with a very slow camera should compare frame by frame in 15 different tests and give us a definitive answer

>>102660948
>The best dither is the monitor chip dither
the "monitor dithering" is not dithering in the way youre talking about. it's a constantly enabled temporal dither, which is why some monitors are classified as 6+frc or 8+frc instead of true 8bit and true 10bit displays

>>102661011
What are you confused about? Let me give you an analogy you can software decode a codec or can hardware decode it, the logic is the same but the dedicated hardware chip does it faster, it's akin to having way higher FPS in a video game vs. someone else playing the exact same game.

>>102660948
You should never disable dithering as even the conversion to RGB will create high bitdepth data.

>>102660948
>>102660948
>On MPV I always have it disabled because I'm watching 8 bit sources on a 8 bit monitor 99.99% of the times.
by default mpv still dithers 8bit to 8bit. does it really make a difference in that case and is it best to disable? will it improve image quality and performance?

>The best dither is the monitor chip dither, because it accelerates the best software solutions there are beyond what the software could do
so I should just disable dither in mpv? from what I've seen HDR -> SDR with or without dither look the same but without dither performs worse.

>>102661023
...anon are you retarded? yes, monitors have dedicated hardware for temporal frc dithering. but the dithering is applied throughout the entire panel because it requires actual display itself to generate intermediate color values (higher perceived bit-depth) by displaying two different pixel values rapidly, which makes the two colors mix on an LCD display. that's why its called "FRC" (frame rate controlled)

mpv's dithering is entirely static, and dithering in applications is to combat bit-depth mismatch and avoid color banding. your monitor's "dithering" is always active on an FRC display, and is not controlled per application

>>102661078
You're autistic.

>>102661078
Lots of text to say nothing and be schizophrenic.

>>102661086
>>102661096
>open 16-bit content
>open mpv
>"should be fine because my monitor is dithering!!!"
>see color banding
>"wtf!!! but my monitor is dithering!"
this is you

>>102661054
no dither = truncation
I guess if you do anything to the image like altering colors, color correction and applying gamma luts you should probably round the values at least.

>>102661115
>creates a strawman
>not retarded himself
you feed your monitor 8 bit data and it's not dithering it if your panel is 8 bit, you feed it 10 bit data and it dithers everything on the monitor continuously to 8 bit, stop being this much of a sperg and learn to read and think

>>102661160
>you feed your monitor 8 bit data and it's not dithering it if your panel is 8 bit, you feed it 10 bit data and it dithers everything on the monitor continuously to 8 bit
you absolute retard, you cumguzzling faggot. your monitor is not applying any dithering there, your compositor is, which is entirely gpu-accelerated (there are even built-in gpu driver functions for different dither patterns, which are applied only to the application window)
>stop being this much of a sperg and learn to read and think
you're not even thinking, you're just actually retarded. you should read and learn instead, and then come back to this thread once you've moved your IQ up by a fraction of a z-score

>>102661141
>what is an FBO format

>>102661221
Raw 8-bit source can be converted to a higher bit depth and then back to 8-bit deterministically with no impact if no other processing is done in between. Truncation, when done correctly, achieves this.
When you convert from 8-bit to a higher bit depth (e.g., 16-bit), you're essentially adding more bits to represent the same value. These extra bits are usually filled with zeros (or ones, depending on the implementation).
When you truncate back to 8-bit, you simply discard the extra bits. Since these bits were either zeros or a direct extension of the original 8-bit value, discarding them doesn't change the original information.
In conclusion converting to a higher bit depth and then truncating back to the original bit depth without any intermediate processing is a lossless operation.

I've abstained hard from not insulting you, instead I gave you facts and logic, basic math.

>>102661275
What a worthless pseud post.

>>102661057
TL;DR FAQ four you that simplifies.
Do I use color conversion or any processing of the image?
Yes use dithering.
Do I watch movies in their raw format, the player just gives me back what's inside the container?
Disable dithering.

>>102661524
>Do I use color conversion or any processing of the image?
>Yes use dithering
so... use dithering 100% of the time... i dont think mpv can even output raw yuv video without a user shader

>>102661542
Just load an image in paint and compare it with MPV. In ideal conditions and settings MPV is just an ordinary media player, that's a bit faster because of being written in C.

This isn't for ffmpeg anymore? I wanted to know what the best way to encode my anime (bluray) is for quality & space saving.

>>102661880
remux or bust

>>102661896
But then I'd need a 12TB drive just for anime.

>>102661919
you dont already have that?

>>102662075
No, it's 8TB, thus why I want to save space.

>>102662106
buy download more tb, problem solved

or delete all moeshit from your current drive, only keep kino

>>102661919
>>102662106
Meanwhile I do fine with 1 TB for media.
My own personal anecdote conforms with papers on encoding strategies to maximize perceived quality for the least file-size so I do not need anything more than ~7gb per a movie as that roughly gives it the proper bitrate.

>>102662234
7gb for a movie is bitstarved lol, even for anime.
for live action, you're basically watching the movie at a worse quality than most streaming services. may as well just paypig

>>102651600
I hate this pajeet so much you have no idea

>>102662275
That's just like your opinion. Mine is different and mine also overlaps with the empirical findings. You're also exaggerating in order to be the clown of this conversation.
https://streaminglearningcenter.com/codecs/bandwidth-savings-vp9-hevc-av1.html

>>102662275
If you download bit by bit the highest quality streams from say Netflix they are about in that range for movies. Anyone can download them even if they are encrypted, what you see on torrents "WebDL" is those but decrypted.

>>102662106
tips from someone probably much older than you (not anime specific):
* 95%+ of what you already watched, you will never watch again.
* Keeping popular stuff on your disk is useless, especially when you're thinking about re-encoding that content, so you don't care about maximum quality. You will always find that stuff available online if the need ever arises.
* So don't keep anything around unless they are rarities or remuxes. Make common sense exceptions as you go. And always prioritize what you didn't watch yet over what you already watched.
* Bonus: Keep track or a record of everything you watch and everything you delete.

>>102663724
forgot to mention that i'm a recovering hoarder.

since correct-downscaling got a mention, i have a question.
is my observation wrong that disabling it helps performance when absolutely nothing is getting downscaled?
can't mpv auto-set this to no in these cases?
is it because it's a filter setting that gets set before filter chains are created, and recreating the chain twice is for some reason undesirable?
or did i misunderstand what downscaling even means in this context?

Why is it not possible to make jxl screenshots with the first party MSVC build?

>>102663950
>meme OS
>meme build
>meme screenshot format
quintessential nu-mpv user

>>102663950
cumguzzling cattle retard
fucking kys
t. windows user

spoonfeeding OP and its consequences

Will shinchiro builds officially be deprecated now?
https://github.com/mpv-player/mpv/pull/14984

>>102666809
if you like maximally backdoored builds, sure.

>>102666911
You don't have to download mscv schizo

>>102666959
linked PR only relates to MSVC builds not-schizo

>>102666911
If you're using windows, you're already using a backdoored OS anyway so why fret?

>>102667074
yes. "maximally" was the operative word in that sentence.

>>102667074
he's a larper (known schizo who posts on this thread 24/7). he probably uses some meme win10 debloater script and thinks he's not backdoored or something, so using m$ clang compiler seems like a security threat to him. this is a big tell that he's larping and daily drives windows (just like how he claimed that he used 0.29.1, but doesnt actually)

>>102646522
I still haven't figured out how to loop videos in MPV. At this point I'm convinced this basic and much needed feature simply doesn't exist in MPV and the dev refuses to implement it.

t. been using it for 8 years

^
bait

>>102630767
>no graphical equalizer
lol

>>102667470
Bait.

>>102667676
It's a pretty barebones video player, and I guess that's why RAM-lets love it.

>>102667676
why would a media player provide an equalizer? doesn't your audio mixer already have one?

>>102667888
lavfi has EQ's
>>102667280
>he's a larper
k
>known schizo who posts on this thread 24/7
i like where this is going
>he probably uses some meme win10
lmao
haven't touched a m$ system since 2004.
how old were you back then?
negative numbers allowed.
>debloater script
what the fuck is that lol
>and thinks he's not backdoored
on the fence with that one.
one can't prove a negative without indulging in some hubris.
>so using m$ clang compiler seems like a security threat to him.
"warning users about being extra backdoored is harmful, and the only explanation is deflection."
we are reaching average jeet intelligence levels with this one.
>this is a big tell that he's larping and daily drives windows
>big tell
lmao
> claimed that he used 0.29.1, but doesnt actually
not saluting the increasingly winfested nu-mpv with its aptly described kindergarchy doesn't mean i'm not following it closely on the code level.
and i build my own packages thank you very much.

% mpv --version
mpv v0.39.0-43-gXXXXXXXXXX-dirty Copyright © 2000-2024 mpv/MPlayer/mplayer2 projects
 built on Oct  2 2024 14:50:55
libplacebo version: v7.349.0 (v7.349.0-15-gXXXXXXXX-dirty)
FFmpeg version: N-117296-gXXXXXXXXXX
FFmpeg library versions:
   libavcodec      61.21.100
   libavdevice     61.4.100
   libavfilter     10.6.100
   libavformat     61.9.100
   libavutil       59.41.100
   libswresample   5.4.100
   libswscale      8.4.100
% uptime
 17:04:12 up 85 days,  7:04,  7 users,  load average: 0.94, 0.87, 1.05
% uname -a
Linux arch64 6.6.38-1-lts #1 SMP PREEMPT_DYNAMIC Tue, 09 Jul 2024 11:08:50 +0000 x86_64 GNU/Linux

commit hashes scrubbed because i have my own patches applied to all three projects (other than the non-committed "-dirty" part).
here goes your "windows using, mpv 0.29.1 baiting, larping schizo" theory.
who is the real schizo who is sitting here 24/7 to the point of confusing multiple anons as the same person now?
>[incoming] you took the bait lol
yes. i knowingly did

>>102666809
the brapschizo will love this attack vector
>>102668659
your uptime tells me you have big testicles and automatically won

>>102668659
>>102668926
samefagging is against the rules

what should my mpv config be? i don't want any borders. i mostly watch pirated movies while getting black out drunk to cope with my alcoholism

>>102668975
--no-border

and you probably need adaptive-sharpen.glsl shader to combat the blurry vision from alcoholism

is apple sabotaging intel hardware? or mpv/ffmpeg nudevs have broken something again?
>can't play webms vp9 with videotoolbox anymore mpv just freezes without any errors
>can't play h264/h265 4k@60fps with videotoolbox anymore mpv drops every second frame

also s o y dev is now spam lol

>>102669395
for the second one, make sure your display is not running in a 59.94Hz mode or something

pedotroon thread

https://github.com/mpv-player/mpv/pull/14984#issuecomment-2392034631

BASED BRAPFAN5!!

>>102669550
i also have an m3 macbook pro, and it doesn't have any of these issues

>>102669395
did your device enter the planned obsolescence period?
that reminds me of this song:
https://www.youtube.com/watch?v=7sVc8lkTfrk
religion was declared obsolete because we have fortran now. good thing no one did that with rust.

>>102663724
I doubt you're older than me, and I'm just dumb. But it makes sense to archive only remuxes, that's true. I'll just have to invest in more space.

>>102663724
>tips from someone probably much older than you (not anime specific):
I don't know what is the average age on this board to judge that. Also old != smart :)
>* 95%+ of what you already watched, you will never watch again.
True, I learned this over the years. I would keep my favorite content, but overtime I may go back to only a few. And most of the time I could just re-download. Though it is sometimes nice to find this random movie in your library and just watch it (again) without thinking. Otherwise you might never remember it or want to watch it.
>* Keeping popular stuff on your disk is useless, especially when you're thinking about re-encoding that content, so you don't care about maximum quality. You will always find that stuff available online if the need ever arises.
True, popular stuff is accessible, so unless you have abundant of space, there is no point in hoarding that.
>* So don't keep anything around unless they are rarities or remuxes. Make common sense exceptions as you go. And always prioritize what you didn't watch yet over what you already watched.
Of course solution is always get more space. There is no good rule for removing stuff, common sense, but still.
>* Bonus: Keep track or a record of everything you watch and everything you delete.
Bonus: Add good indexer to your library, you would be impressed what you can find sometimes hidden deep.

>>102673109
>Bonus: Add good indexer to your library, you would be impressed what you can find sometimes hidden deep.
my scripts are all i need. i replace removed files with empty* files containing limited "metadata" about them in their name. and anything beyond `locate` for indexing is bloat.

* actually equally-sized sparse ones.

How can I yt-dlp and get a specific sub embedded into my file? --sub-lang ja/'ja' doesn't give me japanese subs, I have to do --all-subs which downloads a bunch of crap sub files and then I have to delete them and manually add the ja one to my file. Am I doing it wrong or is my yt-dlp broken?

>>102667676
Buy a good audio interface you broke ass nigger

>>102630767
Question, I noticed that when you play a set of files on mpv, the program skips some based on their filename.
I'm not sure if it's a problem with the program itself or the autoload.lua script though.
Do you guys know if there's a known fix somewhere?

To test it try grabbing 3 .mp4 video files and name them: 000D364-720P, 00A0cd982643b23e6275bcddd2943bb6 and 1.
When 000D364-720P is done playing and/or when you click the arrow that takes you to the next video, the program will jump to 1.mp4 and skip 00A0cd982643b23e6275bcddd2943bb6.

>>102630767
>>102674204
Forgot to add, I use Shinchiro's build, does that matter?
https://sourceforge.net/projects/mpv-player-windows/files/64bit/mpv-x86_64-20240929-git-c3d9243.7z/download

>>102674204
>>102674223
Oh, I think the problem lies in autoload.lua?
It must be related to how it generated a playlist.
Turns out mpv with autoload is putting 00A0cd982643b23e6275bcddd2943bb6.mp4 BEFORE 000D364-720P.mp4 in its own playlist.
I just realized the arrow that normally takes you to the previous video wasn't grayed out so I checked it, and it effectively takes you from 000D364-720P to 00A0cd982643b23e6275bcddd2943bb6 for whatever reason, even though on Windows 000D364-720P comes first and 00A0cd982643b23e6275bcddd2943bb6 comes second.
Is this a UNIX vs Windows thing maybe?

File: file.png (160 KB, 1001x450)

160 KB PNG

>>102674650
Alphanumsort is pretty arbitrary, of course they'll will be differences between implementations unless you go out of your way to make them compatible. I never even noticed that there was a non ascii sort order in all these programs desu

>>102674650
>Is this a UNIX vs Windows thing maybe?
That's strange. Sounds like a bug to me. Sorting should work the same way.

>>102631385
mpv-chan's so lucky...

>>102630767
I look like this

>>102676991
No, you don't.

>>102677418
he meant on VTube

is there an audio transcription plugin? e.g. write subtitle for the media as it is playing?

>>102678550
yes but I could never get it to work well
https://github.com/GhostNaN/whisper-subs

>>102678564
>https://github.com/GhostNaN/whisper-subs
thank you senpai.
I will study the codes and see if it works on my machine.

>>102678564
hm it cannot do real-time transcription on my machine.
But, it seems to work out of the box for mp4s.

>>102630767
I don't look like this and post this.

https://github.com/mpv-player/mpv/issues/8705#issuecomment-2394443543
new meme dropped

File: file.png (50 KB, 954x286)

50 KB PNG

What did he mean by this?

https://github.com/mpv-player/mpv/pull/14995

new bloat and enshittifaction (loss of two free keys)

>This should help new users to view mpv's keybindings as it's easier to discover than shift+i and 4.

and how exactly are they supposed to discover this? by slamming their head into the keyboard? it takes the same amount of effort (reading the manual) to actually figure out where this page is

>>102686891
>how exactly are they supposed to discover this?
well ? is a fairly intuitive mapping used by lots of other programs. I've already had it remapped to show binds for ages now. F1 would also have been a good option.

>>102686891
>how exactly are they supposed to discover this?
The same way how the discover that space pauses, now ask yourself how would they know?

>>102686988
VLC has no keybind for ?

>>102687099
by pressing their keys randomly? absolutely circular logic... totally worth getting rid of the h key for this

>>102687144
You must have never used any terminal interface or any interface that is controlled by keybinds. Even web players use ?. But you know better
>totally worth getting rid of the h key
What do you mean? You can still rebind it...

>>102686891
>>102686988
agreed.
i actually already have 'h' bound to something else, and that will not break for me or anyone else.

and the '?' is also good since probably no one is using that, except for maybe the exact same purpose.

"loss" of two "free keys" is probably the most retarded thing i read today.
like if you're savvy enough to know about input.conf, then you should you even care.
or are you worried about mythical users who are somehow capable enough to know how to use `mp.add_key_binding`, but somehow retarded enough to not know when when they should be using `mp.add_forced_key_binding`?
(=) Anon approved these change

>>102687550
literally only youtube does this. why are you lying?

>>102630975
I bind the shaders to a keybind cause you dont always need it (for non-anime)

what would be the best way to batch download videos

will gibplacebo ever support vdpau/vaapi?

File: 9958f0ca0.png (263 KB, 1134x922)

263 KB PNG

mpv --show-profile=high-quality
Profile high-quality:
 scale=ewa_lanczossharp
 dscale=mitchell
 scale-antiring=0.6
 hdr-peak-percentile=99.995
 hdr-contrast-recovery=0.30

what should I use for video-sync, tone-mapping and tone-mapping-param? should I enable deband?

>>102688962
you can do whatever you want my man, the world is your oyster

>>102688962
defaults

>>102689010
>--video-sync=audio
lol

>>102689087
>too poor for vrr
many such cases..

>>102630767
SEX

>>102689099
>vrr
cool meme

>>102689208
it's only a meme when you don't have access to it. ever heard of sour grapes?

>>102689222
>ever heard of sour grapes?
unfortunately no.
maybe because i wasn't a Juicero customer like you.
describe it to me.

>>102689129
Anon that's a media player...

>anoooon ANOOOOOOOOOOOOOOOOOOOOOOOOONNNNNNNNNN ERP sex saar sexo smooch sexi trannypozzparty Anon,Anon,Anon,Anon,Onion, anooooooooonnnnnnnn........ ..........

>>102645978
touch it?

>>102691326
meds now

Name one thing that you would change or fix in mpv.

>>102692074
Add a big dick futanari doing the helicopter on every startup

>>102692074

--screen

so it's linked to the actual monitor instead of Windows' retarded number scheme that sometimes just rejiggers itself.

>>102692225
complain about it in every thread and kasper will eventually fix it

File: __megumin_kono_subarashii(...).jpg (121 KB, 1468x978)

121 KB JPG

>>102692074
changing speed in range of 1...2 wouldn't drop frames

>>102694041
get a better pc

>>102694122
no imo it's some issue with audio filters, where buffers get dropped and equivalent amount of video frames get dropped

>>102694319
if you already deduced the problem then go send a pr

>>102694330
it's observation, idk how to fix

>>102694343
ok so you're literally just hallucinating and making shit up, cool

>>102694362
no but it's impossible to fix

>>102694362
find me better reason than audio buffers getting flushed from speed change and video sync code not giving a fuck and moving the same amount that was flushed

>>102694569
find me a reason to care

if it was a real problem and you were saying this with complete conviction, then you would know how to fix it. if you dont know how to fix it, then its not a real problem and you're flailing at the void

It's a real problem. Fixing it is probably not really possible other than allowing video and audio to desync for a while before catching up.

>>102694916
find me a reason to care

>>102694952
betterment of the mpv