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. 1 month ago
    Anonymous

    >ITT: IQfy in 2010

  2. 1 month ago
    Anonymous

    Why haven't we made this a general and made a copy yet?

    • 1 month ago
      Anonymous

      lurk more newbie

  3. 1 month ago
    Anonymous

    >mpv
    Overrated meme.
    >ffmpeg
    Good.
    >yt-dlp
    Good.

    • 1 month ago
      Anonymous

      Name a single better application with which to consoom media with.
      I'll w8

      • 1 month ago
        Anonymous

        plplay just works. no bloat, just pure video quality

        • 1 month ago
          Anonymous

          I don't know what this is but I mostly use mpv for audio

          VLC

          VLC introduces a split second pause between tracks which is extremely distracting to the type of media I consoom. MPV wins by a mile

          • 1 month ago
            Anonymous

            use mpd

          • 1 month ago
            Anonymous

            That is quite literally a million times more complicated than using mpv and it adds nothing that I care about.
            Why do I /need/ some stupid interface that displays all my shit in a totally separate application when I can install one (1) program and play anything I want with
            >mpv /path/to/whatever

          • 1 month ago
            Anonymous

            mpv can't into cue files

          • 1 month ago
            Anonymous

            I don't know or care what a cue file is

          • 1 month ago
            Anonymous

            just say that you're moronic then, even firefox would be a good enough audio player for you

          • 1 month ago
            Anonymous

            >I have a media player that is simple to use, does everything I need, and never breaks
            >NO NO NO you NNNEEEEDDD to install the bloated daemon and start it with soistemD and run a totally separate client to browse your media in a grid like a dumb cow
            >I'm not going to do that I already have a media player and it's better than that
            >you're moronic
            OK

            >I don't know or care what a cue file is
            what the frick? why do you need a music player then you moron

            >Have music files on HDD
            >want to listen to music
            >need a music player

            mpc

            Again, what exactly do I gain from having a separate application just to look at my music folder? I already have everything organized how I want it. I know exactly where every file I want to play is. What do I gain other than 2 extra programs when I already have 1 that does it just fine?

          • 1 month ago
            Anonymous

            >I already have everything organized how I want it.
            I spend 0 hours organizing my music and mpd does it for me based on metadata.

          • 1 month ago
            Anonymous

            >assuming my media files all have metadata
            So half your collection is completely unorganized? Or you purchase / otherwise acquire all your media through totally legitimate or at least semi-professional methods? Or your collection isn't on my level?
            At least most of my collection has either no metadata or completely random metatadata. For example one artist I have a nearly complete collection of would be split into a dozen or so different 'genres' based on metadata.
            Is that really the entire appeal of this thing? That you don't have to organize your music collection? You just let some vague algorithm (which will change if you need to change your music application at some point) do it for you? And I'm the one who's moronic??

          • 1 month ago
            Anonymous

            Why are you afraid of reddit spacing?

          • 1 month ago
            Anonymous

            if you dont have any .cue files then you don't care about music. just use yt-dlp to listen to lo-fi pop or something through mpv, it fits your usecase

          • 1 month ago
            Anonymous

            >if you dont have any .cue files then you don't care about music.
            That is an extremely callous statement. I am done with this thread I can tell the people here are IQfy type morons who can't get enough of the smell of their own farts.
            Get back to me when you hit 8 TB of FLAC files

          • 1 month ago
            Anonymous

            >I don't know or care what a cue file is
            what the frick? why do you need a music player then you moron

      • 1 month ago
        Anonymous

        VLC

      • 1 month ago
        Anonymous

        mpc

  4. 1 month ago
    Anonymous

    mp librempeg yt-dlp

    • 1 month ago
      Anonymous

      >mp
      explain

      • 1 month ago
        Anonymous

        https://github.com/haasn/mp you need special access to recent updates

  5. 1 month ago
    Anonymous

    ayy nice one OP kek'd super hard at the pic lmao!

  6. 1 month ago
    Anonymous

    Frick Jules

    • 1 month ago
      Anonymous

      what did he do this time

      • 1 month ago
        Anonymous

        https://github.com/mpv-player/mpv/issues/13313

        • 1 month ago
          Anonymous

          >OP gets answer in first post
          >Jules turns it into a 26 post issue because he made ravu hook to chroma instead of luma even though it's a luma prescaler
          Classic day on mpv git

        • 1 month ago
          Anonymous

          >Chroma is not downscaled at any point
          If you resize the window to a smaller resolution than the chroma channel, then what happens? Not trying to suggest it's actually downscaled, but I'm guessing it simply drops information and thus is not considered "downscaling" in the typical sense? Could that be comparable to nearest downscaling then?

          • 1 month ago
            Anonymous

            just scroll down
            >Without user shaders, mpv will always upscale the chroma plane to match the luma plane (assuming that the chroma is subsampled), merge them, then covert them to a single RGB plane. After this, the RGB image will be scaled to a target resolution. There's no way to scale the luma and chroma planes independently to meet a target resolution (and there's no reason to outside of performance reasons, as this will result in a quality reduction).

          • 1 month ago
            Anonymous

            >this will result in a quality reduction
            I actually don't think that's true.
            If you scale an image over and over again, you will eventually notice quality loss.
            why wouldn't the same principle apply to chroma/luma? the image would be preserved better if you scale chroma directly to the target resolution rather than matching the luma channel first.
            regardless, the difference would be visually imperceptible, but we're all about that last 0.01% of quality amirite mpvbros?

          • 1 month ago
            Anonymous

            We should probably test this once and for all shouldn't we?

          • 1 month ago
            Anonymous

            https://slow.pics/c/7IwavwxQ
            chroma_rgb.png - MAE: 0.007043695675608207, PSNR: 32.86079619612017, SSIM: 0.9734039300511528, MS-SSIM: 0.9964616465881063

            luma_chroma.png - MAE: 0.007068069668550288, PSNR: 32.75355233637886, SSIM: 0.9724927731996953, MS-SSIM: 0.9963344950437554

            Shaders: https://pastebin.com/HgbA1YJA

            Using this because I already had them written for some reason (I can't remember). The filter is equivalent to the "Lanczos2 approximation" found in FSR (it's actually a polar sinc-sinc). I can change this to any bicubic spline or something like this but even using this shitty filter we can already see that chroma->rgb is better.

          • 1 month ago
            Anonymous

            i'd be interested in the following comparison:
            >reference 4:2:0 1080p frame downscaled to half the resolution
            >chroma channel unscaled, luma channel downscaled to half the resolution to match unscaled chroma
            curious if there's a perceptible difference

          • 1 month ago
            Anonymous

            1/2

          • 1 month ago
            Anonymous

            2/2

          • 1 month ago
            Anonymous

            Bonus

          • 1 month ago
            Anonymous

            2/2

            isn't this still upscaling chroma? If the luma channel is 1920x1080 then a 4:2:0 chroma channel would be 960x540

          • 1 month ago
            Anonymous

            Turns out doing it with a less sharp filter makes the difference way smaller. These were done with polar mitchell:

            Chroma->RGB

          • 1 month ago
            Anonymous

            Luma->Chroma

          • 1 month ago
            Anonymous

            Difference map

          • 1 month ago
            Anonymous

            Luma->Chroma

            Difference map

            polar_mitchell_chroma_rgb.png - MAE: 0.008176129465096223, PSNR: 30.970645929956248, SSIM: 0.9654901112384638, MS-SSIM: 0.9951366041997107

            polar_mitchell_luma_chroma.png - MAE: 0.008148925072054263, PSNR: 30.993195292775336, SSIM: 0.9657777819897051, MS-SSIM: 0.9951870184611428

            Scaling them separately got better scores with polar mitchell, I guess this is a more intricate matter than it seems. Logically speaking though, I think scaling the channels separately should be generally better as you'll only be introducing scaling artifacts once (per channel).

          • 1 month ago
            Anonymous

            Why would you match the chroma to a downscaled version of the luma? Your luma is your original high resolution image. The entire point of chroma scaling is to reconstruct the 4:4:4 image as best as you can. Having both the luma and chroma scale independently is not correct.

          • 1 month ago
            Anonymous

            >Having both the luma and chroma scale independently is not correct.
            According to who? It is just a different strategy. Sure, upscaling to 4:4:4 first is closer to reverse pipeline of encoding process. But nothing wrong with doing it differently.

          • 1 month ago
            Anonymous

            >According to who?
            GODshi, the alpha and omega of video rendering. someone posted his quotes last thread. i trust GODshi (pbuh) with my life and have an ENVY shrine dedicated to him

          • 1 month ago
            Anonymous

            madshi and haasn vs wm4

            wm4 implemented the original vo_gl to do luma and chroma scaling separately. madvr and gpu_next merge them before doing anything else.

          • 1 month ago
            Anonymous

            >wm4 implemented the original vo_gl to do luma and chroma scaling separately
            based performance optimizer

            cscale should also ALWAYS be set to bilinear

        • 1 month ago
          Anonymous

          >It's just the RGB version of the Ravu-Zoom-AR shader but with this:
          //!DESC RAVU-Zoom-AR (rgb, r3)
          //!HOOK CHROMA
          //!BIND HOOKED
          //!BIND ravu_zoom_lut3
          //!BIND ravu_zoom_lut3_ar
          //!WIDTH OUTPUT.w 2 *
          //!HEIGHT OUTPUT.h 2 *
          //!OFFSET ALIGN

  7. 1 month ago
    Anonymous

    profile=high-quality
    Rate my config.

    • 1 month ago
      Anonymous

      It's shit.

  8. 1 month ago
    Anonymous

    real thread

    [...]

    [...]

    [...]

    • 1 month ago
      Anonymous

      Fake ass thread.

    • 1 month ago
      Anonymous

      based and oldgaypilled thread

      diaper fetishist thread

  9. 1 month ago
    Anonymous

    why did you people have to ruin these threads

    • 1 month ago
      Anonymous

      just take a break, nothing major is going on anyway until haasn is done wageslaving with his yuvj removal patchset for ffmpeg. trolls will stop making this thread once nobody posts in it

      • 1 month ago
        Anonymous

        when's a good time to check back in? How do other people know when to tune in? Are they checking out the github commits daily?

        • 1 month ago
          Anonymous

          https://desuarchive.org/g/thread/98462225/#98464338

          • 1 month ago
            Anonymous

            was mpv good until 0.36.x? why was it garbage after version number increment?

          • 1 month ago
            Anonymous

            no 0.36 was horrible, 0.37 is the final nail in the coffin

          • 1 month ago
            Anonymous

            It was good until 0.30. Then wm4 left and all the colorautists took over, and basic features stopped working

          • 1 month ago
            Anonymous

            don't spawn Him you idiot

      • 1 month ago
        Anonymous

        even worse. haasn does it for FREE

        • 1 month ago
          Anonymous

          he's actually not doing it for free I think, it's contract work for VideoLAN

  10. 1 month ago
    Anonymous

    does anyone know how i can change the specified framerate of the container without affecting playback at all (boxed in red)? ffmpeg's IVTC (fieldmatch,decimate) filter is speeding up this video and I'm confident it's because whoever remuxed this specified double the framerate for whatever reason.

    • 1 month ago
      Anonymous

      Have you tried the pullup filter?

      • 1 month ago
        Anonymous

        all that one ever does for me is introduce judder. "fieldmatch,decimate" is the ivtc filter that just werks if the video is telecined predictably.

  11. 1 month ago
    Anonymous

    >start a SubsPlease encode with --cache
    >press u
    >subtitles freeze
    mpv 0.37 is shit

    • 1 month ago
      Anonymous

      >SubsPlease encode

      • 1 month ago
        Anonymous

        Aren't those just crunchy roll rips? They don't do any re-encoding (at least they shouldn't).

      • 1 month ago
        Anonymous

        It breaks with all mkv files. I specified SubsPlease because they're widely available, which makes reproduction easier.

        • 1 month ago
          Anonymous

          I just stick to 0.37.0
          It's very stable

        • 1 month ago
          Anonymous

          werks for me
          post config and log

          • 1 month ago
            Anonymous

            Used --no-config.

  12. 1 month ago
    Anonymous

    ffmpeg is too strong bros..... I found out today I can use it to convert rtsp to hls to make a nice local web front end for my security cameras with ptz controls.

    • 1 month ago
      Anonymous

      ffmpeg doesn't even have software defined radio, it's not complete yet

  13. 1 month ago
    Anonymous

    merge le hec-- you know what.. nevermind... what's the point.......

  14. 1 month ago
    Anonymous

    Explain chroma sampling location to me, how can someone sample chroma between two pixels, what does that even mean?

    • 1 month ago
      Anonymous

      https://en.wikipedia.org/wiki/Subpixel_rendering

      • 1 month ago
        Anonymous

        What does that have to do with chroma subsampling?

        • 1 month ago
          Anonymous

          read this
          https://en.wikipedia.org/wiki/Chroma_subsampling

          • 1 month ago
            Anonymous

            I already did and didn't understand the "left" sampling method that samples the midpoint of the left side of 2x2 pixel square.

          • 1 month ago
            Anonymous

            that is because the wiki infographic is wrong, this is more accurate. chroma loc determines which "box" you take chroma information from

          • 1 month ago
            Anonymous

            But ffmpeg really does have "bottom", "top" and "center" as valid values for chroma_sample_location option, what are those for?

          • 1 month ago
            Anonymous

            not him, but where the chroma samples are in relation to the luma samples can be different depending on the format used
            think of it like how chroma samples are aligned with luma samples
            https://www.mir.com/DMG/chroma.html

          • 1 month ago
            Anonymous

            What does the MPEG-2 variety mean if I have a 4:4:4 video that I'm subsampling into 4:2:0 with that variety?

          • 1 month ago
            Anonymous

            i believe full-frame mpeg2 chroma location is the most common nowadays
            full frame meaning progressive scan, or not interlaced, most video isn't interlaced nowadays, unless you're ripping a dvd or anything older

          • 1 month ago
            Anonymous

            That's not what I asked

          • 1 month ago
            Anonymous

            then i don't understand what you're asking

            why do I just KNOW that this screenshot is from a korean live show

            it's from a short of piers morgan

          • 1 month ago
            Anonymous

            If I'm writing code to convert yuv444p to yuv420p without ffmpeg, how would I actually do that? Which pixel out of four pixels in the 2x2 square would I pick as a chroma for that 2x2 square? The MPEG-2 variety in the page you linked, as I understand it, says that the chroma between two left pixels need to picked. How is that possible?

          • 1 month ago
            Anonymous

            by blending them

          • 1 month ago
            Anonymous

            why do I just KNOW that this screenshot is from a korean live show

  15. 1 month ago
    Anonymous

    if you think about it, psychovisually, our eyeballs have subsampled chroma

  16. 1 month ago
    Anonymous

    https://haasn.dev/posts/2016-12-25-falsehoods-programmers-believe-about-%5Bvideo-stuff%5D.html
    >> Falsehoods programmers believe about.. image scaling
    >chroma and luma can/should be scaled separately

    • 1 month ago
      Anonymous

      hackernews shit on this post when it was posted there, I trust them more than your fake god

      • 1 month ago
        Anonymous

        >hackernews shit on this post when it was posted there
        isn't this literally just plebbit 2.0? post the thread

        • 1 month ago
          Anonymous

          https://news.ycombinator.com/item?id=13259686

          • 1 month ago
            Anonymous

            i don't see anyone shitting on this article though, why did i waste my time reading this. it seems like they just dont like the meme format because it doesn't provide examples

    • 1 month ago
      Anonymous

      it depends on his justification.
      if it's "it affects image quality by .01%" then that means it can and does make a difference.
      if it's "it makes no difference whatsoever", then that remains to be proven

      • 1 month ago
        Anonymous

        It makes a difference, scalers don't behave the same on all scaling factors and they also don't behave the same on all colour spaces.

        Whether or not you should always scale chroma first is debatable though, it doing them separately provides a meaningful performance improvement then I think the quality drop is justified.

        • 1 month ago
          Anonymous

          if doing*

    • 1 month ago
      Anonymous

      haasn is a piece of art... I believe madshi more

      • 1 month ago
        Anonymous

        both haasn and madshi agree on this topic, your choices are haasn and madshi or wm4

        • 1 month ago
          Anonymous

          yet madshi implemented it, because he believed it is useful for performance reasons

          • 1 month ago
            Anonymous

            we need madshi's input on this as we all know he posts on this thread

      • 1 month ago
        Anonymous

        sister.... both madshi and haasn believe the same thing....

        >Unchecking that option always upscaled chroma first, then converts to RGB, then scales the RGB image to the final resolution. Checking that option scales Y and the CbCr channels separately, if doing so improves performance. Quality is slightly better when always upscaling chroma first.

        >See the madVR trade quality option "scale chroma separately, if it saves performance". You'll get slightly higher quality if you upscale chroma to 4:4:4 first, then downscale both luma+chroma at the same time. But if you want to save performance, you can use that trade quality option. Then madVR won't upscale chroma in your specific situation.

  17. 1 month ago
    Anonymous
    • 1 month ago
      Anonymous

      Stark is such a lucky guy

      • 1 month ago
        Anonymous

        silly weeb

        >just frick my depth perception up pham

        yes, frick you!

    • 1 month ago
      Anonymous

      >just frick my depth perception up pham

    • 1 month ago
      Anonymous

      config?

      • 1 month ago
        Anonymous

        niggy?

        • 1 month ago
          Anonymous

          anime4k monk

          • 1 month ago
            Anonymous
        • 1 month ago
          Anonymous

          do you intentionally use the vomit dithering every time you post? pretty easy to track you

          • 1 month ago
            Anonymous

            how do I avoid the dithy?

          • 1 month ago
            Anonymous

            just use fruit instead of making everything look like pic?

          • 1 month ago
            Anonymous

            jesus christ this is why error diffusion is a meme

            another observation that GODshi correctly made years ago when he said static and evenly distributed dither patterns are superior

          • 1 month ago
            Anonymous

            >another observation that GODshi correctly made years ago when he said static and evenly distributed dither patterns are superior
            wrong, he never made that claim

          • 1 month ago
            Anonymous

            his personal advice has always been to use a static ordered pattern. (which means disabling "change dither for every frame"). error diffusion has always been pointless bloat

          • 1 month ago
            Anonymous

            but I have nothing enabled for dithering, so fruit is the default?

          • 1 month ago
            Anonymous

            >so fruit is the default?
            yes

          • 1 month ago
            Anonymous

            >yes
            so why the frick is Anon hissing at me if I'm using what he's telling me to use??

            [...]
            [...]
            [...]
            [...]
            [...]
            if this visual diarrhea looks terrible in still-shots, i cant imagine how bad it looks in motion

            reeeeeeeeeeeeee

          • 1 month ago
            Anonymous

            I force dithering and I've never seen anything like that. I'd check --no-config

    • 1 month ago
      Anonymous

      silly weeb
      [...]
      yes, frick you!

      someone please help me pick a release (I'll get all the episodes)

      • 1 month ago
        Anonymous

        9volt > Breeze (mini 9volt) > CR > Judas
        Anime4K will probably like Judas more if you're into that disgusting soulless shit for Frieren.

        • 1 month ago
          Anonymous

          What's the deal with that 9volt release? They're like 3 GB an episode. Is that encoded from a blu ray?

          • 1 month ago
            Anonymous

            Grain + transparent encode autism

          • 1 month ago
            Anonymous

            is it artificial grain? and if so, why?

          • 1 month ago
            Anonymous

            Ask Sam. I assume they regrained when they probably rescaled.

            unironically the one that's just a straight rip of the version on crunchy roll. I think that's SubsPlease but I don't watch seasonals so double check that.
            i'm assuming there's no BD yet, so if that's the case then anything else is just going to be a miniencode.

            >miniencode.
            Minis can unironically be superior to the web releases sometimes. rescaling, good debanding, and other artifact removal and filtering help a lot if you're willing to sacrifice minimal detail loss.

          • 1 month ago
            Anonymous

            probably. I'm out of the loop on seasonals so I don't know how those are. All I know is i'm pretty jaded on most mini-encodes for classic anime so I try to hunt down remuxes, or an encode from JySzE/SoM affiliates.

      • 1 month ago
        Anonymous

        unironically the one that's just a straight rip of the version on crunchy roll. I think that's SubsPlease but I don't watch seasonals so double check that.
        i'm assuming there's no BD yet, so if that's the case then anything else is just going to be a miniencode.

      • 1 month ago
        Anonymous

        i like subsplease though afaik their subtitles are hardcoded

        • 1 month ago
          Anonymous

          They aren't

    • 1 month ago
      Anonymous

      silly weeb
      [...]
      yes, frick you!

      niggy?

      how do I avoid the dithy?

      but I have nothing enabled for dithering, so fruit is the default?

      if this visual diarrhea looks terrible in still-shots, i cant imagine how bad it looks in motion

  18. 1 month ago
    Anonymous

    Any webm.lua gays?
    How do I edit the script so vp9 does 10bit encodes. You can't just pass pix_fmt to mpv through the ovcopts. Apparently mpv gets its color space from applying a videofilter but I don't really understand the script or mpv enough to make it stop applying 8bit colorspace every encode.

  19. 1 month ago
    Anonymous

    >mfw when all my sm screenshots are missing a line of pixels, fixed simply by an aspect ratio override
    guess i gotta take them all over again

    • 1 month ago
      Anonymous
  20. 1 month ago
    Anonymous

    madshi was right about ringing
    madshi was right about spline36
    madshi was right about chroma

    why haven't you apologized yet IQfy? you should buy an ENVY right now

    • 1 month ago
      Anonymous

      qrd on those statements?

      • 1 month ago
        Anonymous

        madvr's ultimate scalers is jinc3ar and mpv has only recently been able to replicate it after a decade
        madshi correctly identified that spline36 is inferior to lanczos years ago, and mpv only recently replaced spline36 with lanczos
        see above discussion for chroma

        you can either interpret this as madshi being ahead of mpv at all times, or mpv being one step behind at all times. either way, mpv keeps chasing after madvr's superiority

        • 1 month ago
          Anonymous

          except spline36 vs lanczos always seems to look better in mpv image comparisons?

          • 1 month ago
            Anonymous

            spline36 is not only not as sharp as lanczos, but it's also more aliased than lanczos. it's a useless meme scaler. madshi is always right baby

          • 1 month ago
            Anonymous

            spline was as sharp as lanczos with less ringing every single time i've seen a comparison

          • 1 month ago
            Anonymous

            same

          • 1 month ago
            Anonymous

            yes it rings less because it's blurrier and more aliased than lanczos, this isnt an astute observation

        • 1 month ago
          Anonymous

          The strangest part is that IQfy both circlejerks madshi but then went apeshit after people started to push replacements for spline36. I guess if you aren't madshi IQfy doesn't care what you think even if you're right.

  21. 1 month ago
    Anonymous

    I heard the spline36 schizo finally killed himself last night.

    • 1 month ago
      Anonymous

      oh wait, he's still alive just above, nevermind

    • 1 month ago
      Anonymous

      never forgetti

      • 1 month ago
        Anonymous

        based
        i think he lost his mind after this and started spamming fat porn

        • 1 month ago
          Anonymous

          speaking off, can someone post, I need for science...

  22. 1 month ago
    Anonymous

    could someone make a turbozoomed slow.pics of the french embassy image (perfect) spline v. lanczos i'm fricking tired of this shit

    • 1 month ago
      Anonymous

      post image
      and yeah, ime spline36 always produces a sharper image with less haloing compared to lanczos

      • 1 month ago
        Anonymous

        >ime spline36 always produces a sharper image
        this is literally mathematically impossible outside of your schizophrenic delusions

        • 1 month ago
          Anonymous

          it looks just as sharp

          • 1 month ago
            Anonymous

            you're eyes are not a metric

      • 1 month ago
        Anonymous

        >post image
        https://en.wikipedia.org/wiki/File:(Castelmaurou)_-_Fa%C3%A7ade_de_la_mairie.jpg

        • 1 month ago
          Anonymous

          https://slow.pics/

        • 1 month ago
          Anonymous

          https://slow.pics/

          https://slow.pics/c/rEROVRHv
          320x194 image upscaled to 1781x1080. also threw in nearest and bilinear.
          as usual, no perceivable difference between spline36 and lanczos.
          in scenarios in which there is a perceivable difference, spline36 is typically sharper with less haloing (animation or pixelated content). Most real life pics show no difference.

          • 1 month ago
            Anonymous

            Ahh yes lets compare kernels with an image already full of ringing.. fricking moron lmao.

          • 1 month ago
            Anonymous

            the tree branches look sharper with lanczos, i dont really see any difference in ringing, both of them ring pretty harshly

            lanczos wins basically as most people already surmised

          • 1 month ago
            Anonymous

            >'sharper'
            ringing le good

          • 1 month ago
            Anonymous

            >in scenarios in which there is a perceivable difference
            where?

          • 1 month ago
            Anonymous

            heavily pixelated content
            in such cases, spline36 has visibly less haloing.
            https://slow.pics/c/G7BRRTVO

          • 1 month ago
            Anonymous

            jesus christ spline36 is so disgustingly blurry, may as well use catrom at that point

          • 1 month ago
            Anonymous

            spline36 and lanczos are nearly identical in that comparison except for a slight halo around lanczos.

          • 1 month ago
            Anonymous

            That's why you use AR with lanczos

          • 1 month ago
            Anonymous

            post best AR value
            mine: 0.97813578463387157835125347.548723!@353

          • 1 month ago
            Anonymous

            your comparison just shows that they both ring but spline36 is blurrier and more aliased. you keep pivoting from comparison image to comparison image and getting btfo constantly

          • 1 month ago
            Anonymous

            I have yet to see this "btfo" you speak of. Is it in the room with you right now?

          • 1 month ago
            Anonymous

            put your avatar back on you moron and stop bypassing my filters

          • 1 month ago
            Anonymous

            you choose to read, and you choose to engage

  23. 1 month ago
    Anonymous

    >someone posts the exact same thread every day
    >people come in to reply to it every day
    >people have the same exact conversations in it every day
    grim

    • 1 month ago
      Anonymous

      What can we talk about, women? We have none.

    • 1 month ago
      Anonymous

      what actually happens:
      >someone posts one of the nth concurrently running generals on IQfy
      >people reply in it every day
      >newbies ask the same moronic questions and have the same exact moronic misconceptions every thread
      >general lurkers have to keep correcting them every thread or else the thread will get overrun by moronic newbies repeating moronic info at each other every thread
      this is how /mpv/ died and was the source of a bunch of moronic cargocults btw. thankfully our prophet sizumam corrected the record and killed all cargocults forever

    • 1 month ago
      Anonymous

      have you ever even had ONE hobby

  24. 1 month ago
    Anonymous

    libplacebo release when

  25. 1 month ago
    Anonymous

    hot chroma on luma action

  26. 1 month ago
    Anonymous

    was heckin cache merged?

    • 1 month ago
      Anonymous

      libplacebo release when?

  27. 1 month ago
    Anonymous

    I still can't believe haasn died.

    Wonder how used his account to merge these two PRs last week?

    • 1 month ago
      Anonymous

      automerge probably

    • 1 month ago
      Anonymous

      https://github.com/mpv-player/mpv/issues/13323#issuecomment-1902076342
      haasn died and got replaced by a russian

      • 1 month ago
        Anonymous

        What is haasn's opinion on the Ukraine war?

        • 1 month ago
          Anonymous

          I don't care what your personal politics are, but you have to be seriously deranged if you don't want peace ASAP. The war has done nothing but completely destroy Ukraine as a country and an entire generation of Ukrainian men.

          • 1 month ago
            Anonymous

            How would you achieve peace? Ask Putin nicely?

          • 1 month ago
            Anonymous

            Be an adult and concede several territories to Russia and ban Ukraine from ever joining NATO. Not really complicated. The war, which was never winnable by Ukraine obviously, could have ended with minor concessions to Russia in spring of 2022, but the bloodthristy maniacs that run the west couldn't have that of course.

          • 1 month ago
            Anonymous

            >just consneed territory to every invading force
            not even a nafo shill but surely you see the issue with doing this

          • 1 month ago
            Anonymous

            That's not what I said. It was unwinnable from the getgo unless you literally want to start WWIII over some backwards slav shithole.

          • 1 month ago
            Anonymous

            >giving Russia what it wants
            >not considering that Russia is pretty greedy
            You're a cuck.

            if the US didn't refuse to even sit down at the negotiating table, as putin called for several MONTHS before the war started, conceding territories wouldn't have been necessary. this is by no means a sudden outlash by a crazy dictator as many seem to think, this has been brewing for years, with the euromaidan coup 2014 being major escalation

          • 1 month ago
            Anonymous

            homosexual, all of Europe tried to convince Putin not to be a moron.

          • 1 month ago
            Anonymous

            and putin insisted that he wanted to negotiate with the US, which they refused to

          • 1 month ago
            Anonymous

            How is the US even related to the Ukraine, moron.

          • 1 month ago
            Anonymous

            uh did you miss the part where the US is defacto in charge of NATO and that billions of US taxpayer dollars went to ukraine. We even pay the salaries of their politicians

          • 1 month ago
            Anonymous

            how can you be this naive? i mean, just see the related link

            [...]
            if the US didn't refuse to even sit down at the negotiating table, as putin called for several MONTHS before the war started, conceding territories wouldn't have been necessary. this is by no means a sudden outlash by a crazy dictator as many seem to think, this has been brewing for years, with the euromaidan coup 2014 being major escalation

            i don't get why you're holding on to some sort of belief that the war is winnable when there's tons of discussion about peace talks within a group of allied countries (see meloni prank call and NEWS et al) and the mainstream media narrative has done an 180, a ton of unecessary suffering later

          • 1 month ago
            Anonymous

            >giving Russia what it wants
            >not considering that Russia is pretty greedy
            You're a cuck.

        • 1 month ago
          Anonymous

          war?

          • 1 month ago
            Anonymous

            sorry I meant special operations, Nikolai Haas

  28. 1 month ago
    Anonymous

    How do I find out which default has changed after a build? Everything past mpv-x86_64-v3-20230917-git-181eddc is fricked.
    >pic related
    top: mpv-x86_64-v3-20230924-git-140d018 - mpv-x86_64-v3-20240114-git-bd35dc8
    bottom: mpv-x86_64-v3-20230917-git-181eddc

    • 1 month ago
      Anonymous

      They use the same config:
      hwdec=d3d11va
      vo=gpu-next
      gpu-api=d3d11
      gpu-context=d3d11

      Both do h/w decoding, but one of them has some autism default enabled I assume.

    • 1 month ago
      Anonymous

      They use the same config:
      hwdec=d3d11va
      vo=gpu-next
      gpu-api=d3d11
      gpu-context=d3d11

      Both do h/w decoding, but one of them has some autism default enabled I assume.

      profile=fast

      • 1 month ago
        Anonymous

        Yes, I just discovered it. Thank you! I like to leave MPV playing shit when I sleep, and I need it to work fanless.

    • 1 month ago
      Anonymous

      ENJOY!

    • 1 month ago
      Anonymous

      mpv --list-options
      Then use a text diff thing.

      • 1 month ago
        Anonymous

        Than's a good idea for future. Right now it's fixed and time for me to go back to sleep. It's middle of the night, the fans literally woke me up.

  29. 1 month ago
    Anonymous

    The nu-mpv experience:
    >Load up your favorite show
    >mpv takes 6 gorillion nanoseconds to start
    >fans start blazing thanks to bugplacebo
    >Hah sucks for potato pc users. Serves you right for not updooting!
    >Show starts. Hey wait a minute, wasn't her dress supposed to be red?
    >Try to seek back just to confirm, but bugmanguy's commits broke seeking
    >Open up console.lua and furiously start typing away, adjusting the target response curve.
    >Screen can't keep up since osc refresh rate is still tied to video fps.
    >Hey but at least we have autocomplete.

    • 1 month ago
      Anonymous

      Simpler:
      >enable cache
      >press u
      >subtitles freeze

  30. 1 month ago
    Anonymous

    Lanza vs Eva_Lanza ??

  31. 1 month ago
    Anonymous

    imagine using diaper36 in 2024

  32. 1 month ago
    Anonymous

    https://github.com/mpv-player/mpv/pull/12802
    bugs incoming, brace yourselves

    • 1 month ago
      Anonymous

      nam flashbacks to statusline which still isn't fully fixed

    • 1 month ago
      Anonymous

      giving away even more control to libplacebo when we're in le cache fiasco

  33. 1 month ago
    Anonymous

    whats the best way to watch anime on a Thinkpad x200?

  34. 1 month ago
    Anonymous

    anyone know which ffmpeg filter can be used to scale to an aspect ratio? (not by providing hard x and y values, but by providing an aspect ratio like 4:3 and it's smart enough to scale)

    • 1 month ago
      Anonymous

      you can use math with the scale filter
      scale=(ih*(4/3)):ih (adjusts width to make the video 4:3 with square pixels)
      though i would generally recommend just setting the aspect ratio with "-aspect 4:3" instead of scaling the video unless you really need to scale the video

      • 1 month ago
        Anonymous

        i needed an ffmpeg filter so that webm.lua would catch the scaling. sank yuuu.

        • 1 month ago
          Anonymous

          that is an ffmpeg filter

  35. 1 month ago
    Anonymous

    >switch to vo=gpu-next
    >screenshot-directory="..." no longer works
    ??????

    • 1 month ago
      Anonymous

      you're a homosexual moron get the frick out of IQfy

    • 1 month ago
      Anonymous

      had nothing to do with gpu-next
      it seems some settings only work without making a mess when they're at the very end of mpv.conf

  36. 1 month ago
    Anonymous

    i was originally using video-aspect-override to change aspect ratio to 4:3 like so:
    A cycle-values video-aspect-override "-1" "4:3"

    are there any downsides to replacing it with the following ffmpeg filter instead?
    A cycle-values vf "" "scale=(ih*(4/3)):ih"

    • 1 month ago
      Anonymous

      yes, you may reduce the quality of the video
      for example if you ran 720x480 dvd video through it to adjust it to 4:3 (square), then the video will become 640x480, before being scaled to your screen, you'd reduce 720 to 640, throwing away information

      • 1 month ago
        Anonymous

        yes, i caught that it's only scaling x. i'm only using it to fix the bad scaling output on this sm remux (767 width instead of 768 for whatever reason)

        • 1 month ago
          Anonymous

          they may have cropped a bad line, i would just leave it alone

          • 1 month ago
            Anonymous

            nah, it's scaled output. source video is 720x576 and is supposed to be scaled to 768.

          • 1 month ago
            Anonymous

            i see
            still not really sure why you need to scale anything
            the video stream should be flagged as 4:3 and that's all you should need to do

          • 1 month ago
            Anonymous

            the container itself is marked as 767x576. so it outputs at the wrong resolution unless i correct it

          • 1 month ago
            Anonymous

            so it's 720x576 pixels, with a DAR of 767x576?
            i would just use ffmpeg or mkvmerge(gui) to change the DAR to 4:3 in the files

          • 1 month ago
            Anonymous

            that is;
            ffmpeg -i source.mkv -c copy -map 0 -aspect 4:3 destination.mkv

          • 1 month ago
            Anonymous

            yes

            that is;
            ffmpeg -i source.mkv -c copy -map 0 -aspect 4:3 destination.mkv

            is there a way to do this recursively for every file in a folder?

          • 1 month ago
            Anonymous

            mkdir out; for i in *.mkv; do ffmpeg -i "$i" -c copy -map 0 -aspect 4:3 out/"$i"; done

            windows (had to look it up, haven't touched cmd in ages, so may have mistakes)
            mkdir out && for %i in (*.mkv) do ffmpeg -i "%i" -c copy -map 0 -aspect 4:3 out"%i"

          • 1 month ago
            Anonymous

            Thanks.
            Also, does the scale filter in ffmpeg work the same as setting -aspect (I'm using the filter with vf= in mpv) or is there a different filter for that?

          • 1 month ago
            Anonymous

            aspect set a flag to say how the video should be displayed, but doesn't change the pixels stored in the file
            scale changes the pixels in the file
            so like in clip related, i used scale to change the video from 1920x1080 to 1440x1080, and then aspect 16:9 to make it display as 16:9 (ffmpeg preserves the aspect by default anyway, but you get the idea)

  37. 1 month ago
    Anonymous

    sorry to interrupt your totally inane autism filter discussion, but yt-dlp, actually important software, has been unusable on youtube for two days

    • 1 month ago
      Anonymous

      unironically werks on my machine

    • 1 month ago
      Anonymous

      So apparently the problem is that one of youtube's servers is broken. Youtube provides stream URLs with both a main server plus a fallback server included in the parameters, and while youtube's website player will correctly use the fallback server, yt-dl just fails forever. Without actually fixing yt-dlp's extractor, a workaround can simply pass %(url)q and %(filename)q to an external script run via --exec to fix the URL and fetch from fallback server before yt-dlp attempts to download like a moron

  38. 1 month ago
    Anonymous

    I hate this sailor moon poster shithead

    • 1 month ago
      Anonymous

      Lanza vs Eva_Lanza ??

    • 1 month ago
      Anonymous

      We all do. Tripgays ruin the general

    • 1 month ago
      Anonymous

      has the sailor moon moron been banned? I hate that fricking thirdie so much thank god he's offline during white hours

      funny you say that, not gonna give too many details but I'm a white mid-western American. I thought you were a thirdie because you started posting when I had to go to sleep.

      • 1 month ago
        Anonymous

        nig when tf do you sleep it's almost 14:30 now

        • 1 month ago
          Anonymous

          went to sleep around 3:30 am and woke up around 11:30 am cuz it's saturday (yes i am an adult with a well paying job)

          • 1 month ago
            Anonymous

            I can smell the immigrant in you, you are not white.

  39. 1 month ago
    Anonymous

    mpv idiot here, in the last week I updated mpv on an Arch distro but now I've noticed two things
    >Mouse scroll changes volume now instead of scrolling through video time like it always used to
    >Screenshots are saving at a much higher file size
    Did something change lately or did my conf frick up? Never even looked at it in years so wondering if they changed something or I fricked something

    • 1 month ago
      Anonymous

      Default scroll wheel behavior was changed, you can change it back if you want. I'd assume the same for screenshots, but I haven't heard that specifically yet.

      • 1 month ago
        Anonymous

        Ok thanks for the intel, reverted the scroll behaviour using input.conf.
        If the screenshots changed then something must be different if
        screenshot-format=png
        screenshot-png-compression=0.2

        or whatever suddenly saves 1080p screenshots at 8+MB instead of 4MB. Will post here if I find anything

        • 1 month ago
          Anonymous

          my screenshots save at 26~28MB with compression=0
          (´꒳`)

          • 1 month ago
            Anonymous

            >26~28MB
            o_O
            For some reason only one 1080p anime is screenshotting at 8MB, the others are normal so there's nothing wrong with the config after all
            Guess we'll never know...

    • 1 month ago
      Anonymous

      Ok thanks for the intel, reverted the scroll behaviour using input.conf.
      If the screenshots changed then something must be different if
      screenshot-format=png
      screenshot-png-compression=0.2

      or whatever suddenly saves 1080p screenshots at 8+MB instead of 4MB. Will post here if I find anything

      The screenshot thing is because it was actually bugged and wrong for years. The higher size is correct. You can use --screenshot-high-bit-depth=no instead.

  40. 1 month ago
    Anonymous

    https://github.com/mpv-player/mpv/pull/13317#issuecomment-1902077322

    • 1 month ago
      Anonymous

      librempeg support when?

  41. 1 month ago
    Anonymous

    has the sailor moon moron been banned? I hate that fricking thirdie so much thank god he's offline during white hours

  42. 1 month ago
    Anonymous

    How do I compress below a specific size? let's say getting a video down under 10MB?

    • 1 month ago
      Anonymous

      trial and error

    • 1 month ago
      Anonymous

      You can set the target bitrate to somethings that should get under it for a given length, but it's not a guarantee.

  43. 1 month ago
    Anonymous

    112 prs, is mpv saved sisters?

    • 1 month ago
      Anonymous

      We're still at 900+ issues sister

      if it hits 1000 the repo explodes

      • 1 month ago
        Anonymous

        DMG saved us

        • 1 month ago
          Anonymous

          another catastrophe avoided

        • 1 month ago
          Anonymous

          rofl nice necro post. do you have a bot that alerts you or something?

          • 1 month ago
            Anonymous

            >necro
            this isn't a forum you moron, it's an imageboard

          • 1 month ago
            Anonymous

            >necro
            You use that term on a forum when someone revives a thread that's been dead for years. Not a new post after a few hours on an imageboard.

            >same anon

          • 1 month ago
            Anonymous

            >necro
            You use that term on a forum when someone revives a thread that's been dead for years. Not a new post after a few hours on an imageboard.

    • 1 month ago
      Anonymous

      Need to go under 100 again

  44. 1 month ago
    Anonymous

    >update mpv
    >speed up video
    >terminal gives video progress as [current normal speed time] / [total video time when sped up]
    whose idea was this shit

    • 1 month ago
      Anonymous

      im pretty sure mpv has done this for years

    • 1 month ago
      Anonymous

      im pretty sure mpv has done this for years

      https://github.com/mpv-player/mpv/pull/13326#issue-2092223892

      • 1 month ago
        Anonymous

        I think I might agree with kasper on this one.
        1x speed:
        00:01:00 / 00:10:00
        2x speed:
        00:01:00 / 00:05:00
        proposed 2x speed:
        00:00:30 / 00:05:00

        • 1 month ago
          Anonymous

          fo sho

  45. 1 month ago
    Anonymous

    https://github.com/mpv-player/mpv/issues/11739#issuecomment-1902225805

    WINDOWS WON
    D3D11 >>>>> VULSHIT
    D3D11VA SUPREMACY

  46. 1 month ago
    Anonymous

    mpv ffmpeg yt-dlp

  47. 1 month ago
    Anonymous

    meme shaders peaked at ArtCNN. Has anyone compared it to NGU yet?

    • 1 month ago
      Anonymous

      Go and look at benchmarks, loser.

      • 1 month ago
        Anonymous

        Artoriuz's blog used to have metrics for NGU but he removed those years ago.

    • 1 month ago
      Anonymous

      >trained on the Manga109 dataset
      yep, it's useless

      • 1 month ago
        Anonymous

        That was my thought as well :c

      • 1 month ago
        Anonymous

        name a better public dataset of anime content that's not legally dubious.

        • 1 month ago
          Anonymous

          madshi doesn't focus only on anime and this is his success, providing tools to watch real content and not kids cartoons.

          • 1 month ago
            Anonymous

            Couldn't find a better dataset?

          • 1 month ago
            Anonymous

            You don't even need a large dataset. Madshi trained on a few dozen images, including a mix of patterns, test images, movie frames and lineart. It's really not hard to create your own dataset, and if you use what is already available on the internet, you won't produce anything new or interesting.

      • 1 month ago
        Anonymous

        what should be trained on instead

      • 1 month ago
        Anonymous

        better use sharp part of this https://seungjunnah.github.io/Datasets/reds.html

  48. 1 month ago
    Anonymous

    damn i used to like mpv but after learning that haasn is a Z operative, ill have to switch back to MPC-HC

  49. 1 month ago
    Anonymous
  50. 1 month ago
    Anonymous

    I understand that common video file formats like the ones inside webm and mp4 basically store the video data as high quality keyframes every few seconds and store the other frames as delta from the last keyframe to save file size

    to losslessly cut out a part of video without re-encoding you have to start the cut exactly at a keyframe since other frames depend on last keyframe. Makes sense. No wait, it doesn't. Let's say I have 9 second video with keyframes at 0s, 3s, 6s and want to losslessly cut out a part from 4s to 7s. But lossless cut limits me to 3s-7s cut right? But why, why can't the video contain raw data from 3 to 7 but the container metadata would mark 4s as the start time so I end up with a video that has 4s of raw data but when you play it it shows 3?

  51. 1 month ago
    Anonymous

    nice bug tracker

    • 1 month ago
      Anonymous

      mpv is probably the only video player that can get direct feedback from a community of mentally ill coonsoomers yet it's still somehow buggier than VLC

      how does this happen?

      • 1 month ago
        Anonymous

        >how does this happen?
        overwhelmingly moronic devs
        autism
        lack of leadership

        • 1 month ago
          Anonymous

          >lack of leadership
          blame jeeb (or as i like to call him, the multimedia anti-christ)

          • 1 month ago
            Anonymous

            >blame jeeb
            don't call jeeb out from the woods. He will destroy us all if he gets mad. He holds all the keys, need to be careful.

      • 1 month ago
        Anonymous

        >buggier than VLC
        can you seek in VLC without it glitching yet

        • 1 month ago
          Anonymous

          you dont NEED to seek backwards... always be forward-looking

  52. 1 month ago
    Anonymous

    >VLC
    I dunno bros. Just browsing through the issues here, this looks 1000x worse than mpv bugs.
    https://code.videolan.org/videolan/vlc/-/issues

    • 1 month ago
      Anonymous

      mpc-be is the least buggy media player according to your metric
      https://github.com/Aleksoid1978/MPC-BE/issues

      only SEVEN bugs, how do mpvoids cope?

      • 1 month ago
        Anonymous

        I mean yes I'd expect MPC-HC and friends to be less buggy. Not really sure where you're going with this.

      • 1 month ago
        Anonymous

        https://github.com/Aleksoid1978/MPC-BE/issues/359

        here's your typical MPC-BE bug report and dev "response"
        >hey your video renderer has garbage tonemapping
        >uh just turn up the brightness
        >fixed and closed

  53. 1 month ago
    Anonymous

    igv seething ITT

  54. 1 month ago
    Anonymous
    • 1 month ago
      Anonymous

      I could work with her.

  55. 1 month ago
    Anonymous

    https://github.com/mpv-player/mpv/issues/12752
    is this doable yet

  56. 1 month ago
    Anonymous

    /R2ITA/i

    thank me later anons

    • 1 month ago
      Anonymous

      see picrel
      thank me later

      • 1 month ago
        Anonymous

        hahah good one dude

  57. 1 month ago
    Anonymous

    https://github.com/mpv-player/mpv/commit/a39f9b6181ee54732050d64540703ca2e2c90a8d
    bugmanguy's latest commit. what are the implications?

  58. 1 month ago
    Anonymous

    https://github.com/mpv-player/mpv/pull/13331

    i don't know if i would necessarily say that bwdif is better than yadif, it just produces sharper results

Comments are closed.