/mpv/ - the Sup Forumsreatest media player

Install mpv:
mpv.io/installation/

Wiki:
github.com/mpv-player/mpv/wiki

Manual:
mpv.io/manual/master/

User Scripts (including opengl shaders):
github.com/mpv-player/mpv/wiki/User-Scripts

input.conf:
github.com/mpv-player/mpv/blob/master/etc/input.conf

High quality video playback profile:
profile=opengl-hq
video-sync=display-resample
interpolation
tscale=oversample

Other urls found in this thread:

mpv.srsfckn.biz/
bitbucket.org/rorgoroth/mingw-w64-cmake/downloads/
sourceforge.net/projects/mpv-player-windows/files/
sourceforge.net/projects/mpv-player-windows/files/experimental/
jenkins.maeyanie.com/job/mpv/
youtube.com/watch?v=04F4xlWSFh0
github.com/shinchiro/mpv-winbuild-cmake/blob/master/packages/ffmpeg.cmake#L21
github.com/igv/FSRCNN-TensorFlow/tree/fsrcnnex
github.com/mpv-player/mpv/pull/5033
patchwork.ffmpeg.org/patch/5686/
github.com/mpv-player/mpv/blob/master/DOCS/man/lua.rst
superuser.com/questions/42039/change-windows-sound-volume-via-the-command-line
github.com/mpv-player/mpv/blob/master/DOCS/compile-windows.md
github.com/igv/FSRCNN-TensorFlow/releases
twitter.com/NSFWRedditGif

So, is mpv kill now?

what haasn is busy with right now? still working on libplacebo? Cant wait for new interpolation technique!

its all over

wat

B-but I'm switching to GNU.
There's no MPC-BE on GNU.

Windows Builds:
Stable Portable: mpv.srsfckn.biz/
Stable Installer: bitbucket.org/rorgoroth/mingw-w64-cmake/downloads/
Dev Snapshots: sourceforge.net/projects/mpv-player-windows/files/
WIP d3d11/VulkanR2: sourceforge.net/projects/mpv-player-windows/files/experimental/
WIP d3d11/VulkanR2 Installer: bitbucket.org/rorgoroth/mingw-w64-cmake/downloads/
Vapoursynth: jenkins.maeyanie.com/job/mpv/

Config for D3D11/Vulkan:
gpu-api=d3d11 |..or..| gpu-api=vulkan
hwdec=auto-copy
profile=gpu-hq
vd-lavc-dr=yes

THREAD THEME: youtube.com/watch?v=04F4xlWSFh0

BIG DICK WM4
>you will never be this alpha

libavfilters dev did exactly the same long ago

So, how is this mess gonna affect mpv on Arch?

Error parsing option gpu-api (option not found)
Error parsing option profile (option parameter could not be parsed)

ffmpeg: Cannot load libnvcuvid.so.1
ffmpeg/video: h264_cuvid: Failed loading nvcuvid.
vd: Could not open codec.

If you're using mpv-build just change the repo in the pkgbuild and makepkg

What if you just use pacman?

Nothing will change because the people who need to deal with this are the Arch package maintainers for ffmpeg/mpv.

changing this line will be enough?
github.com/shinchiro/mpv-winbuild-cmake/blob/master/packages/ffmpeg.cmake#L21

Yes.

You will probably wait 5 generations of GPU before using this new interpolation in realtime.
So you will be probably dead. ;)

Im not waiting for THAT kind of interpolation. haasn said he will try to make something similar but much less demanding!

To get your money on patreon...
Don't trust him too much.

Revert "build: require our own ffmpeg repo"

Finally! You can't trust those ffmpeg retards.

>tumblr

>github.com/igv/FSRCNN-TensorFlow/tree/fsrcnnex
Is it supposed to be faster than fsrcnn?

Im interested too.

>it's a wm4 going ballistic episode again

ok. thanks!

Is it just me or is wm4 massively over-reacting about all this?

github.com/mpv-player/mpv/pull/5033
^ Literally a 4 line patch fixes this in mpv and brings it up to speed with a fix made in FFmpeg.

Apply that? NO! Fork FFmpeg!
How long can he keep this up, how often will he sync changes in to his fork, is he going to pull all or fucking look at each one first?

>muh API breakage
Who gives a fuck, how often does this happen? Not often so get over it lmao.

Wm4 is getting old and doing retarded things.

>muh API breakage
Well that's understable. If you use some kind of library that's supposed to work in a way and someone goes "Oh let's change this 'cause I just feel like it" you won't be happy.

Still no one I guess?

The FFmpeg guys are already working on a patch to make this harmless and not break shit: patchwork.ffmpeg.org/patch/5686/
Posted today.

Also posted in FFmpeg mailing list is that 0 as EOF was never documented as public API in the first place:
>I do not consider this a public API change because 0 was never
documented as a valid return value for a read_packet callback, while
AVERROR_EOF has been around for a long time.


The change to clients for this is so small there are really no excuses not to update for it.

I see, thanks for the explanation.
Then wm4 is going full retard ¯\_(ツ)_/¯

>Error parsing option gpu-api (option not found)
>Error parsing option profile (option parameter could not be parsed)
>ffmpeg: Cannot load libnvcuvid.so.1
>ffmpeg/video: h264_cuvid: Failed loading nvcuvid.
>vd: Could not open codec.
PLS RESPOND WHY I GET THESE ERRORS?
I USE ARCH LINUX

Use git not the official repo!

Why is wm4 such a retard?

He is getting old...

How old is he?

Fuck you.

Older than you.

just compiled a build with the new FFmpeg.

Feels faster!

anyone got the pic comparing anime on several players? I think it was MPV & VLC

Go to bed wm4.

Faster what?

>Initial ra_d3d11 implementation PR
FUCK YEAH

>before vulkan rev2
but why

vulkan is a joke right now in terms of stability compared with d3d11

and d3d11 is a joke right now in terms of performance compared to vulkan and ogl

not really. my only problem is dropping frames in windowed mode.

what the fuck lads

what's the point of speed if you can't even resize the window without freezing the video

how fucking often do you manually resize a window? seriously you're doing something horribly wrong if this is actually an issue for you.

The mpv devs don't read this. You should make a github issue. The best way to handle this is probably to allow passing --seed which will call srandom() on startup

poe's law m8

That was a shit movie

What movie?

Hopefully there is wm4 to fix all haasn's and rossy's mess.
Mpv has never been that unstable.
I begin to love the madshi's and nev's way aka optimized for only one platform because it's already a fucking mess.

eat shit poozoor

Werks great for me on Xubuntu and Win 10!

>no NGU
Into the trash it goes.

He's just making things messier. What hell happened to him?

fate/GO : First Order

not really. Just need to revert ffmpeg's commit

Just like everything else TM up until now.

Its a special. Not movie.

Is there a reason mpv can't do -c:v copy like ffmpeg? Other output codecs and containers are supported - so why not the stream copy?

First for MPLayer

It was probably overlooked whenever encoding mode was implemented

Just us ffmpeg...

ok thats it. im going back to windows media player.

NGU - next gen upscaler
There is 2 next gen upscalers in mpv already - RAVU and FSRCNN
Eat shit fag.

FSRCNN literally is NGU

FYI if you're using Gentoo with the 9999 ebuilds for mpv and ffmpeg, you can create a local repo and just replace the EGIT_REPO with the mpv fork in ffmpeg and it builds.

man,dont start yet another ffmpeg-forked

best settings for linux/nvidia gt 740?
hwdec=gpu
profile=gpu-hq
what else?

i have ytdl-format set to a youtube-only format so when i try to play videos from another site i get the 'requested format not available' error. how can i get mpv to ignore the setting if the site isn't youtube? why isn't that he default behavior?

Add a /best on the end.
ytdl-format=bestvideo[height

I don't know why it's not the default, but that's youtube-dl's thing not mpv's.

not him but how to increase cache when playing?

my bigcache profile
[bigcache]
cache=yes
cache-default=524288 # size in KB
cache-backbuffer=524288 # size in KB
cache-initial=1024 # start playback when your cache is filled up with x kB
cache-seek-min=1000
cache-secs=120 # how many seconds of audio/video to prefetch if the cache is active

It's likely not the default because when you're using youtube-dl to save videos it wants to alert you that your requested format wasn't enabled. With the /best option at least you have to tell ytdl it's ok.

so if I copy your profile, is this right?
mpv --profile=gpu-hq,bigcache video.mkv

> is this right?
*to use it

That will work. I personally like to specify most of it in my mpv.conf. I have one profile I call manually with the "open with" extension for Firefox for youtube videos and such.

oh thanks

Every fucking time when I come back from a movie I gotta remember to lower my volume like 50% to listen to music, cause all movies are just so fucking QUIET in comparison. Is there any feature in MPV that could help this shit? Last time I looked I could set audio of all videos to be something like 120% but it would distort the shit out of it...

You could use a different profile for music and videos.

Or you could set the volume based on file extension like
[extension.mp3]
volume=50

[extension.ogg]
profile=extension.mp3
Just change the extension to match yours

Oh, yeah, definitely, but I mean just the volume of my computer in general. I actually use foobar for music.

A lua script could maybe be written to adjust your windows/linux volume on mpv start and exit but I'm unfamiliar with them.

Actually, yeah, that's a good idea! I'll look into it, thanks.

github.com/mpv-player/mpv/blob/master/DOCS/man/lua.rst

I was thinking you could use utils.subprocess(t) to run a batch or bash script that adjusts your volume on start-file event and stop event.

I'd tinker with it myself if I was familiar with lua.

This looks like a good tool for adjusting sound via windows. superuser.com/questions/42039/change-windows-sound-volume-via-the-command-line

github.com/mpv-player/mpv/blob/master/DOCS/compile-windows.md
MSYS2 compilation on Windows 10 doesn't work anymore because of the ffmpeg commits

well, you need to patch ffmpeg in order to make it works again

Can anything in mpv compete with madvr NGU?

Not trolling, wanna know if I should switch back

FSRCNN

How can I even tell if that's enabled? It doesn't show up in stats like many other shaders I've tried. I don't get an error message about file not found in terminal

Just downloaded the latest shinchiro experimental build (went up less than 10min ago)

So compute RAVU is finally working with d3d11?

need some config guide

github.com/igv/FSRCNN-TensorFlow/releases

is it this?

should I disable scale=ewa_lanczossharp
cscale=ewa_lanczossharp
dscale=ewa_lanczossharp
tscale=oversample when using FSRCNN?

Probably, except tscale