/mpvg/ Mpv general

Maki play videos.
Github: github.com/mpv-player/mpv
Manual: mpv.io/manual/master/'
wiki.archlinux.org/index.php/mpv

Rubberband script user made.
Script made by an user: change youtube quality on the fly.
webm scripts by user

Attached: scrot_2018-09-05-19:53:44_1280x720.png (1280x720, 1.17M)

Other urls found in this thread:

github.com/mpv-player/mpv/wiki/User-Scripts
github.com/wm4/mpv_backup
github.com/ekisu/mpv-webm/
github.com/vayan/external-video
addons.mozilla.org/en-US/firefox/addon/play-with/
twitter.com/AnonBabble

My wife

>sudo zypper install mpv
>mpv ~/muh\ anime/yuri\ shit.aac
>doesn't work
wut do

Attached: 1529266664427.png (518x391, 200K)

>.aac
Is that an audio file?

what do the verbose logs say?

You should just link all user scripts and shaders

github.com/mpv-player/mpv/wiki/User-Scripts

I'll do next time, sorry.

And the best basic config for Linux
gpu-context=x11egl
profile=gpu-hq
sigmoid-upscaling=no
keep-open
x11-bypass-compositor=yes
ad-lavc-downmix=no
audio-channels=stereo
slang=en,eng
ytdl-format=bestvideo[height

wrong link, this is the real mpv repository
github.com/wm4/mpv_backup

Are these scripts really better than webm.lua?

github.com/ekisu/mpv-webm/

Hmmmmmm
I think they're just a modified version of webm.lua to get webm with or without sound and get them as close to 3-4mb as possible.

nevermind fixed it

webm.lua does both of those things tho...

>webm without sound
mute the video before encoding. webm.lua recognizes this and gives you a webm without the audio stream.
>webm exactly at the file limit
set target filesize and strict filesize constraint.

>backup

Awful

Not an argument, your post is awful.

Madvr is higher quality

MPV is higher quality

Why disable sigmoid upscaling? Why not explicitly list d3d11va for hwdec? Why use stereo? Why do you have video height twice for ytdl?

Prove it nerd

Prove it nerd

>sigmoid upscaling
meme
>Why not explicitly list d3d11va for hwdec?
i agree, hwdec=d3d11va would be better
Why use stereo?
Without
ad-lavc-downmix=no
audio-channels=stereo

audio is too quiet when watching video with more channels
>Why do you have video height twice for ytdl?
because 1st is for youtube and 2nd is for all other streaming sites

How is sigmoid a "meme"? That's a retarded and lazy explanation by the way.
If you need stereo because the audio is too quiet than you must have shit speakers and/or watch shit encodes. Sound is fine on my 5.1
I have ytdl-format=bestvideo[height

Is it possible to displays all the youtube stats (uploader/upload date/views/likes/description) on the mpv window?

>That's a retarded and lazy explanation by the way.
ok, if you are so smart prove then that sigmoid upscaling is not a meme.
>If you need stereo because the audio is too quiet than you must have shit speakers and/or watch shit encodes. Sound is fine on my 5.1
It has nothing to do with speakers quality, but downmixing quality, if you don't downmix then set it to 5.1 or auto, as I said it's a basic/startimg config not everybody has 5.1 speaker layout.
>I thought this was for all sites.
best doesn't always mean 1080, on some sites best can mean 1440 or 2160.

>Sound is fine on my 5.1
I believe he's saying that if you only have stereo speakers then 5.1 audio will be too quite unless you downmix it to stereo.

So you really dont know how to explain why sigmoid isnt needed using adult words? I'm genuinely curious.

it's a placebo meme, in other words, it doesn't fucking do anything.

Has anyone figured out a way to start caching the next youtube video on the playlist before the current video ends for a perfectly smooth transition yet?

actually it does, it adds a lot of ringing artifacts.

Why would the devs include it in profile=gpu-hq? Are they memeing us?

Attached: su.jpg (1620x197, 54K)

I'd love this too.

I don't know, but I know for sure that that's a misleading description.

>When upscaling, use a sigmoidal color transform to force emphasizing ringing artifacts.
ftfy

what's a good font for subs/osd?

Papyrus.

How hard is it to learn Lua?

Lua is easy. Kind of shitty in a lot of ways, but easy.

How do I make my MPV use pseudo GUI?

Warning: mpv: this formula has no --without-libcaca option so it will be ignored

how do I reinstall without libcaca?

>tfw sad because most of my screenshots are shit and i want to replace it with better quality screenshots

i still have the video and hundreds of screenshots with the timeframe included in the timeframe, how much would it cost someone to make a script that would play the video and take a screenshot based on those timeframes?

>timeframe included in the timeframe
meant to type "timeframe included in the filename"

im using a 4:3 monitor but i want to watch 16:9 subbed anime. how can i move the subtitles to the black black bar on the bottom?

Looks like sigmoid-upscaling is broken in mpv, it works correctly on rgb images, but not on YUV images/videos. In madvr it works correctly on both YUV and RGB.

Which one?

mpvizzle plays videolicious

You could pretty easily do this with ffmpeg.
ffmpeg -ss -i input.file -vframes 1 output.png
Makes and saves the frame at the specified time. You'd only have to figure out how to parse the filename.

Holy shit, mpv doesn't even convert YUV to RGB correctly, this is unacceptable. Meme player.

Kill yourself spic subhuman.

sub-use-margins=no

Is this an older version of mpv? It's from a movie called the innkeepers (2011).

Attached: mpv-shot0003.jpg (1920x800, 142K)

Rude

Attached: Cowboy Bebop 04 BDRip [Lat_Cast_Jap + Sub Esp] (1440 x 1080_x264) por camus10x (00:24:32.471) 0008.p (1440x1080, 1.6M)

I also use downmixing, so i'm interested in ad-lavc-downmix
Why do you disable it?

but ffmpeg isn't mpv

or are you saying that ffmpeg would open the video with mpv

Perhaps you should clarify. Are your screenshots shit because you saved them as low quality JPGs or because they don't show your current filters/shaders?
My ffmpeg suggestion was aimed at the first scenario.

I don't remember, seems unnecessary.

>Are your screenshots shit because you saved them as low quality JPGs
yes, why else would they be shit

i just want to watch a video with mpv, and let it automatically take screenshots accordingly to the timeframe of my screenshots in a folder

I use Source Sans Pro Semi bold.

>yes, why else would they be shit
Some people are pretty autistic about their shaders and declare anything other than their placebo settings as shit.
If it's just about the image quality, then why not use ffmpeg? Loop through all of your original screenshots, parse each filename to get the timeframe and let ffmpeg extract the new screenshot as png, tiff, bmp or whatever lossless compressed image type you want. It doesn't get easier than this and only takes a few seconds for 100 screenshots.

what's the difference between mpv's png and ffpmeg's png?

these are my screenshot settings on mpv, I don't like png because the file size is too big (3-6 MB) and I find these settings to be nearly as good as PNG
screenshot-format=jpg
screenshot-jpeg-quality=100
screenshot-tag-colorspace=yes
screenshot-high-bit-depth=yes
screenshot-jpeg-source-chroma=yes

Attached: 1515964061681.jpg (1920x1080, 1.74M)

What is /mpvg/ prefered add on to send youtube videos from firefox to mpv?

Play With MPV? Is there something better?

Is there any hentai with that redhead smug bitch?

yes

Attached: a13014ab32fd3e2258a218d006a54afc04027eb545101b7e5a0ed06d9cf25689.jpg (850x1200, 311K)

github.com/vayan/external-video
for linux/firefox

any youtube video you click on opens with mpv automatically

I'm on waterfox with openwith, the pre 7.0 version. I saw that the newer versions are a bit hacky. You can make multiple entries with various arguments for audio only, resolutions and cache, etc.

...

addons.mozilla.org/en-US/firefox/addon/play-with/

The difference is that you can easily automate ffmpeg, while mpv's screenshot function can't be executed via a script.
ffmpeg can also extract JPGs, although you have to set the q:v parameter, because the default quality is shit. Alternatively you can extract lossless screenshot and then use a utility like Image-/GraphicsMagick to batch convert them to whatever lossy format you like.

That's what I kinda wanna do with mpv. Either way, I have ffmpeg and it's ready to go. How do use your command line to automate this for all of my screenshots?

Pic related, to the left, we have my shitty screenshots with the timeframe as the filename, and in the middle is the video itself, and I want new screenshots to be dumped to the right

Attached: 1533858934107.png (1920x1046, 392K)

I fucked up, I forgot to open the folder to the left. Here it is.

Hours_Minutes_Seconds_Milliseconds.jpg

Attached: 1513831127318.png (1920x1049, 245K)

That's where I can't help you further. I only know bash.
Ask in

isn't this possible with mpv though? i'm pretty sure you can play/stop a video in cmd, so shouldn't it be possible to fast forward to a certain time and take a screenshot?

The problem is you can only take screenshots with mpv manually. So you can either write a semi-automatic solution, where the script opens mpv in a paused state at the given timeframe and you manually take the screenshot and close the program every time or you write an automatic makeshift solution, where you take a screenshot of the window itself and instead of the footage.

is it possible to use ffmpeg with powershell?

Yeah, that should certainly be possible.

??? It works for me!

1) Real NGU Sharp Medium/High alternative when? FSRCNNX is shitty!
2) New madshi's secret algorithm alternative when?
3) Acceptable HDR tone-mapping when?
4) Is mpv dead without haasn?
5) Is linux still relevant for the new generation on desktops?

unix babby here

I installed mpg with brew it didn't show up then I installed it with cask and now it shows up

is it only installed once or do I have two different installations now

all this time doing OPs and you're still only at the third episode?

New FSRCNNX with CSFM super resolution BTFO NGU Sharp.

what does async screenshot do

'screenshot' is an input mode command so you should be able to use it from a userscript, or from the JSON IPC api. It can be automated, but either solution is more involved than simply using ffmpeg.

>Setting the async flag will make encoding and writing the actual image file asynchronous in most cases. (each-frame mode ignores this flag currently.) Requesting async screenshots too early or too often could lead to the same filenames being chosen, and overwriting each others in undefined order.

What is asynchronous? srs question

"not synchronous"

In practical terms this means that mpv won't wait for the screenshot to finish (that is to say, it won't synchronize) before the screenshot command returns. I haven't checked the source but this is PROBABLY implemented by having the screenshot work in a separate thread while the execution of mpv continues.

I tried it with FFmpeg and it prints out my screenshots with the filename "output.png" which is lame, and it takes a lot of time to use the command for each screenshot

I don't mind paying for a script that would work with mpv (if it's possible to get it to work with mpv)

or as the documentation explains it:

> async - Allow asynchronous execution (if possible). Note that only a few commands will support this (usually this is explicitly documented). Some commands are asynchronous by default (or rather, their effects might manifest after completion of the command). The semantics of this flag might change in the future. Set it only if you don't rely on the effects of this command being fully realized when it returns.

key part:
> effects might manifest after completion of the command
The `async screenshot` command can/will return before the screenshotting operation has actually completed.

is it a bad thing or good thing to keep "async screenshot" in input.conf?

I doubt it will hurt anything unless you are mashing the screenshot key a lot and you run into that "one screenshot overwrites the other" race condition (which would happen if two parallel screenshot operations check for an available screenshot file name at more or less the same time and neither sees that the other already claimed that name, so both screenshots get written to the same file and whichever one came slightly later overwrites the first).

On the plus side it probably means your video won't stutter for a moment when you take a screenshot. I assume anyway.

It's up to you, I don't think the upsides or downsides are very significant either way.

won't this solve that override problem anyways

screenshot-template='D:\Downloads\Screencaps\%F\%F (%P) %n'

why is mpv slow to open youtube links? can it be fixed?

how slow? it takes 3 seconds for the video to start playing for me.

I'll see what I can do, but absolutely no promises. And I don't want your money for it.

I need a clarification on your filenames, they seem to have the format:
>hh_mm_ss.ttt
where hh is hours, mm is minutes, ss is seconds, and ttt is thousands of a second. Is that correct?

Also, if I do make this script then for simplicity I'm going to require you to place the video file into the same directory as your screenshots before running the script. That's so that you won't have to enter any sort of configuration about where to find the screenshots, the script will just look in the same directory.

Also don't count on this being done before this evening, I've got other things I'm working on too.

>filename has milliseconds
that would probably solve the override problem, unless you screenshot twice during the same millisecond, which seems unlikely.

The output.png in the command was simply an example. I took a look at how you loop through files with cmd. You would put this code in a .bat file, which should be located in the same directory as your old screenshots. Executing it would take the screenshots for one video. You could extend it to do all videos in one go, but this is the basic concept.

@echo off

FOR %%A IN (*.jpg) DO (
< convert hh_mm_ss_ms to hh:mm:ss.ms for ffmpeg to understand it >
ffmpeg -loglevel panic -ss hh:mm:ss.ms -i path\to\input.file -vframes 1 "destination\for\new\screenshots\%%~nA.png"
echo "Current screenshot: %%~nA.png"
)

sometimes even more then 20 seconds. could it be the config? I haven't update it in a while, but it should be fine.

I finally managed to get Vulkan working on Fedora.

Let me know if this solution works for you: If so, I won't bother with the lua script. This bat file solution is the 'proper' solution.

I'd update youtube-dl. It shouldn't take that long however, and I believe it is mostly due to a slow connection. I do notice it takes longer at seemingly random times.