mpv ffmpeg yt-dlp

mpv ffmpeg yt-dlp

Beware Cat Shirt $21.68

Rise, Grind, Banana Find Shirt $21.68

Beware Cat Shirt $21.68

  1. 3 weeks ago
    Anonymous

    1fps + display-resample how?

    • 3 weeks ago
      Anonymous

      2fps works btw

    • 3 weeks ago
      Anonymous

      here's a sample homosexuals
      https://files.catbox.moe/lze3ls.mov

      • 3 weeks ago
        Anonymous

        unironically encode that to 23.976 fps using frame doubling

        • 3 weeks ago
          Anonymous

          i want it to just work

          • 3 weeks ago
            Anonymous

            in mpv config try
            vf=fps=24

          • 3 weeks ago
            Anonymous

            [...]
            or unironically replace 24 with your screen's native fps

            isn't it a bug though 1fps @ min 2.0 speed works, 2fps+ also works

          • 3 weeks ago
            Anonymous

            Can you restate exactly what the problem is? Not sure which posts are yours in the previous thread.

          • 3 weeks ago
            Anonymous

            display-resample doesn't work on 1fps

          • 3 weeks ago
            Anonymous

            >display-resample
            cargo cult. don't use it.

          • 3 weeks ago
            Anonymous

            one of best features of mpv tbh senpai

          • 3 weeks ago
            Anonymous

            it's placebo

          • 3 weeks ago
            Anonymous

            not for me, vsyncs pretty much perfectly, without it video lags behind while resizing

          • 3 weeks ago
            Anonymous

            there's something wrong with your system

          • 3 weeks ago
            Anonymous

            >vsyncs pretty much perfectly
            what are you talking about? you can't possibly be seeing screen tearing on 30 fps videos.

          • 3 weeks ago
            Anonymous

            it's not really about screen tearing, audio mode is late in resizing video

          • 3 weeks ago
            Anonymous

            have you tried not resizing your video so often that this became a problem? Why don't you pause the video, resize, then resume. And do this only once because how can you focus on anything while you're constantly resizing?

          • 3 weeks ago
            Anonymous

            no, i enjoy resizing windows so i am display-resample enjoyer

          • 3 weeks ago
            Anonymous

            cheers anon. can't help you there but enjoy your resizing experience.

          • 3 weeks ago
            Anonymous

            >30 year old resoooooozer

          • 3 weeks ago
            Anonymous

            linux moment

          • 3 weeks ago
            Anonymous

            decent workaround so far

          • 3 weeks ago
            Anonymous

            [...]
            or unironically replace 24 with your screen's native fps

            hm even with 24fps video and vf=fps=60 it seems like an improvement, what's the downside of this?

          • 3 weeks ago
            Anonymous

            probably none. it's an ffmpeg filter. if you were re-encoding it you'd use the same exact parameter.

          • 3 weeks ago
            Anonymous

            in mpv config try
            vf=fps=24

            or unironically replace 24 with your screen's native fps

  2. 3 weeks ago
    Anonymous

    damn, after 3 good threads we're back to this

  3. 3 weeks ago
    Anonymous

    Finally. Actual sovl.

  4. 3 weeks ago
    Anonymous

    haasn did nothing wrong

  5. 3 weeks ago
    Anonymous

    the duality of man

  6. 3 weeks ago
    Anonymous

    diaper thread

    • 3 weeks ago
      Anonymous

      living all day rent free in your head

  7. 3 weeks ago
    Anonymous

    Hourly reminder that the mighty huhn remains undefeated.

    • 3 weeks ago
      Anonymous

      huhn blundered massively with bt.601

      • 3 weeks ago
        Anonymous

        standards mind broke huhn

        no, hunh is 100% right.

    • 3 weeks ago
      Anonymous

      huhn blundered massively with bt.601

      explain, I need to know who won in that issue

      • 3 weeks ago
        Anonymous

        standards mind broke huhn

        • 3 weeks ago
          Anonymous

          [...]
          no, hunh is 100% right.

          >The transfer characteristics for the BT.601 spec are the same transfer characteristics as BT.709. Anything else is either historic or outside of spec. If they share the same transfer characteristics, it only makes sense to decode both of them with BT.1886.

          >which is wrong because they don't have a EOTF they have the same OETF which is completely irrelevant. bt 709 got an official EOTF with bt 1886. befoer that it was the wild west and later calibrator defined 2.4 for a Batcave.

          so which of these is correct??

          • 3 weeks ago
            Anonymous

            he's correct in that retroactively applying the newly defined EOTF is nonsense on principle. However because before bt1886 there wasn't a single standardized transfer, you can't really know what EOTF "should" be used. Moreover bt1886 tries to "emulate" roughly some CRT transfer (at infinite contrast it gives you 2.4 gamma). So if you are indeed on such a display then there's actually no difference between bt1886 and assuming 2.4 gamma.

  8. 3 weeks ago
    Anonymous

    how do i make mpv look like VLC? the boys are coming over for movie night and i dont want them to make fun of me

    • 3 weeks ago
      Anonymous

      scale=nearest

    • 3 weeks ago
      Anonymous

      scale=nearest

      I thought VLC used catmull rom

      • 3 weeks ago
        Anonymous

        vlc uses (b=0.0,c=0.6)

        it doesnt have cubic convergence unlike catrom because b+2c=1.2

  9. 3 weeks ago
    Anonymous

    i liked the anime girl from the last general

    • 3 weeks ago
      Anonymous

      yeah i really love mpv-san too

  10. 3 weeks ago
    Anonymous

    sfan was the ffmpeg4 + mpv 0.29 user all along?

    https://github.com/mpv-player/mpv/issues/13454

    • 3 weeks ago
      Anonymous

      he always was, you didn't realize? you can tell from his posting style

      he also revealed in a thread that he uses vo=gpu for content that needs fast init times and vo=gpu-next only sometimes. he basically hates libplacebo but only mentions it ITT because its anonymous obviously

      • 3 weeks ago
        Anonymous

        Schizo theory.

    • 3 weeks ago
      Anonymous

      0.29 user is a macgay
      t. lore grandmeister

    • 3 weeks ago
      Anonymous

      he always was, you didn't realize? you can tell from his posting style

      he also revealed in a thread that he uses vo=gpu for content that needs fast init times and vo=gpu-next only sometimes. he basically hates libplacebo but only mentions it ITT because its anonymous obviously

      https://github.com/mpv-player/mpv/issues/13451#issuecomment-1936970510
      deboonked

      • 3 weeks ago
        Anonymous

        no point in using anything between 0.29.1 and master

  11. 3 weeks ago
    Anonymous

    [...]

    You're saying sizumam is still around? interesting

    • 3 weeks ago
      Anonymous

      zewia lurks IQfy of course he's still around

  12. 3 weeks ago
    Anonymous

    Anyone know a way to confirm if VRR is working on an AMD laptop with freesync? I'm using the same config that I know triggers VRR on my desktop monitor (as confirmed by the OSD refresh rate display) but no such thing exists on a laptop. Do I just have to do it by eye (i.e. look for judder) or is there some hint in the mpv stats overlay that will give it away, or maybe some external tool?

    • 3 weeks ago
      Anonymous

      move your mouse around

      also VRR is a confirmed meme. just use video-sync=audio and dont worry about it

      • 3 weeks ago
        Anonymous

        >move your mouse around
        jfc... why didn't I think of that.
        >just use video-sync=audio
        But that's exactly what I use on my desktop to trigger VRR.
        >VRR is a confirmed meme
        Avoiding both judder and smoothmotion blur/smearing is a meme? If you say so.

        • 3 weeks ago
          Anonymous

          >Avoiding both judder and smoothmotion blur/smearing is a meme? If you say so.
          its a meme because the high jitter with video content causes uneven framepacing compared to regular vsync, this is old news already. the VRR meme is dead and should no longer be recommended.

          basically, madshi won

          >But that's exactly what I use on my desktop to trigger VRR.
          im obviously talking about without VRR, aka the default

          >Avoiding both judder and smoothmotion blur/smearing is a meme? If you say so.
          there are ways of doing this that dont involve VRR or resampling
          also the problem with oversample isnt the fact that it "smears", it doesn't "smear". the problem with oversample right now is that its just fricking broken.

          • 3 weeks ago
            Anonymous

            >basically, madshi won
            Why that?

          • 3 weeks ago
            Anonymous

            i already mentioned it in my post
            https://github.com/mpv-player/mpv/issues/12005#issuecomment-1902597586

          • 3 weeks ago
            Anonymous

            He says that the jitter doesn't matter.

          • 3 weeks ago
            Anonymous

            post a quote, because thats not what he says.

          • 3 weeks ago
            Anonymous

            The VRR has random outlier values which is the problem and probably confirms that dokuro wasn't hallucinating.

          • 3 weeks ago
            Anonymous

            >probably confirms that dokuro wasn't hallucinating.
            how do I acquire dokuro's eyesight to become a true pixelpeeper?

          • 3 weeks ago
            Anonymous

            >there are ways of doing this that dont involve VRR or resampling
            May I see them?

          • 3 weeks ago
            Anonymous

            no

      • 3 weeks ago
        Anonymous

        >move your mouse around
        jfc... why didn't I think of that.
        >just use video-sync=audio
        But that's exactly what I use on my desktop to trigger VRR.
        >VRR is a confirmed meme
        Avoiding both judder and smoothmotion blur/smearing is a meme? If you say so.

        Oof, still 60fps mouse movement. Now that I think about it VRR wouldn't trigger on this laptop from 24fps video content because it has a VRR range of 40-60hz with no LFC. Is there some way to get mpv to send the same frame twice in a row to trigger 48hz VRR?

        • 3 weeks ago
          Anonymous

          >Is there some way to get mpv to send the same frame twice in a row to trigger 48hz VRR?
          --vf=fps=48/1.001

        • 3 weeks ago
          Anonymous

          https://github.com/mpv-player/mpv/issues/12005#issuecomment-1647766614

    • 3 weeks ago
      Anonymous

      go here:
      https://www.rtings.com/tv/tests/motion/24p

      download the test video and follow the instructions. you'll need a smartphone to do it (or an actual digital camera, who the frick owns one of them anymore, lol)

    • 3 weeks ago
      Anonymous

      move your mouse around

      also VRR is a confirmed meme. just use video-sync=audio and dont worry about it

      >move your mouse around
      jfc... why didn't I think of that.
      >just use video-sync=audio
      But that's exactly what I use on my desktop to trigger VRR.
      >VRR is a confirmed meme
      Avoiding both judder and smoothmotion blur/smearing is a meme? If you say so.

      Well even with video that falls in the VRR range I still can't get it to trigger (tested with a 41fps video and mouse was still moving at 60fps). Is it garbage AMD drivers or something? Don't have this issue on my nvidia desktop.

      • 3 weeks ago
        Anonymous

        your vrr range is probably 48-60. in which case your attempts are futile to begin with

        • 3 weeks ago
          Anonymous

          >range is probably 48-60
          i know this feeling. The good thing is that you can just make modelines and it will work but utterly useless otherwise.

          • 3 weeks ago
            Anonymous

            >modelines
            wtf are those? are they a linux thing?

          • 3 weeks ago
            Anonymous

            no, you can set up modelines on windows as well. nvidia and amd have custom resolution tools where you can specify a pixel clock. or alternatively you can use CRU

          • 3 weeks ago
            Anonymous

            oh
            yea, ive used CRU

        • 3 weeks ago
          Anonymous

          No it's 40-60 as reported in the driver panel.

          • 3 weeks ago
            Anonymous

            i dunno turn off hwdec see what happens

          • 3 weeks ago
            Anonymous

            Still nothing. Seems like a futile effort unfortunately

    • 3 weeks ago
      Anonymous

      nope. without an overlay from the display its literally impossible to tell

  13. 3 weeks ago
    Anonymous

    why are VRRtroons so lazy? just set up a modeline and you wont have to worry about any of this shit

  14. 3 weeks ago
    Anonymous

    if you're this ocd about vrr then just buy a high frame rate screen. on a 144 hz screen you'd see <7ms of jitter max.

    • 3 weeks ago
      Anonymous

      thats not how that works, the "7ms" frametiming number comes from the fact that the screen is running at 144hz. obviously if when you're using VRR, your screen is either going to drop to a true 24hz value (in which case it will be 40ms+), and if it's using LFC to go to something like 72hz, you'll have ~14ms frametimings. considering the fact that. 60hz is 16ms and plenty of people can visibly see stuff like 3:2 frame judder, its not unrealistic to assume that you can see a repeat frame that sticks on screen for ~14ms, or you can see uneven framepacing caused by jitter at the 14ms mark.

      BUT what you're saying is true if you're not using VRR, and you just use a typical vsync. at even more absurd refreshrates (240hz, 360hz) repeat frames may as well be nonexistent even when they do occur. i would argue that its not very visible at 144hz either but some schizos would obviously argue otherwise

      • 3 weeks ago
        Anonymous

        my monitor is precisely 143.856 hz which divides perfectly even with 23.976 fps. This means I can use 23.976 fps content as a baseline in case I notice judder with any other framerate. I have never seen judder with 30 fps content.

        • 3 weeks ago
          Anonymous

          cool but i dont see what that has to do with my post, this is the solution im advocating for anyways, not VRR.

    • 3 weeks ago
      Anonymous

      We were talking about laptops lil bub.

      • 3 weeks ago
        Anonymous

        then you are watching on hardware that categorically will not deliver the optimal viewing experience. just have fun watching the content, it doesn't matter what config you run on your chinkpad.

  15. 3 weeks ago
    Anonymous

    Can anyone explain why vulkan hwdec is so slow

    • 3 weeks ago
      Anonymous

      softwarechuds are powerful hardwarecels are weak

    • 3 weeks ago
      Anonymous

      let me guess, RDNA2 and windows?

      • 3 weeks ago
        Anonymous

        Yes

        • 3 weeks ago
          Anonymous

          known issue. vulkan hwdec has insane overhead compared to d3d11va for some godawful reason. kasper claims it doesn't happen on RDNA1 so the issue is mostly with RDNA2.
          it will probably never be fixed because lol who uses vulkan hwdec on windows unironically (mpv is literally the only software i know of that supports vulkan hwdec).

          AMD isn't going to give a flying shit about this until something like Chrome starts supporting vulkan hwdec

          • 3 weeks ago
            Anonymous

            Thanks Dokur0

  16. 3 weeks ago
    Anonymous

    dokuro? More like DORKuro!

  17. 3 weeks ago
    Anonymous
  18. 3 weeks ago
    Anonymous

    [...]

    didn't you accuse the braphog poster of being sfan the other day? total schizo

    • 3 weeks ago
      Anonymous

      >t. thread raider from the sfan telegram chat

      • 3 weeks ago
        Anonymous

        i'm just going to assume you're trolling

        • 3 weeks ago
          Anonymous

          you got exposed, now frick off and stop derailing the threads with your fat porn

          • 3 weeks ago
            Anonymous

            back up what you're saying or frick off yourself, schizo moron

  19. 3 weeks ago
    Anonymous

    >yt dlp
    Made for computer illiterate zoomers. I laugh every time I see a dumb ytdlp weeb struggle with hls encryption or widevine content because he can’t code

    • 3 weeks ago
      Anonymous

      wtf is a widevine

  20. 3 weeks ago
    Anonymous

    Seeking maxes out my gpu to the point of stuttering and locking up mpv

    Why this never happened before?

    • 3 weeks ago
      Anonymous

      0.29.1 status?

      • 3 weeks ago
        Anonymous

        ????
        mpv is 0.37 stable

        • 3 weeks ago
          Anonymous

          does it work on 0.29.1 THOUGH

          • 3 weeks ago
            Anonymous

            homie i aint got a copy that old laying around

          • 3 weeks ago
            Anonymous

            compile it and let us know
            >homie
            calm down with the language

    • 3 weeks ago
      Anonymous

      profile=fast

  21. 3 weeks ago
    Anonymous

    Is there a filter that can remove/reduce chromatic aberration?

    • 3 weeks ago
      Anonymous

      post example?

      • 3 weeks ago
        Anonymous
        • 3 weeks ago
          Anonymous

          i was thinking of frieren.
          i have not watched it but that filter is kino.
          regardless, try searching the ffmpeg filter list. they're applied via vf=
          https://ffmpeg.org/ffmpeg-filters.html

          • 3 weeks ago
            Anonymous

            I actually like it. It goes overboard in some scenes though and can be distracting in motion.

          • 3 weeks ago
            Anonymous

            modern anime is so trash

          • 3 weeks ago
            Anonymous

            >spend countless hours drawing art
            >smear with meme filters
            I don't understand this practice

          • 3 weeks ago
            Anonymous

            >spend countless hours setting up mpv.conf
            >smear it with meme shaders
            I don't understand this practice

          • 3 weeks ago
            Anonymous

            very true

          • 3 weeks ago
            Anonymous

            >spend countless hours drawing art
            >smear with meme filters
            I don't understand this practice

            it's a callback to a phenomenon that used to be more prevalent, caused by anything from poor lens focus to flawed film telecine to bad encoding (chroma haloing). The old CCS BDs had this effect, and while it is technically a "flaw", I prefer the look of the old BD as an overall package over the newer 4k master. There are many old animu with a bit of rainbowing that don't have a better release than that in existence, and that effect becomes part of how it's remembered. it's probably calling back to that sort of thing.

  22. 3 weeks ago
    Anonymous

    what is that? I just have novideo, can't see what written there

  23. 3 weeks ago
    Anonymous

    0.37 was still half baked 0.38 is when we're back to being based

    • 3 weeks ago
      Anonymous

      here's what 0.38.0 needs so far

      >(x) cache fixed
      >( ) oversample fixed
      >( ) subtitle colorspace fixed
      >( ) a/v sync improved so no more Dropped: 1
      >( ) no more (slow!)
      >( ) polar downscaling unnecessary subpixel offset fixed
      >( ) gpu-next png gamma fixed so that its finally huhnpilled
      >( ) deband that no longer adds banding
      >( ) deband removed from high-quality
      >( ) catromschizo PR to make catmull_rom the default downscaler with high-quality
      >( ) orthogonal antiringing fixed for downscaling

      • 3 weeks ago
        Anonymous

        >>(x) no more (slow!)

        • 3 weeks ago
          Anonymous

          (slow!) still happens. just load an absurdly large shader like nnedi3 nns256 and you'll see that the compilation times are still identical.

          • 3 weeks ago
            Anonymous

            Kasper only make the built-in shaders faster by not unrolling the loops. It had nothing to do with user shaders.

          • 3 weeks ago
            Anonymous

            *built-in ORTHOGONAL shaders

            polar shaders still take the same amount of time to compile. my point is we can still go faster

      • 3 weeks ago
        Anonymous

        https://github.com/mpv-player/mpv/wiki/Stuff-to-do

      • 3 weeks ago
        Anonymous

        Good fix list. The last and second item especially.

    • 3 weeks ago
      Anonymous

      0.29 is not only perfectly baked but also well seasoned

  24. 3 weeks ago
    Anonymous

    screenshot-tag-colorspace=no

    any downsides?

    • 3 weeks ago
      Anonymous

      wrong gamma. but i guess there's not much you can do if you're a firefox cuck

  25. 3 weeks ago
    Anonymous

    >We need help with converting mpv from C to a more secure and modern language with a CoC. This will probably, and unfortunately, be Electron based (For the UI, and plug-ins have already moved to Javascript with great success), and the C(nile) core rewritten in Rust. (Other language suggestions, if appropriate, are welcome. This could for example be something like Ada.) This would probably go as follows:

  26. 3 weeks ago
    Anonymous

    vgh, upscaling perfect

    • 3 weeks ago
      Anonymous

      VGH...

      • 3 weeks ago
        Anonymous

        diaper status?

      • 3 weeks ago
        Anonymous

        VGH?

  27. 3 weeks ago
    Anonymous

    pick one IQfy

    • 3 weeks ago
      Anonymous
      • 3 weeks ago
        Anonymous
        • 3 weeks ago
          Anonymous

          choose

          • 3 weeks ago
            Anonymous

            which scalers are these? wanna see if i can match all 4.
            i'm gonna guess up front that #3 is bilinear

          • 3 weeks ago
            Anonymous

            >which scalers are these? wanna see if i can match all 4.
            that's the point of the game
            >i'm gonna guess up front that #3 is bilinear
            nope, bilinear looks like this

          • 3 weeks ago
            Anonymous

            is #3 catmul_rom?
            nobody is good enough to match them in a blind test to all the known scalers of the universe.
            #3 looks to me to be a bit more blocky in the way it samples the pixels.
            The other 3 perform the same with varying degrees of sharpness, probably doesn't matter which one you use.

          • 3 weeks ago
            Anonymous

            maybeh

            [...]
            [...]
            [...]
            4 & 1 > rest
            4 has a slight edge over 1 in sharpness but also has a lot more haloing

            interesting interpretation, ill wait a couple more minutes

          • 3 weeks ago
            Anonymous

            [...]
            [...]
            >1
            scale=lanczos

            >2
            scale=spline36

            >3
            scale=catmull_rom

            nice job guys
            >4
            scale=lanczos
            scale-window=hanning

            this is why i said [...] your interpretation is interesting because the hanning window makes sinc less sharp, with less ringing, and more aliasing (just look at the metal sheen on the bumper, the difference in aliasing between all of them is pretty obvious)

            >this is why i said

            [...]


            [...]
            [...]
            4 & 1 > rest
            4 has a slight edge over 1 in sharpness but also has a lot more haloing(You)# your interpretation is interesting
            i rather quickly compared them in firefox going off the sky and pole but the zoom in on them were slightly off despite saying the same, disregard. use slow.pics next time

          • 3 weeks ago
            Anonymous

            [...]
            [...]
            [...]
            4 & 1 > rest
            4 has a slight edge over 1 in sharpness but also has a lot more haloing

            [...]
            [...]
            [...]
            I have no idea what the frick these are but I'm going to go 1 > 4 > 2 > 3

            >1
            scale=lanczos

            >2
            scale=spline36

            >3
            scale=catmull_rom

            nice job guys
            >4
            scale=lanczos
            scale-window=hanning

            this is why i said

            [...]
            [...]
            [...]
            4 & 1 > rest
            4 has a slight edge over 1 in sharpness but also has a lot more haloing

            your interpretation is interesting because the hanning window makes sinc less sharp, with less ringing, and more aliasing (just look at the metal sheen on the bumper, the difference in aliasing between all of them is pretty obvious)

          • 3 weeks ago
            Anonymous

            cool. that's why i wanted just the names of the scalers because I bet I could have matched them up.
            interesting how scale-window=hanning basically turns lanczos into spline36.

          • 3 weeks ago
            Anonymous

            >scale=catmull_rom
            >nice job guys
            dscale still good

    • 3 weeks ago
      Anonymous

      choose

      4 & 1 > rest
      4 has a slight edge over 1 in sharpness but also has a lot more haloing

    • 3 weeks ago
      Anonymous

      choose

      I have no idea what the frick these are but I'm going to go 1 > 4 > 2 > 3

    • 3 weeks ago
      Anonymous

      choose

      [...]
      [...]
      >1
      scale=lanczos

      >2
      scale=spline36

      >3
      scale=catmull_rom

      nice job guys
      >4
      scale=lanczos
      scale-window=hanning

      this is why i said [...] your interpretation is interesting because the hanning window makes sinc less sharp, with less ringing, and more aliasing (just look at the metal sheen on the bumper, the difference in aliasing between all of them is pretty obvious)

      third opinion here, agree with this

      cool. that's why i wanted just the names of the scalers because I bet I could have matched them up.
      interesting how scale-window=hanning basically turns lanczos into spline36.

      1 > 2 = 4 >>> 3
      spline and hanning window trade places depending on where on the picture. really pixel peeping with over 3x zoom. but i have to admit that lanczos is the overall slightly better performer
      the moral of the story is that catmul got btfod so hard it's not even funny, thought it was fricking bilinear

  28. 3 weeks ago
    Anonymous

    jesus christ spline36 looks so blurry, aliasy AND ringy. why the frick were we using this unironically for years

    • 3 weeks ago
      Anonymous

      wait until you hear high-quality uses ewa_lanczossharp which is ever blurrier.

      • 3 weeks ago
        Anonymous

        substantially less ringing and basically no aliasing though, hence why it's "high quality", it's also the sharpest polar filter that has all of these characteristics, any sharper and you start weighing the balances in terms of more aliasing or more ringing

        in terms of metrics they usually score pretty close together, at least back when spline36 was still in the meme metrics...

        realistically your only sane choices are catmull_rom, lanczos and ewa_lanczossharp for upscaling. they all offer something the other doesn't. there are some other meme selections like hanning-sinc and ewa_hanning which perform nicely for their own niches, but i cant think of any reason why i would daily drive spline36 tbh.

        He wins yet again

        performs well in subjective tests AND objective tests... i kneel profusely

        >scale=catmull_rom
        >nice job guys
        dscale still good

        ye

        • 3 weeks ago
          Anonymous

          https://slow.pics/c/G7BRRTVO
          ewa_lanczossharp rings more than spline36 and is blurrier.

          • 3 weeks ago
            Anonymous

            this is a pretty crap test tbh, ringing is mostly measured by intensity and amount of halos. you can have something that only has one halo but rings A LOT (polar sharp cubic for example), in which case the ringing is very noticeable despite having a smaller halo.

            ewa_lanczossharp has two halos just like spline36, but ewa_lanczossharp has less intense ringing because it's polar and exhibits better blending since it's less aliasy. orthogonal filters with two halos looks harsher, hence why people always meme about lanczos looking "deep fried" and ewa_lanczos looking "more natural" or looking "more analog". this is typically preferable for most people, but ewa_lanczossharp is slow obviously

            [...]
            [...]
            [...]
            [...]
            third opinion here, agree with this [...]
            1 > 2 = 4 >>> 3
            spline and hanning window trade places depending on where on the picture. really pixel peeping with over 3x zoom. but i have to admit that lanczos is the overall slightly better performer
            the moral of the story is that catmul got btfod so hard it's not even funny, thought it was fricking bilinear

            catrom offers utility in being really fast with basically no ringing, but yes the aliasing is pretty bad as you go up the scale factors... i wouldnt really recommend it for anything beyond +0.50x tbh, but even then lanczos is a little better

            what's the performance difference between spline and lanczos?

            they both have a radius of 3, no difference because the scalers are not actually ran on the GPU as math, they're cached as LUTs. if they weren't spline36 would be substantially faster because its a polynomial function

            [...]
            >this is why i said [...](You)# your interpretation is interesting
            i rather quickly compared them in firefox going off the sky and pole but the zoom in on them were slightly off despite saying the same, disregard. use slow.pics next time

            >use slow.pics next time
            sorry i was lazy 🙁

          • 3 weeks ago
            Anonymous

            i don't need anyone to tell me what i should see in my test images. Maybe it is in-fact a bad test case. But ewa_lanzossharp has more ringing in my crappy test case than spline36. simple as.

          • 3 weeks ago
            Anonymous

            zewia stop trolling with spline36

          • 3 weeks ago
            Anonymous

            >i don't need anyone to tell me what i should see in my test images.
            there isn't really any contrast, its just not that great. i recommend dot.png if all you're trying to do is count the amount of halos, but it wont actually tell you how harshly something rings because its quite literally just 1px dot upscaled to 1000x++ scale factor. though its not like your test image really shows that either. at least it has better contrast than what you're using tho

            I just use spline36 because it's mpv 0.29 gpu-hq default. simple as.

            fake oldchad, use opengl-hq with prescale

          • 3 weeks ago
            Anonymous

            >dot.png
            post pls

          • 3 weeks ago
            Anonymous

            its one of those test images that youll have to make yourself. its just a white 1px dot centered on a grid. i recommend a black or grey background

  29. 3 weeks ago
    Anonymous

    He wins yet again

  30. 3 weeks ago
    Anonymous

    what's the performance difference between spline and lanczos?

  31. 3 weeks ago
    Anonymous

    I just use spline36 because it's mpv 0.29 gpu-hq default. simple as.

  32. 3 weeks ago
    Anonymous

    You guys thought haasn was the mastermind... But it was Wiwaz, author of crop_resize

    • 3 weeks ago
      Anonymous

      >cropping
      uohhhh

  33. 3 weeks ago
    Anonymous

    ravu-lite-ar-r3 vs ewa_lanczossharp??

    • 3 weeks ago
      Anonymous

      artcnn sirs

      • 3 weeks ago
        Anonymous

        too many frame drops, also

        fvck vulkan on nvidia gpu's, I'm using d3d11 now and I can comfortably use deband iterations=16 on my gt 1030 instead of a hard limit of 6 on vulkan because that gave me too many stutters and jitters.

        • 3 weeks ago
          Anonymous

          what the FRICK are you doing anon

          • 3 weeks ago
            Anonymous

            watching my mongolian cartoons and asian movies

            In theory Vulkan should be faster. d3d11 is SM5. Yet in most test d3d11 wins in performance on both NVIDIA and AMD. I have no idea, but something is wrong in Vulkan implementation, because it should be in theory faster.

            something is very wrong with the vulkan implementation, I'm also surprised how d3d11 is giving me much better performance than vulkan

            [...]
            what is the use case for vulkan hwdec? is it just performance?
            d3dva11 increases performance substantially at absolutely no quality loss on my gpu. I've never seen a need for anything else that may be potentially unstable.

            >d3dva11
            how is this one different from d3d11?

        • 3 weeks ago
          Anonymous

          In theory Vulkan should be faster. d3d11 is SM5. Yet in most test d3d11 wins in performance on both NVIDIA and AMD. I have no idea, but something is wrong in Vulkan implementation, because it should be in theory faster.

          • 3 weeks ago
            Anonymous

            wait for haasn's d3d12 and metal backends

          • 3 weeks ago
            Anonymous

            metal may be worth it. d3d12, a lot of work for not much reword. Few minor improvements, but overall nothing that would be a killer feature to be worth investing into it

          • 3 weeks ago
            Anonymous

            it would probably almost have the same performance as d3d11 but the shader compilation speeds would be much faster at least

            doesnt ffmpeg currently fully feature d3d12va?

        • 3 weeks ago
          Anonymous

          In theory Vulkan should be faster. d3d11 is SM5. Yet in most test d3d11 wins in performance on both NVIDIA and AMD. I have no idea, but something is wrong in Vulkan implementation, because it should be in theory faster.

          what is the use case for vulkan hwdec? is it just performance?
          d3dva11 increases performance substantially at absolutely no quality loss on my gpu. I've never seen a need for anything else that may be potentially unstable.

          • 3 weeks ago
            Anonymous

            The use case for vulkan hwdec is being able to decode videos using vulkan, which means you can render with vulkan without having to use a copy-back hwdec.

          • 3 weeks ago
            Anonymous

            i get the idea of using an api just for its own sake but i feel like high-quality kino video decoding is a completely different use case than something like rendering a video game. I would categorically stick to directx just for its safety.

          • 3 weeks ago
            Anonymous

            Video games have videos too

          • 3 weeks ago
            Anonymous

            vaapi hwdec doesn't work on my card if I use vulkan output, but vulkan hwdec does. Thank you amd

  34. 3 weeks ago
    Anonymous

    https://github.com/mpv-player/mpv/pull/13458

  35. 3 weeks ago
    Anonymous

    dscale=lanczos
    dscale-antiring=1

    endgame?

  36. 3 weeks ago
    Anonymous

    GNVME totem just werks

    • 3 weeks ago
      Anonymous

      what the frick is that logo

      • 3 weeks ago
        Anonymous

        SOVL

        • 3 weeks ago
          Anonymous

          move out of the fricking way, max soul incoming

          • 3 weeks ago
            Anonymous

            >https://mpc-hc.org/
            >We are 100% spyware
            >Please switch to something else.

          • 3 weeks ago
            Anonymous

            <hanna> https://mpc-hc.org/2017/07/16/1.7.13-released-and-farewell/
            <hanna> it's official
            <hanna> mpv killed MPC-HC
            <hanna> RIP madshi
            <hanna> what happened to the MPC-HC developers though
            <hanna> did they stop watching video
            <hanna> or did they start using other video players
            <wm4> maybe life happened to them
            <wm4> it's what drains open source projects of power
            <hanna> yes but that still raises the question
            <hanna> they still watch videos during their downtime, no?
            <wm4> maybe they watch stuff on their TV now
            <hanna> do they use VLC now or what
            <wm4> using the netflix app etc.
            <hanna> https://github.com/dragons4life/MPC-HC-config-for-MPV/blob/master/input.conf wew
            <hanna> also whoever edited that has a great font
            <vBm> alex just fell from face of earth, u78 got tired of coding at home since he does code for a living, armada moved on to different things (and switched to mpv iirc :D), kasper is also MIA
            <hanna> frick
            <hanna> all that MIA
            <hanna> crunchyroll got them
            <hanna> curse you daiz
            <Armada> can confirm, mpv killed MPC-HC for me 😉
            <wm4> lol
            <Armada> When I looked at what needed to be done for MPC-HC to have a future, I concluded it basically needed to become MPV 😛
            <wm4> Armada: what does that mean?
            <Armada> wm4, well to start there's no real reason anymore to have a Windows-only video player
            <Armada> and we needed to get rid of the antique UI, preferably replacing it by something as simplistic
            <wm4> that's funny to hear, because mpv having no "GUI" (only config files) is apparently a big deal

          • 3 weeks ago
            Anonymous

            >open source projects of power
            something about that wording doesn't sit right with me

      • 3 weeks ago
        Anonymous

        valid logo

    • 3 weeks ago
      Anonymous

      TRVTHNVKE

  37. 3 weeks ago
    Anonymous

    it's over

    • 3 weeks ago
      Anonymous

      explain

  38. 3 weeks ago
    Anonymous

    why the frick does my sound crash when i use mpv. i have to then kill and restart it and when it comes back the volume is super low and cannot be adjusted..

    • 3 weeks ago
      Anonymous

      stop using pulse

      • 3 weeks ago
        Anonymous

        frick!

        • 3 weeks ago
          Anonymous

          you can try pulse-latency-hacks=yes and see if that works

  39. 3 weeks ago
    Anonymous

    how do I make a comparison between scalers? I want to see with my own eyes which one I like better.

    • 3 weeks ago
      Anonymous

      use these keybinds
      u cycle-values scale nearest bilinear spline36 lanczos ewa_lanczossharp
      U cycle-values scale nearest bilinear spline36 lanczos ewa_lanczossharp

      add any scaler you want to them

      • 3 weeks ago
        Anonymous

        how do I add a glsl shader to those options?

        • 3 weeks ago
          Anonymous

          probably a different keybind cycling through glsl-shader

    • 3 weeks ago
      Anonymous

      use these keybinds
      u cycle-values scale nearest bilinear spline36 lanczos ewa_lanczossharp
      U cycle-values scale nearest bilinear spline36 lanczos ewa_lanczossharp

      add any scaler you want to them

      these are not organic posts... smells like discord

  40. 3 weeks ago
    Anonymous

    also any good test images? I just want to compare ravu-lite-ar-r3 and ewa_lanczossharp

    • 3 weeks ago
      Anonymous

      boku_no_pico.png for anime
      goatse.png for live action

    • 3 weeks ago
      Anonymous

      rose.png

  41. 3 weeks ago
    Anonymous

    help out a vlc re(tard)fugee, what the frick is v3/gcc? which one do i use? are dev just untested builds?

    • 3 weeks ago
      Anonymous

      https://github.com/zhongfly/mpv-winbuild/releases/tag/2024-02-10-f5c4f0b

      grab "mpv-x86_64-v3-20240210-git-f5c4f0b.7z"

    • 3 weeks ago
      Anonymous

      use mpv-x86_64-v3

    • 3 weeks ago
      Anonymous

      >dev
      shared library, not application itself
      >v3
      clang AVX2 enabled build
      >gcc
      gcc builds, for Windows 7 compatibility
      > without suffix
      normal x64 clang build

  42. 3 weeks ago
    Anonymous

    anyone have a guide for hooking up audio to plplay? i want to have ffplay playing audio ONLY in the background and plplay playing video in sync

  43. 3 weeks ago
    Anonymous

    >mpv ffmpeg yt-dlp

  44. 3 weeks ago
    Anonymous

    On vo=gpu where (code-wise) does mpv convert to srgb even when using tag-csp=yes?

  45. 3 weeks ago
    Anonymous

    Also why did wm4 say screenshots are broken if you use target-trc?

    https://github.com/mpv-player/mpv/commit/9f595f3a80eeaa0bfbda5702031f054f684f6123

    >Will break for users who use --target-trc and similar options.

  46. 3 weeks ago
    Anonymous

    Analyzed by haasn here https://github.com/mpv-player/mpv/issues/7840#issuecomment-645874716 - unles it changed since then, on vo-gpu this means taking screenshots will always tag as srgb even if the screenshot data (as retrieved from gpu tex) was for a different csp

    • 3 weeks ago
      Anonymous

      yes it's broken on vo_gpu basically. it always forces sRGB as mentioned in the huhn issue.

      its not too surprising, until very recently it was figured out that vo_gpu actually applies mpv's dithering shader to screenshots, so "high bit depth" screenshots weren't actually "high bit depth", hence why vo_gpu had smaller file sizes. lots of broken screenshot related stuff

      • 3 weeks ago
        Anonymous

        so if things are always tagged srgb then what exactly is the point of tag-csp option, since this existed before gpu-next

        • 3 weeks ago
          Anonymous

          it's literally broken like most things related to vo_gpu. why do you think haasn rewrote it? did you actually fall for the memes?

          • 3 weeks ago
            Anonymous

            it's an easy fix, the rewrite to have screenshot download from gpu happened in 0.29, so likely wm4 just never got around to fixing it. In fact he noted in his commit that this breaks target-csp. Before then screenshots always used the software path (libswscale), and I think in this case the source csp info is properly passed along. (And window screenshots also had a hack to propagate CSP information).

            On gpu-next I think if you set an ICC profile screenshots are still broken: mp_image only has enums for csp, so I can't see how image_writer will ever tag the written image with the ICC profile that was used.

          • 3 weeks ago
            Anonymous

            on vo=gpu if you use target-trc=srgb it ends up accidentally working.

            I think this means that for vo=gpu unless you fix it yourself (it's an easy fix btw, big brain vo=gpu chads can probably figure it out) there are 2 options for "correct" screenshots:

            Use target-trc=srgb to always force conversion to srgb for the rendered image, then things end up working accidentally since tag always uses srgb.
            Use sw screenshots instead of hw screenshots, in this case profile info is passed along properly. Of course now it falls prey to the issue mighthuhn mentioned where a bt709 image will retartedly be tagging with a decode gamma of 1.961. ironically this actually gives you compatibility with quicktime (but only for screenshots) on mac.

          • 3 weeks ago
            Anonymous

            >the rewrite to have screenshot download from gpu happened in 0.29
            0.28 will be the new meme

          • 3 weeks ago
            Anonymous

            If you want 0.28 behavior, disable hw screenshots. Non issue.

          • 3 weeks ago
            Anonymous

            >disable hw screenshots
            how? Can this be done with latest mpv?

          • 3 weeks ago
            Anonymous

            --screenshot-sw=yes

  47. 3 weeks ago
    Anonymous

    should I use the options
    target-trc=gamma2.2
    icc-profile-auto

    in my mpv.conf?

    • 3 weeks ago
      Anonymous

      >icc-profile-auto
      this doesn't do anything

      • 3 weeks ago
        Anonymous

        it does, go check on your mpv shift + I page 2

        "colorspace conversion, ICC 3DLUT, color encoding" shows up when using that option

    • 3 weeks ago
      Anonymous

      no point specifying both, icc management will override target-trc if icc profile is present. I guess the only use-case is if you want a "universal" profile and have it fall back to 2.2 gamma if icc info not available. Also most computer monitors might be closer to srgb instead of 2.2, check rtings review for yours.

      • 3 weeks ago
        Anonymous

        >check rtings review for yours.
        rtings just gives an average gamma value of 2.19. does that mean it's actually gamma2.2?

        • 3 weeks ago
          Anonymous

          Do they have the precise curve there? Technically srgb also averages to 2.2ish. Or another way to do things might be to try the test in https://www.shadertoy.com/view/7sjBWD

          • 3 weeks ago
            Anonymous

            ok it seems like it's a sRGB curve (im assuming the yellow line is sRGB reference)? so technically i shouldnt have to do anything, right? i have a 1000:1 contrast ratio so whether or not i use target-trc=srgb should not matter

          • 3 weeks ago
            Anonymous

            yeah the yellow target is an srgb curve

      • 3 weeks ago
        Anonymous

        Colors: 16.7M Color Space Support: 8 bit (6 bit + Hi-FRC) Color Gamut: NTSC: 72% size (Typ), sRGB: 103% size (Typ)

        no rtings review but this is what I got from their website. worth using "target-trc=gamma2.2" and "icc-profile-auto"?

        • 3 weeks ago
          Anonymous

          still doesn't really tell you much, lots of monitors claim srgb but are actually 2.2 or vice-versa. Only way to know for sure is if someone profiled it and reported the full curve (instead of just average).

  48. 3 weeks ago
    Anonymous

    how to take screenshot automatically converted to srgb so I don't have to worry about color compatibility with viewers/editors?

    • 3 weeks ago
      Anonymous

      https://mpv.io/manual/master/#options-screenshot-tag-colorspace

      also most imageviewers (like jpegview on windows) ignore colortags anyways

    • 3 weeks ago
      Anonymous

      on vo=gpu if you use target-trc=srgb it ends up accidentally working.

  49. 3 weeks ago
    Anonymous

    btw chrome and edge (both chromium obviously) will display bt.1886 screenshots correctly

    in fact, chromium actually properly tonemaps HDR images to SDR as well. it's unironically over for firefox

  50. 3 weeks ago
    Anonymous

    if mpv is built on ffmpeg how does the mpv-webm encode script encode so much faster than native line ffmpeg

    • 3 weeks ago
      Anonymous

      possibly different compression options? Mpv does have its own encoding mode but it should ultimately call into libavcodec at the end of the day, so if all options are matched I don't see how mpv would be slower. Btw for the weenies who keep saying to remove mpv's encoding mode, it's the only way to make use of mpv-specific demuxers.

    • 3 weeks ago
      Anonymous

      because it's no mystery that mpv's encoding is lower quality. It's why people want it gone.

      possibly different compression options? Mpv does have its own encoding mode but it should ultimately call into libavcodec at the end of the day, so if all options are matched I don't see how mpv would be slower. Btw for the weenies who keep saying to remove mpv's encoding mode, it's the only way to make use of mpv-specific demuxers.

      Mpv's encoding mode should ultimately be kept but all encoding related tasks should just be an interface for ffmpeg.

      • 3 weeks ago
        Anonymous

        >It's why people want it gone.
        well those people are homosexuals

      • 3 weeks ago
        Anonymous

        >interface for ffmpeg
        But isn't that what it currently is? It's not like they reimplement an encoder or anything, they just take the demuxed packets and sent it to libavconv. Otherwise you can't use mpv's superior demuxers.

        • 3 weeks ago
          Anonymous

          is it?

  51. 3 weeks ago
    Anonymous

    i dont see a difference between srgb and bt1886, is it over for me bros?

    • 3 weeks ago
      Anonymous

      they are basically the same at 1k contrast ratio (which is what gpu-next defaults to)

  52. 3 weeks ago
    Anonymous

    listen, 2.2 just looks washed out (im not gonna use it haha)

  53. 3 weeks ago
    Anonymous

    Instead of messing around with specs written by boomers, why don't we just take the icc profile for some commonly used reference monitor and assume that as the source profile for bt709 video. Then let the CMS handle conversion to your monitor's profile. Simple as that.

    • 3 weeks ago
      Anonymous

      how about you just whatever looks best. much simpler, no software solutions required

    • 3 weeks ago
      Anonymous

      Btw I'm serious about the proposal, why can't we just do this and relegate bt1886 to the trash bin. We have enough power to do CMS conversion on the fly, so thinking in terms of EOTF is boomer behavior.

  54. 3 weeks ago
    Anonymous

    target-trc=gamma2.4
    brightness=-5
    contrast=3
    saturation=4
    gamma=-13

    try it out, btfos every setup for me

  55. 3 weeks ago
    Anonymous

    >ywn use mpv on an IMAX screen
    why even live

  56. 3 weeks ago
    Anonymous

    Now that the dust has settled who won between the VRRgays and fixed refresh gays?

  57. 3 weeks ago
    Anonymous

    y'all are great at identifying problems but not so good at figuring out solutions. I now know that red is too orange and screenshots have wrong colors, but have no idea how to fix any of it. I guess I'll just keep using defaults.

  58. 3 weeks ago
    Anonymous

    why is there no mpv on winget?

    • 3 weeks ago
      Anonymous

      microshaft knows that vlc is betterer

  59. 3 weeks ago
    Anonymous

    mpv needs to add jxl screenshot support!

    • 3 weeks ago
      Anonymous

      it already has it

      • 3 weeks ago
        Anonymous

        what the... you're lying!

        • 3 weeks ago
          Anonymous

          it does, but it's done through ffmpeg so your ffmpeg needs jxl support to be enabled

    • 3 weeks ago
      Anonymous

      if you use jxl you are supporting troonypora

      • 3 weeks ago
        Anonymous

        AVIF is the troonest codec of all time.

  60. 3 weeks ago
    Anonymous

    whats the next PR that we want merged sisters

    • 3 weeks ago
      Anonymous

      https://github.com/mpv-player/mpv/pull/13172
      lgtm
      https://github.com/mpv-player/mpv/pull/13357
      lgtm

      • 3 weeks ago
        Anonymous

        explain to a techlet what spdif is

        • 3 weeks ago
          Anonymous

          "pure" audio feed that you plug into your 1000$ DAC

          it's snake oil, basically

        • 3 weeks ago
          Anonymous

          the old way of transporting digital audio, it's much simpler than hdmi, so it has stuck around

        • 3 weeks ago
          Anonymous

          rca cable but magic 1s n 0s travel through instead

  61. 3 weeks ago
    Anonymous

    I copied this some time ago from here. Might be helpful to someone.

    > Best FFMPEG Scripts for IQfy <

    >Convert Video to WEBM
    ffmpeg -i INPUT.mp4 -o OUTPUT.webm
    >The -o flag is NOT required

    >Lower Quality Before Conversion
    ffmpeg -i INPUT.mp4 -crf 48 OUTPUT.webm
    >The range is from 0 to 51 where the higher the number, the lower the quality

    >Cut a Specific Portion of a Video
    ffmpeg -i INPUT.mp4 -ss 00:30 -to 01:53 OUPTPUT.webm
    >The -ss flag signifies the start point and -to as the end point

    >Adjust Video Size
    ffmpeg -i INPUT.mp4 -vf scale=1280:720 OUTPUT.webm
    ffmpeg -i INPUT.mp4 -vf scale=iw/2:ih/2 OUTPUT.webm
    >Self-Explanatory; the second one halves the Width and Height or divides by two

    >Combine Video and Audio File
    ffmpeg -i INPUT.mp4 -i INPUT.mp3 -map 0:v -map 1:a OUTPUT.webm

    >Disable Audio Track
    ffmpeg -i INPUT.mp4 -an OUTPUT.webm

    >Convert GIF into a Continuously Looping Video
    ffmpeg -r 7 -ignore_loop 0 -i INPUT.gif -t 30 OUTPUT.webm
    >The -r flag is the Frame Rate; the -t flag is the amount of seconds you want the video to be

    >Overlay an Image at a Specific Time
    ffmpeg -i INPUT.mp4 -i INPUT.jpg -filter_complex "[0:v][1:v] overlay=0:enable='between(t,0,1)'" OUTPUT.mp4
    >You only need to worry about the (t,0,1) which focuses on the length of time the image will be on the screen

    >Crop a Video File
    ffplay -vf "crop=x:y:w:h" OUTPUT.webm
    ffmpeg -vf "crop=500:500" OUTPUT.webm
    >The ffplay program allows us to preview the video as it is being rendered so we can stop it and adjust accordingly
    >Once we are satisfied with the results, we can use FFMPEG to render the actual video file

    >Export Individual Frames to Image Files
    ffmpeg -i input.mp4 %04d.png
    >The %04d indicates how many sequencial numbers you want to use. %04d goes to 9999 while %03d goes to 999
    >Essentially, if you want to export more than 1,000 frames, you'll need to use %04d

    • 3 weeks ago
      Anonymous

      unironically ChatGPT is made for this kind of task

      • 3 weeks ago
        Anonymous

        Yep, but I think reposting such information every once in a while does not hurt, even if it is basic stuff.

  62. 3 weeks ago
    Anonymous

    How can I get ffmpeg to restore the quality of a YIFY encode back to bluray quality?

    • 3 weeks ago
      Anonymous

      ffmpeg -i yify.mp4 --restore-to-bluray-quality -o bluray.mkv

    • 3 weeks ago
      Anonymous

      ffmpeg -i yify.mp4 --restore-to-bluray-quality -o bluray.mkv

      Kek

  63. 3 weeks ago
    Anonymous

    I just found out the sailor senshi are all middle schoolers.

    Why'd they make them so sexy?

  64. 3 weeks ago
    Anonymous

    >https://flathub.org/apps/com.warlordsoftwares.youtube-downloader-4ktube
    yt-dlpxisters????

    • 3 weeks ago
      Anonymous

      command line yt downloaders are tinkertroon shit but
      >flatpak

  65. 3 weeks ago
    Anonymous

    icc works, screenshots with icc work, huhn lost

  66. 3 weeks ago
    Anonymous

    https://github.com/mpv-player/mpv/issues/13439#issuecomment-1937625058
    fcranekino is back on the table

    • 3 weeks ago
      Anonymous

      >What could be the reason for lip-sync issues? Why was this never a problem before? Something must have changed in the last few versions that could affect this...
      Ever heard of the tale of bugmanguy the wise?

      • 3 weeks ago
        Anonymous

        last good version according to fcrane
        >mpv-dev-x86_64-v3-20230903-git-67368ac

        [...]

        at this point he must be bullshiting

  67. 3 weeks ago
    Anonymous

    what's the proper way to watch chinese cartoons nowadays? mpv+anime4k?

    • 3 weeks ago
      Anonymous

      mpv with
      ao=null
      vo=null

    • 3 weeks ago
      Anonymous

      --vo=gpu-next
      --profile=fast
      --deband=no
      --video-sync=display-resample

      • 3 weeks ago
        Anonymous

        Non-meme answer:
        >vo=gpu-next
        >profile=high-quality
        >deband=no
        >ArtCNN or Ravu if your PC is good enough

        >deband=no when most anime has banding out the ass

        • 3 weeks ago
          Anonymous

          Isn't deband broken rn?

          • 3 weeks ago
            Anonymous

            It has always worked like that

        • 3 weeks ago
          Anonymous

          Post example from a title older than 1999 and not a miniencode

        • 3 weeks ago
          Anonymous

          watching material as intended

          • 3 weeks ago
            Anonymous

            The kind of banding that deband seeks to eliminate is most likely not intentional but a result of poor encoding. Doesn't mean adding on a meme filter improves anything, I also keep it off.

        • 3 weeks ago
          Anonymous

          you're being fooled, they have none at all but deband=yes actually adds banding

        • 3 weeks ago
          Anonymous

          i dunno i downloaded an Amazon rip of a show the other day and it had no banding. this is exclusively a crunchyroll/hidive problem, but making an autoprofile for ripper groups is trivial

          Isn't deband broken rn?

          deband-threshold=64
          deband-grain=16

          come home

          The kind of banding that deband seeks to eliminate is most likely not intentional but a result of poor encoding. Doesn't mean adding on a meme filter improves anything, I also keep it off.

          >meme filter
          it gets rid of banding, what's the problem exactly?

          • 3 weeks ago
            Anonymous

            >what's the problem exactly?
            Find a better source.

          • 3 weeks ago
            Anonymous

            How exactly are you supposed to find a better source for a simulcast? Do you have braindamage?

          • 3 weeks ago
            Anonymous

            Accept the banding. It's a web-dl.

          • 3 weeks ago
            Anonymous

            Why should I do that when I can remove it? This just sounds like poorgay/GPUlet cope

          • 3 weeks ago
            Anonymous

            My pc is plenty powerful. There are a few principles here such as:
            >minimalism
            >you can't polish a turd
            >fell for the meme

          • 3 weeks ago
            Anonymous

            poorgay cope

          • 3 weeks ago
            Anonymous

            Do you unironically find reasons to run mpv at 90% CPU/GPU utilization to prove you're not poor? moronic justification whether your answer is yes or no.

          • 3 weeks ago
            Anonymous

            >deband
            >90% CPU/GPU utilization
            anon, it's time to upgrade your thinkpad

          • 3 weeks ago
            Anonymous

            deband barely makes a dent on my GPU utilization and it makes a tangible and visible difference (gets rid of banding)

            if deband makes your cpu and gpu go to 90%, then yes, you're poorgay coping right now

            deband doesn't affect utilization at all on my PC. my point is, if you use deband to avoid "poorgay coping" then that is a very stupid reason.

          • 3 weeks ago
            Anonymous

            >deband uses 90% gpu
            >a-a-actually no it doesnt! haha i was just exaggerating!
            kek poorgay cope

          • 3 weeks ago
            Anonymous

            deband barely makes a dent on my GPU utilization and it makes a tangible and visible difference (gets rid of banding)

            if deband makes your cpu and gpu go to 90%, then yes, you're poorgay coping right now

    • 3 weeks ago
      Anonymous

      Non-meme answer:
      >vo=gpu-next
      >profile=high-quality
      >deband=no
      >ArtCNN or Ravu if your PC is good enough

      • 3 weeks ago
        Anonymous

        thanks bruv

        • 3 weeks ago
          Anonymous

          you fell for meme

      • 3 weeks ago
        Anonymous

        >Non-meme answer
        >ArtCNN or Ravu
        why do you lie

  68. 3 weeks ago
    Anonymous

    bros, I miss -ss, --start= sucks.

  69. 3 weeks ago
    Anonymous

    These are my video settings:
    profile=high-quality
    vo=gpu-next
    gpu-api=vulkan
    scale-antiring=0.6
    deband=no
    Do I need gpu-api=vulkan? I use Linux and have an AMD GPU.

    • 3 weeks ago
      Anonymous

      yes you dumb moronic homosexual
      specify dscale=catmull_rom
      use artcnn + cfl + pixelclipper

  70. 3 weeks ago
    Anonymous

    artoruiz where the hell is the live action specific artcnn???

  71. 3 weeks ago
    Anonymous

    [...]

    [...]

    [...]

    • 3 weeks ago
      Anonymous

      I've seen literal gay porn on this board up for hours and this gets deleted instantly.

      • 3 weeks ago
        Anonymous

        one of the jannies is a resident mpvgay

  72. 3 weeks ago
    Anonymous

    [...]

    [...]

    [...]

  73. 3 weeks ago
    Anonymous

    profile=high-quality
    Rate my config.

Your email address will not be published. Required fields are marked *