...
/mpv/ - the 2018 media player
Other urls found in this thread:
github.com
rorgoroth.bitbucket.io
github.com
madvr.com
cloud.blender.org
youtube.com
youtube.com
github.com
ffmpeg.org
twitter.com
Hanna, please save mpv again!
Why is C so over complicated?
github.com
And why there is no docstrings in the mpv code source?
Because mpv is not ffmpeg. And because mpv is a shitfest.
>mpv is best play-
It's nice that you support legacy software, but there is no active development.
All I want from my mpv is to give me proper filenames and save as png when I take screenshots.
I assume you mean doxygen-style comments, since docstrings are a Python thing.
/**
* It's because mpv isn't fucking garbage. Just write normal comments lmao.
*
* @returns All Doxygen gives you is the godawful Doxygen web interface, which
* is worse than just reading the header file.
*/
That’s a fairly meme C interface.
Would it be possible to implement NND-like scrolling text with a lua script or would the perfs be absolute garbage?
It's possible, read the manual.
screenshot-format=png
screenshot-template="%f_snapshot_%P_%#02n"
Why does my i3wm get screen tearing after using mpv?
dwm.
[1080p]
profile-desc=cond:get('height',0) < 2160 and get('height',0) >= 720
[720p]
profile-desc=cond:get('height',0) < 720
why is 720p still being trigger at the beginning, it defeats the purpose of it
This was explained yesterday, you need to use a default value (second arg to get) bigger than the right hand side of the comparison.
720p is "less than 2160 wide and less or equal to 720 high"
you want to remove the "=" from near the end
oh, also noticed you have 2160 being checked as a height, not width
>2018
>not in 1970 anymore
>normal comments
How do I get .srt in a folder to work. say i have /videos/subs.
right now i have sub-auto=fuzzy
but it doesn't work :(
Why not this?
rorgoroth.bitbucket.io
Pst. New thumbnail release github.com
Like the release and commit tell, the thumbnail feel is now greatly improved.
Next version will likely get rid of the manual worker scripts and spawn them automatically, which is also a step or two closer to the progressbar.lua API dealio.
There's your no-progress.
why not just drag the bar?
thumbnails are beneficial for low bandwidth links, as they load faster than the actual video, but this isn't the case for local videos
So.. still no GUI?
Why not using mpc-be + madVR? D3D11?
Dunno. This is what people wanted and it was a neat challenge. This does do network paths as well, anything seekable, but like I've said, it's not anywhere near the efficiency it would or should be with a real built-in implementation. Why not just do profile=gpu-hq and enjoy video instead of going on and on about tuning it?
Link me to the thread
It does not matter, the profile still triggers.
What's this weird shit you're watching?
Damn why did he develop a custom resolution creator if custom resolutions are not well supported by current hardware?
madvr.com
It's not the hardware which is the problem but the drivers. Might be improved in future drivers. Or not. But in theory it's a good solution. If it works.
not him, but pretty sure it's Sintel, one of the Blender Open Movies: cloud.blender.org
Why? Is it only because audio resampling is a resource power waste?
Don't wanna shit on blender because I like it, but that particular movie looks pretty bad.
Audio resampling costs a bit of audio quality. And it doesn't work if you want to use bitstreaming. Bitstreaming is necessary if you have an Atmos or DTS:X setup because current ffmpeg decoders can't decode the additional Atmos or DTS:X channels.
>Sintel.2010.1080p.mkv in two places
>pretty sure it's Sintel
missed that
Same answer as always. Because DirectShow is garbage and madVR is closed-source and single platform. Most people in these threads couldn't use madVR if they wanted to, because they have operating systems that madshi is too stupid to program for.
I suppose it's also kind of funny that madVR hasn't upgraded from a sixteen year old graphics API to a ten year old graphics API, but it's not the point.
help anyone?
Is resampling noticeable for humans?
Are those additional atmos/dts:x channels useful for headphones users?
sub-file-paths=subs
sub-auto=fuzzy
Can MPV downmix 5.1 or 7.1 into 2.0 properly yet? Any time i listen to anything that is 5.1 or 7.1 i get a terrible ballance on the chanels when played on my stereo headphones. Can that be properly adjusted yet?
[2160p+]
profile-desc=cond:get('height', -math.huge) >= 2160
interpolation=no
profile=nogpu-hq
[1440p+]
profile-desc=cond:get('height', math.huge) < 2160 and get('height',0) >= 1440
[1080p+]
profile-desc=cond:get('height', math.huge) < 1440 and get('height',0) > 720
[720p-]
profile-desc=cond:get('height', math.huge)
thank you buddy
Is there a way to make discord 'game' status capture the video file name in mpv? i tried looking at the options but no avail.
>profile=nogpu-hq
???
I suppose with good headphones you would have a better chance of noticing any potential quality degradation than with average quality speakers. That said, most people don't even hear a difference between MP3 and lossless, so probably most people wouldn't notice the fairly small resampling quality loss.
That said, even if it's only a very small improvement, why not avoid audio resampling if you can, while still getting perfectly smooth playback? And that's what custom resolutions could achieve. At least in theory.
It's never perfect with custom resolutions because everything can fuck it.
Is the madshi's customres tool only working with madvr or does it work with mpv too?
Works perfectly for me, but I know I'm lucky, it works for some users and completely fails for others.
Should work for mpv, too (if it works at all)!! But only on Windows, obviously.
[nogpu-hq]
scale=bilinear
cscale=bilinear
dscale=bilinear
dither-depth=no
correct-downscaling=no
sigmoid-upscaling=no
deband=no
Try this :
audio-channels=stereo
audio-normalize-downmix=yes
What kind of toaster are you running? I can run 4k with SVP without dropping frames using mpv.
That seems to work to normalize the audio across all channels, yes. Is there a way i can individually specify a percentage or dB value for each chanel? Some sources have terrible volume balance, and normalization just averages them all out, which is undesirable if two channels are loud as fuck and the rest is rather quiet. I would especially like the channel that has the spoken lines in it to be slightly above the SFX on other channels.
3570k+750ti
You can always donate me a GTX 1080 but actually just give it to rossy since he deserves it more.
My motherboards sound card has a driver that allows precisely for this. If 5.1 or 7.1 is played over stereo, it gives you a channel mixer that lets you adjust as you see fit.
I think you can do this with some ffmpeg filters but I don't precisely how to do it
Still much more stable than any of the meme garbage like mpv.
In mpv, Running this
youtube.com
Would play the playlist from the first video to the last, however if I play any other video of the playlist
like this youtube.com
it would play that video only and exit when finished.
What I want is to be able to play the next video in the playlist when using this URL youtube.com
Is it possible? mpv doesn't seem to know how to read the &index=3 part of the URL.
I'm trying to save bandwidth by playing lower resolution, however when I use this ytdl-format=bestvideo[height
bestvideo sounds like it's grabbing the highest resolution video it can find
and yet it, with madVR, is still better than mpv
github.com
Stumbled across this the other day. Doesn't really look usable yet, but the concept is cool, and the dev seems to be actively working on it.
I like to use Sintel and BBB for demoing video stuff, because they're used to demo video stuff. What with being "free". A user of a FOSS video player should be at least aware of the Blender movies.
I knew BBB but not this one.
ffmpeg.org
Is this what you look for?
>GPU-accelerated motion interpolation
Madshi or hanna? Who will be first?
Hopefully neither. I don't want to live in a world where motion interpolation is popular. It's bad enough that smart TVs do it.
Dude shut the fuck up, it's not gonna go faster if you sperg about it everyday
>wanting shitty judder on your animu
Has anyone tried to make discord 'game' status capture the video file name in mpv?
Frame interpolation fixes that. The problem with motion interpolation is that the interpolated frames have artifacts. It's especially bad with 2D animation.
>frame interpolation
>motion interpolation
It seems like I've been an idiot all this time by using these interchangeably.
Now that's what I wanna see. I tried the one on the mpv website and then another one elsewhere and they both didn't work. This one does.
you probably forgot to use quotation marks
i just put it together then, referencing the manual
i never bothered to set a template before, since i only ever use screenshots once and not keep them
It worked with user-shaders and neural based upscalers like ravu and fsrcnn. :)
>xf86videoati
>mpv27
>hwdec
>vdpau
>no tearing in fs
when did this sorcery happen
mesa?
mesa/radeonsi improved a /lot/ through 2017
did you finally update to mesa 17.x?
i has it on windows with cccp ^_^
this player is pretty good, but why do we need constant threads about it?
Because we are on Sup Forums and it's foss!
And 2017 was pretty AWESOME and 2018 will be even more fucking AWESOME.
Yes, as you can see said threads are quite active.
why do we need threads for anything?
just filter it if you don't want to see it
i guess so
since 2017-02-13 apparently
very configurable
where else would you ask?
Because it destroyed the terrible madshi's reign and freed our mind.
Even stable mpv build is more updated than this crap.
I saw the madshi's cave today:
Eulenacker 4 - 22175 Hamburg
Damn it's windy as fuck this morning.
So we can discuss mpv and other players?
i bet you're using ubuntu
>Next version will likely get rid of the manual worker scripts and spawn them automatically
Are you gonna write the script file(s) at runtime? What's the benefit, just simpler installation?
guys please make it easier for encoders to be able to add sub/audio delay in the filenames. This way we don't have to mux/use a subtitle editor every time we want to adjust the times of an audio/subtitle.
I mean, I don't know anything about lua but how hard is it to check the filenames and if there's something like DELAY -1000 at the end then call a function that does the same thing you do with the inpt keys.
Didn't sombody ask for that a while ago? But yeah it should be pretty simple to do.
Relying on magic strings in the filename is error prone and stupid. If you think it's hard to remux subtitles, you're not an "encoder."
Also that's pretty hacky, if you're an encoder you should already have scripts to change to demux, change the delay and remux your subtitles.
how about make it easier for normies that want to sync subs fast and don't have subtitle editing programs?
so you've never demuxed an old dvd and found out it's out of sync end want to sync the audio but you don't want to have to mux it every single time just to check when it works?
I mean, I've encoded and that has happened to me.So I guess I'm not in the cool encoders kid club you are.
What's the keystroke for switching through subtitle tracks?
Nornies should just use the keys to delay the subs, not fuck around with filenames.
j
You're literally relying on a hack to do something that could be done properly.
>So I guess I'm not in the cool encoders kid club you are.
Programmers can use hacks too, that doesn't mean they're good programmers.
Then I hope someone doesn't think like you do and actually makes a simple script for those who want it. if there's a script for other things I find useless, someone might find this useful.