>2017 >CCCP is dead >KCP is dead Is it best to just

>>60748169
You strike me as a VLC kinda guy

>>60748169
>cccp
bloat
>kcp
bloat

Just use mpchc with mad vr.

MPC-HC+LAV Filters+MadVR is the way to go

>>60748270
M8 lav filters have been integrated into mpchc for like 2 years.

>>60748169
>not using mpv

>>60748169
Just use MPC-HC with madVR

File: media_player.jpg (120KB, 750x781px) Image search: [Google]

120KB, 750x781px

so true

>>60748169
>KCP is dead
? Last update was this week.

>>60748169
CCCP is not dead, its just devs dont want to push rc releases as final versions.

>>60748212
>111MB is bloat
must suck being a poorfag

>>60748427
>>60748442
>>60748270
Yeah just ignore the mpv shills. It never ceases to amaze me how many pathetic people there are on /g/ who use an inferior and more inconvenient media player on windows.

>>60748754
>It never ceases to amaze me how many pathetic people there are on /g/ who use an inferior and more inconvenient media player on windows.
Yeah, why don't those people switch to mpv?

You know what else is dead?

MPC-HC

>>60748169
If you're not using mpv by now, you're doing it wrong

File: images.duckduckgo.com.png (490KB, 449x401px) Image search: [Google]

490KB, 449x401px

>>60748779
>he's this dedicated to shilling for a media player

File: mpv.jpg (145KB, 730x730px) Image search: [Google]

145KB, 730x730px

>>60748846

>>60748869
>Can't even navigate between items in the same folder by default
>You actually have to add that ability in via config files
>Still doesn't have seekbar thumbnails in the year 2017

File: 1487026334609.jpg (132KB, 800x484px) Image search: [Google]

132KB, 800x484px

>>60749265

File: madvr.png (2MB, 1920x1080px) Image search: [Google]

2MB, 1920x1080px

CHALLENGE: produce an mpv upscale as sharp with as little noise/artifacting as this MadVR upscale. Source file is [Doki]_Yuru_Yuri_-_07v2_(1280x720_h264_BD_AAC)_[DF080EE5] which can be found at https://nibl.co.uk/bots.php

Several anons attempted and failed in the last mpv thread.

>>60748869
came in like a kekking ball

File: [Doki]_Yuru_Yuri_-_07v2_(1280x720_h264_BD_AAC)_[DF080EE5].mkv_snapshot_01.33_[2017.06.04_19.51.23].png (2MB, 1280x720px) Image search: [Google]

2MB, 1280x720px

>>60750404
Here's a source image for those who don't want to download the whole episode. Also, you must post repeatable mpv settings to qualify for this challenge.

>>60748358
This. NGU AA is a vastly superior upscaling algorithm to anything mpv can use right now. mpv is good for toasters, though, as it's less intensive with equivalent settings than MadVR.

>>60748803
MPC-HC had a new version released in March 2017. It's not dead.

>no youtube-dl support
D R O P P E D
R
O
P
P
E
D

File: smugsawako.jpg (34KB, 359x377px) Image search: [Google]

34KB, 359x377px

>>60750508
>there are actually people who don't have JDownloader2

>>60750516
does it let you stream videos from hundreds of sites with mpc-hc?

>>60750554
You can stream directly to MPC-HC/MadVR with Livestreamer. I've used it to make poor quality soccer streams watchable with NGU AA, which is way better for poor sources than anything mpv can use.

>>60750516
>download manager
>in the year 2017

Let me guess, you also use IRC and screensavers, too?

File: [SSKN] Love Live! Sunshine!! - HAPPY PARTY TRAIN [BD][720p-FLAC][64EF6307].mkv 00_02_13.591.png (2MB, 1920x1080px) Image search: [Google]

2MB, 1920x1080px

https://diff.pics/2LlvhMKcad2W/1

>>60750508
try livestreamer/streamlink

>>60750575
Actually, yes, I still use IRC to download anime, as I'm American and IRC doesn't place my IP address on a tracker for anti-P2P companies to harvest.

>>60750592
>>60750571
doesnt support as many sites
youtube-dl is much better

>>60750649
Considering JDownloader2 usually downloads even the largest Youtube videos in a matter of seconds, there's little to zero functional difference between it and youtube-dl.

>>60748212
>>cccp
>bloat

>>60750658
so it can stream to mpc-hc and be able to seek?

>>60750622
xdcc right

how do you do your searches for episodes, i want to switch my torrent based series downloader to just use xdcc behind the scenes

>>60750715
https://nibl.co.uk/bots.php is the largest available collection of bots searchable via a packlist. Aside from that, if I see a release I want and it's not on that website, I'll generally use a search engine to check whether the subbing group who made the release I desire has their own IRC channel with bots and a packlist, as they often do. For example, I recently downloaded all of seasons 1 and 2 of K-ON from Chihiro's IRC channel, as their K-ON 720p release is the highest quality release available.
For games I usually search on https://ixirc.com/ as well as various search engine results for the search xdcc search.

>>60750590
madvr retards btfo, stay on you inferior mpc+memevr

>>60750590
Congratulations, mpv can produce an equally blurry result as an improperly configured MadVR. Call me when you can complete the challenge here: >>60750404

>>60750404
mpv easily won that last thread

>>60752040
t.mpvdev
Every mpv screenshot posted was either much blurrier or much noisier.

File: mugiconsider.png (328KB, 744x661px) Image search: [Google]

328KB, 744x661px

/g/ is strongly biased towards mpv and will never give MadVR a fair evaluation because mpv is FOSS and MadVR is not.

>>60752177
true and also because mpv is windows only

>>60752215
er madvr is windows only

mpc-be.
mpc-hc fails at 4k h265.

>>60752215
mpc-hc*

>>60752058
t.faggot
>mpv users prove me wrong that you can improve the still of this shit animay using this release. I am the one judge on what looks better. PROVE ME WRONG.

>Get's proven wrong on several accounts. Refuses to take the fedora off and accept that MPV is superior.

File: mfw.gif (349KB, 582x582px) Image search: [Google]

349KB, 582x582px

>>60752230
>he bought a meme resolution display before the 4k standard is even finalized

>>60748169
why do i need newer codecs than my 3yo ones

>>60752269
Post the screenshot and repeatable mpv settings which supposedly fit the criteria laid down by the challenge >>60750404
You will be assumed to be fabricating your claims if you do not.

PotPlayer master race

File: Konachan.com - 200856 blonde_hair brown_eyes brown_hair group headphones instrument katou_hazuki long_hair oumae_kumiko ponytail purple_eyes seifuku short_hair yellow_eyes.jpg (215KB, 1100x777px) Image search: [Google]

215KB, 1100x777px

Has someone compiled comparisons of MPC-HC+madVR vs mpv? I'd love to see downscaling, downscaling, interpolation, etc. on both. I don't have a choice on Lincuck and have grown to love mpv, but I'd like to see if it's a better choice than the competition.

>>60752281
>mpv users prove me wrong that you can improve the still of this shit animay using this release. I am the one judge on what looks better. PROVE ME WRONG.

>>60748358
This.
>>60752269
>shit anime
Hi /v/

>>60752291
It's objectively worse at upscaling, but equivalent at everything else and less demanding than MadVR with equivalent settings. Its subtitle renderer is also superior (faster) to xysubfilter or MPC's internal renderer. MadVR is better for nice PCs and mpv is better for toasters.
NGU AA + adaptive sharpen + crispen edges + denoising (either via video card or ffdshow's denoise3d with time set to 0) is a combination which is more powerful at creating quality upscales than anything mpv can use, specifically because NGU AA + sharpening results in less noise than mpv's upscalers + equivalent sharpening settings.

>>60752297
Oh, so you're just trolling. Got it.

>>60752342
wtf is mpv man

I always used to install CCCP and call it a day. So standalone MPC-HC can play almost everything now? Will I really need to add on MadVR or only if I want to get into audiophile levels of detail? How much more taxing is it on one's machine?

>>60752359
>only if I want to get into audiophile levels of detail
This is >>60750437 upscaled in MPC-HC with MadVR disabled. If you cannot see a massive difference between this and >>60750404 there's something seriously wrong with your eyes or your display.
>How much more taxing is it on one's machine?
For reference, I experience zero frame drops using the settings which produced >>60750404 on my FX-6300/R9 270x machine, which was low-mid-tier a few years ago.

File: Untitled-1.png (1MB, 1920x1080px) Image search: [Google]

1MB, 1920x1080px

>>60752395
>>60752359
Forgot the screenshot.

>>60752271
Nah, I actually work in video production. I don't think I've ever actually watched finished commercial 4k content.

>>60750404
Going to sleep now. I fully expect to wake up to mpv shills posting blurry or noisy screenshots and then samefagging claiming they're better.

>>60752395
>>60752411
That's really nice actually. Is there a recent guide on how to achieve the same thing or at least something that goes through its settings?

>FX-6300/R9 270x
That's reassuring, thanks.

File: 1477817569072.jpg (17KB, 480x347px) Image search: [Google]

17KB, 480x347px

>>60751953
>Improperly configure
Shilling this hard

>>60752478
I don't why you're shilling this hard when that scaling looks like a fucking oil paint

>>60750404
What madvr settings did you use?

File: 1493703285385.gif (1MB, 500x364px) Image search: [Google]

1MB, 500x364px

>>60752342
>Hurr, you're trolling me because you're mocking the stupidity of my request. That means you're retarded and your point is invalid, gotcha.

Do you realize the ridiculousness of your request?

We (MPV users) aren't under any circumstance entitled to have to provide proof for your claim. You made the claim that MPC-HC + MadVR has the best quality upscale given ONE and ONLY ONE screenshot you've hand selected yourself.

The burden of proof lies on you to prove your claim that MPC-HC + MadVR is, without a doubt, better than even the _best_ configured MPV setup.

Literally suck a fat cock and swallow the load.

You're dumb as fuck if you think anyone's interested in trying to prove you wrong. I, nor anybody else, is particularly interested in torrenting your specific release's episode just to tinker with MPV settings to get an upscale that exceeds your biased standards for MPC-HC, because nobody gives a shit about changing your mind over to MPV.

The way you should be going about this is saying "I've experimented with multiple configurations for MPC-HC + MadVR and MPV and these are my findings for X. I've supplied 6 comparative examples. You can verify my samples by using the following releases and the following configurations. The results suggest that MPC-HC + MadVR has better fidelity."

In doing so you make your claim and you follow through on your own burden of proof and do your due diligence on documenting your experimental process.

>>60752774
This anon >>60744149 already proved that NGU and ewa+nnedi3 are almost the same.
These are the comparison the anon made:
https://diff.pics/lRp8gSTrxJH3/1

Just don't try to argue with the madvr fags, they need a gui and use windows(cause they're retarded)

>>60752990
*comparisons

>>60748169
>mpc-hc
>Not mpc-be

>>60745081
How do I correctly configure MadVR?
Because as a madvr user those settings look good to me( I use medium quality and AA tho).

>>60748340
M8 MPC-HC has not been updated for like 3 years.

>>60752334
you are objectively retarded

>>60750404
First we need some mpv user with madvr installed to verify your upscaling in madvr.

>>60753025
So these are not the same thing?
https://mpc-hc.org/
https://sourceforge.net/projects/mpc-hc/

>>60753025
mpc-hc received an update like 2 months ago

>>60752990
Nnedi3 smoothing and ringing? Great combination you mpvfags have.

File: mpv-shot0001.png (3MB, 1920x1080px) Image search: [Google]

3MB, 1920x1080px

>>60750404
ewa_lanczossharp+SSimSuperRes+adaptive-sharpen

SMplayer is mpv without the autism

>>60748169
VLC (which is the best media player of all time) is still alive. Anything else doesn't matter.

>>60753115
I'm curious as to how those edge artifacts translate in motion.

>>60753115
You should use https://diff.pics/ next time
Also where I can get SSimSuperRes and adaptive-sharpen shaders?

>>60753163
Idk, I'm not an autistic, don't even know what artifacts you are talking about
>>60753185
https://gist.github.com/igv/

>>60753233
Thanks

Can MPV do borderless window? It's a feature I use frequently with MPC

>>60753277
https://mpv.io/manual/master/#options-no-border

>>60753312
Awesome thanks

>>60753160
>building font cache all this time

>>60750404
MadVR doesn't show up on screenshots though, unless they changed it recently. And that just looks like a waifu2x upscale which makes the lines look like shit

How do I create a profile in mpv?
I want 720p videos to use certain upscaler and 1080p videos other upscalers

>>60753442
I used this as exemple when creating my profiles, if you can't figure something out just search the manual

https://github.com/Argon-/mpv-config/blob/master/mpv.conf

>>60753442
https://github.com/wm4/mpv-scripts/blob/master/auto-profiles.lua

[SD]
profile-desc=cond:get('height',0) < 720
scale=haasnsoft
cscale=ewa_lanczossoft

>>60753423
>waifu2x
Looks like we have the answer on why nobody has been able to produce an upscaling like this >>60750404 even with NGU.

>>60753531
kys

>>60753560
So tell me the settings that were used in MadVR.
They're still (((undisclosed)))

What has the best video playback objectively?

>>60753663
quality is subjective

>>60753663
VLC

>>60753663
judging by previous thread all you need is algorithm for over-thinning, retards think it makes picture so sharp and clear.

>>60753663
mpv with opengl-hq is more than enough
DO NOT DIVE INTO THE AUTISM

>>60753738
>no linked mkv support
>no proper automatic subtitle selection
>OSD instead of proper controls

>>60753115
>>60750404
>tfw can't see a difference, both are better than the source in the same way
I need to go see an eye doctor...

>>60753738
uh oh... i have no settings on my mpv, it only says "Drop files to play here"

>>60753762
You aren't autistic enough that's all

>>60753767
>You aren't a madvr shill that's all
ftfy

>>60748169
Just use mpv it's literally normie proof

>>60753785
I'm bashing both sides

>>60750437
Are you using mpv out of the box or wat, for me mpv and kcp version madvr look exactly the same

>>60753510
Is this ok?
[SUB1080P]
profile-desc=cond:get('height',0) < 1080

I want to only upscale video that is lower than 1080P

>>60753844
your monitor is 1920x1080?

>>60753880
Yes

>>60753762
mpv's has more ringing. And if you don't know what is it, you have to train your eyes a bit on more severe examples of this kind of artifacts to start noticing it.

>>60753897
then it doesn't upscaling 1080p anyway

>>60753531
I use waifu2x a lot so that's why I recognized it, lets see if he posts the settings he used and if it looks the same for others, but I'm pretty sure he just used waifu2x especially because of how the lines look.

>>60753911
What settings were used in madvr?
Or is a waifu2x upscale and you're just shitposting?

>>60753931
I don't know, it's not me who made it. I just commented on a difference visible to me.

>>60753919
What do you mean by "it doesn't"?
If I watch a 720p video is going to be upscaled to 1080p, right?

>>60753954
yes

>>60753991
Are these profiles well structured?

[DVD] #Only applied to 480P video
profile-desc=cond:get('height',480) < 720
scale=ewa_lanczos
cscale=ewa_lanczossharp
sigmoid-upscaling=yes

[HD] #Only applied to 720P video.
profile-desc=cond:get('height',720) < 1080
opengl-shaders="C:\Program Files\mpv\shaders\nnedi3-nns16-win8x4-yuv.hook,C:\Program Files\mpv\shaders\noise_static_luma.hook,C:\Program Files\mpv\shaders\noise_static_chroma.hook"
sigmoid-upscaling=yes
dscale=ewa_lanczos
correct-downscaling=yes

[FULLHD] #Only applied to 1080P video
profile-desc=cond:get('height',1080) = 1080
cscale=ewa_lanczossharp

>>60753025
So so wrong m80

>>60754112
Why you never use profile=opengl-hq ?

>>60754143
It's applied to all the profiles(I think).

Full conf file:

# Video
vo=opengl
opengl-backend=angle
hwdec=no
profile=opengl-hq
deband=yes
deband-iterations=4

# Audio
volume=100
volume-max=100
audio-channels=stereo

# screenshots
screenshot-format=jpg
screenshot-jpeg-quality=100
screenshot-template="%f %P"

[DVD]
profile-desc=cond:get('height',480) < 720
scale=ewa_lanczos
cscale=ewa_lanczossharp
sigmoid-upscaling=yes

[HD]
profile-desc=cond:get('height',720) < 1080
opengl-shaders="C:\Program Files\mpv\shaders\nnedi3-nns16-win8x4-yuv.hook,C:\Program Files\mpv\shaders\noise_static_luma.hook,C:\Program Files\mpv\shaders\noise_static_chroma.hook"
sigmoid-upscaling=yes
dscale=ewa_lanczos
correct-downscaling=yes

[FULLHD]
profile-desc=cond:get('height',1080) = 1080
cscale=ewa_lanczossharp

>>60754157
>C:\Program Files\mpv\shaders\nnedi3-nns16-win8x4-yuv.hook
you can replace it with just
~~/shaders/nnedi3-nns16-win8x4-yuv.hook

>>60754157
I think HD profile should be before DVD, for example 480 is smaller than 1080, so your DVD profile will never be applied.

>>60753738
that's what I thought at first, but

profile=opengl-hq
scale=ewa_lanczossharp
cscale=ewa_lanczossharp

not complicated, that's all you need

>>60748367
what is it that is so bad about mpv?

>>60754577
Autism in system requirements.

>>60748754
You're right. It's windows itself that is inconvenient and inferior. switching to mpv isn't enough

Try mpv.

how is mpdn?

https://nightly.mpc-hc.org/

MPC-HC hasn't fixed the subtitles malware shit

switched to VLC.
>10 second search time when clicking on the timeline bar to go to the next scene
>no thumbnail maker
>VLC_Snap12412412.png (no option to use the movie's name instead for easy identification when looking at the screenshot 3 years from now).
>movie area is not dragable
>frequently doens't respond to the pause button, continues to play for 10 seconds before pausing

french turd program, but atleast they fixed the subtitles malware

>>60754981
>subtitle malware
I thought MPC-HC didn't have this problem in the first place

>>60755003
It doesn't, he's a tech-illiterate faggot.

>>60755003
>>60755114

source: my ass

if you get source I will apologize for being a douchebag

>>60754981
>subtitles malware shit
You fucking wot m8.

File: 1492228469587.jpg (78KB, 605x454px) Image search: [Google]

78KB, 605x454px

>watching anime

>>60755003
It doesn't because MPC-HC doesn't use a repository like VLC

http://blog.checkpoint.com/2017/05/23/hacked-in-translation/

I just use mpv because it's the only player that can reliably play h.265 on my low end laptop

>>60755228
http://blog.checkpoint.com/2017/05/23/hacked-in-translation/

You are a douchebag that don't have the balls to apologize since MPC-HC was never affected by this in the first place

>>60755418
With vulkan support it will be even better

>>60755323
>>60755426
Even better source, devs themselves: https://trac.mpc-hc.org/ticket/6169

>>60753844
https://mpv.io/manual/master/#options-scaler-resizes-only

>>60750404
I would like to see that with:
image enhancements>
sharpen @4
thin @8
activate anti-bloat @150%
activate anti-ringing

File: before.png (2MB, 1920x1080px) Image search: [Google]

2MB, 1920x1080px

>>60756587
That's what I used to go from here...

File: after.png (2MB, 1920x1080px) Image search: [Google]

2MB, 1920x1080px

>>60756587
>>60756653
...to here

>>60750461
Isn't nnedi better and also available in mpv?

>>60756678
There is a name for that, it's called autism.

>>60756715
Depend on the source. It is much slower than NGU though.

LAV Megamix has superseded KCP for a while now.

Made by weebs for weebs.

>>60748169
CCCP still works for me

>>60750404
Anime looks better a little blurry, having it that sharp makes anime look too hard and not soft enough. I'd recommend decreasing the sharpening enchantments

File: 1453343228954.png (28KB, 186x208px) Image search: [Google]

28KB, 186x208px

>>60748212
>>60748270
>>60748358
I don't know what I am doing wrong but when I use madVR the videos get fucked up with white flashes every now and then during playback and it also causes screen tearing, I followed the how did I played back guide to the letter yet. I went back to ffdshow/ffvideo and the problem disappeared.

>>60757812
maybe your pc is old?

>>60748169
I use the k-lite codec; standard or mega if you want it all.

>>60757984
2011 GPU and 2014 CPU, I doubt it's hardware-related. It happens only with madVR, any other player and decoder work fine.

lately mpc-hc has been giving me issues,

sound will keep going but video will pause.
Trying to seek forward doesn't do anything with the video only sound.

At first I thought this was a one-off thing, but seems to be happening more frequently, did groups switch to some codec that mpc-hc doesn't handle well ?

How do I run a twitch stream in mpv in Windows?
I'm using mpv.exe url in cmd but only sound comes through.

>>60748169
mpv

>>60758275
streamlink --player="mpv" --player-continuous-http <url> 720p,best

>>60757812
What's your gpu?

>>60750404
What settings are you using? My stuff doesn't look this good.

>>60758302
Doesn't work.

>>60757591
this is the equivalent of those people who say videogames should be played at 30fps

>>60758550
The upscale for looks to sharp for my taste

>>60758634
>>60758550
It's a matter of personal taste. Some people like it really sharp, others don't.

File: ScreenShot_20170605103830.png (40KB, 428x392px) Image search: [Google]

40KB, 428x392px

>>60748169
>CCCP is dead
still works

I have a subtitles .ass file and a bunch of fonts that I want to use with the RAW mkv file. How can I make it work in MPV (external subs + fonts) without fiddling with the source file with the mkvtoolnix and whatnot? Anything I should I add to the config file and/or the font folder structure in the video file folder?

mpc-hc wasn't playing some x265 videos for me, switched to mpv and not going back

>>60759210
Learn 2 update your software.

>>60753115
Too noisy. Fail. MadVR still undefeated.

File: 1492812349397.jpg (44KB, 230x272px) Image search: [Google]

44KB, 230x272px

vlc can't properly display evangelion so i dropped that shit so fast.

>>60759537
t. ameridumb

>>60753931
It's not a waifu2x upscale. You could upscale >>60750437 with waifu2x yourself to test that.
Settings for >>60750404 are NGU AA medium (yes, I didn't even bother to set it to very high because I don't even expect mpv to be able to equal medium), and under upscaling refinement adaptive sharpen 1.0 and edge crispening 1.0. Denoising is done via my R9 270x via RadeonMod settings, but I've tested it and hqdn3d on this screenshot and they have the same effect, so an mpv user can use hqdn3d in place of video card denoising if need be. Video card denoising is preferable as hqdn3d tends to blur already blurry content even further, but hqdn3d won't have that kind of a negative effect on this particular screenshot.
As I've been saying all along, NGU AA + higher sharpening settings > NGU Sharp + lower sharpening settings, or any other algorithm + any sharpening settings, really. The primary advantage of NGU AA as an upscaling algorithm is its ability to handle strong sharpening settings without producing a lot of noise.

Is there another player with youtube-dl support? I really just can't stand mpv anymore.

>>60759724
Oh, activate anti-ringing filter under upscaling refinement too.

>>60752276
Get the fuck off /g/ if you don't know what h.265 is.

How long until we can apply waifu2x in real time?

>>60759724
>>60759785
Enjoy your life with autism.

>>60759537
>waifu2x still undefeated.
Fixed

>>60757812
dump pls

File: waifu2x.png (2MB, 1920x1080px) Image search: [Google]

2MB, 1920x1080px

>>60759937
It's not even all that necessary anymore. This is a waifu2x upscale. >>60750404 compares favorably to it.

>>60759967
that retard just can't distinguish thin edges from sharpness, sharpness from noise, just ignore that ameridumb

>>60759967
I gave repeatable settings for the MadVR upscale, anon. Stop trolling and let the adults talk.

>>60756715
>added NGU Anti-Alias algorithm, intended as a potential NNEDI3 replacement
I've done a lot of testing of these two and NGU AA is both better at upscaling and faster. There's a reason madshi put this line in his version notes.

>>60760015
What about shitty sources?
post your settings

>>60760180
With shitty settings just keep it on NGU AA and reduce the sharpening settings until it doesn't look like shit. You may need to completely disable adaptive sharpening and edge crispening depending on the source, but NGU AA in general REALLY shines versus other upscaling algorithms with poor sources. It's like voodoo fucking magic with poor sources and I've had great success porting poor quality soccer streams to MPC with Livestreamer and making them actually fucking watchable with NGU AA.

>>60759724
Youre retarded on so many levels, wow. No wonder the image is completely fucked up in >>60750404
>using RadeonMod and its denoising
>using NGU AA with shitload of sharpening instead of just NGU Sharp
>suggesting to use already very sharp NGU Sharp with additional sharpening
Jesus christ. I hope no one takes seriously all of this shit.
>The primary advantage of NGU AA as an upscaling algorithm is its ability to handle strong sharpening settings without producing a lot of noise.
Youre not supposed to use strong sharpening with NGU AA. If you want more sharp just use Standard and Sharp. You can add a little bit of sharpness with adaptive sharpen to AA, but adding a lot will ruin the looks.

>>60760241
>With shitty settings
With shitty sources. FTFM.

Give me a GUI for mpv.

I just use what that file on nyaa said to use

>>60760384
They had a great guide. I hope you use the last one?

>>60760244
>No wonder the image is completely fucked up
That's your opinion. As >>60758694 pointed out some people prefer more sharpening some people prefer less. I like as much sharpening as I can crank out without producing noise or artifacting. I think crisp, clear lineart looks fantastic. Maasaki Yuasa agreed with me on this when he recently extolled the virtues of vectorized lineart in anime and its ability to produce crisp, clear lines.
>using NGU AA with shitload of sharpening instead of just NGU Sharp
Anon, I have thoroughly tested this and NGU AA with more sharpening produces an equally sharp but cleaner picture to NGU Sharp with less sharpening. NGU Sharp adds a lot of noise and occasionally some very strange artifacts which don't fall under the category of traditional artifacting like ringing, bloating, etc.
>suggesting to use already very sharp NGU Sharp with additional sharpening
I did not do this.
>Youre not supposed to use strong sharpening with NGU AA. If you want more sharp just use Standard and Sharp.
You're entitled to this opinion, but I've found it to be objectively incorrect in my very thorough testing.

>>60758070
>2011 GPU
My previous 670 would shit itself with NNEDI on madVR.

>>60748340
>>60753025
>on 29 Sep 2013
So outdated, m88

>>60760694
MPC-HC's latest version is March 6, 2017.

>>60760241
>he recommends to use adaptive sharpening
it looks like shit in madvr, just leave, you have no idea what you are talking about

>>60760753
I mean since they have lavfilters by default, bro
cccp/kcp is useless for almost 4 years

File: satanialaugh.webm (3MB, 1440x810px) Image search: [Google]

3MB, 1440x810px

>>60760769
>he is this mad nobody has been able to complete the >>60750404 challenge

What's it gotta take for media players to obtain infinitely smooth seeking? I don't want to wait 250ms-500ms for a seek, I want 0ms.

What's the bottleneck in media playback technology?

>>60760440
I don't know, I think I followed it last year

>>60760825
who said nobody has been able to complete it?

>>60760460
>That's your opinion. As >>60758694 pointed out some people prefer more sharpening some people prefer less.
The most correct way is to get upscaled image as close as possible to the ground truth. Of course people have tastes, some like very sharp and some autists like black and white image or who knows what else, its still makes the image look unnatural and is retarded IMO.
>NGU Sharp adds a lot of noise
It doesnt add any noise it reveals the source's flaws in form of "artifacts". The source has those artifacts, NGU Sharp just highlights it.
>You're entitled to this opinion, but I've found it to be objectively incorrect in my very thorough testing.
Same. The image you produced looks like oversharpened unnatural oil painting garbage. I cant stop you from having obscene taste in picture looks but please keep it to yourself and dont give normal people retarded (((suggestions))).

>>60760943
You should update probably. Lots changed in the guide since then. Do you even have NGU in your madVR build?

>>60761265
I don't think I have it

>>60761647
Ah and I never got xysubfilter to work

>>60761265
oil painting meme filter

>>60761171
>The most correct way is to get upscaled image as close as possible to the ground truth.
This logic is the same as refusing to correct large frequency response deviations from accuracy in your speakers/headphones with EQ because "muh signal purity." You're saying that subjective improvement is inherently bad because "muh image purity."
>It doesnt add any noise it reveals the source's flaws in form of "artifacts". The source has those artifacts, NGU Sharp just highlights it.
This is incorrect. IIRC NGU AA was developed precisely because of the strange artifacts I was talking about.
>but please keep it to yourself and dont give normal people retarded (((suggestions))).
I'm sorry, I read this as "HEIL HITLER, AND STOP TELLING PEOPLE TO USE SOMETHING OTHER THAN MPV WHICH I AM ON THE DEVTEAM OF." I'm going to keep posting it and let people make up their own mind whether it looks nice or not.
>>>/pol/

>>60760998
>produce an mpv upscale as sharp with as little noise/artifacting as this MadVR upscale
Nobody has done this. Please note that the challenge was NOT "produce an mpv upscale which you subjectively feel looks better than this MadVR upscale." The challenge is to produce an upscale as sharp with as little noise and artifacting as the MadVR upscale. This challenge has not been completed.

FWIW, for poor sources NGU AA + sharpening is usually better than NGU Sharp, because NGU AA smooths over source artifacts (e.g. aliasing) in such a way that the artifacts don't come back by using post sharpening.

NGU Sharp is totally unforgiving. If the source has artifacts, NGU Sharp shows them very clearly. So for bad quality sources NGU Sharp should not be used. However, for high quality sources, NGU Sharp clearly produces better results than NGU AA + sharpening. I've tested this by downscaling high quality material, then scaling it back up with NGU AA + sharpening vs. NGU Sharp. In such tests NGU Sharp clearly wins over NGU AA + sharpening.

So, as always, pick the right tool for the job.

NNEDI vs. NGU AA is mostly about speed. Both produce a similar look, but NGU AA is significantly sharper at comparable image quality. Sometimes NGU AA also looks better, but it's not a dramatic difference in quality, last time I checked.

BTW, AdaptiveSharpen does appear to look better in mpv than in madVR. Seems igv did a really good job there. madVR seems to use the original AdaptiveSharpen provided by the original AdaptiveSharpen inventor. igv's version looks better, for some reason. Maybe someone could let madshi know about this?

>>60761866
>This is incorrect. IIRC NGU AA was developed precisely because of the strange artifacts I was talking about.
I should add that the artifacts I'm talking about, IIRC, were primarily the result of quadrupling when the original NGU (it was just called NGU) algorithm was in MadVR. These artifacts still occur with NGU Sharp but not with NGU AA, and these artifacts generally don't occur when only doubling with NGU Sharp.

>>60761977
P.S: I meant to write NGU AA is significantly *faster* at comparable image quality. *Faster*, not sharper.

>>60761939
>filthy madvr shill
>incompetent
>ameridumb
your opinion has been disregarded

>>60761977
>In such tests NGU Sharp clearly wins over NGU AA + sharpening.
This has not been the case in my very thorough testing in conjunction with video card denoising. In my testing with high quality 720p BD rips, when I make NGU Sharp and NGU AA equivalently sharp with sharpening settings and video card denoising enabled, NGU Sharp still contains significant noise which is much more apparent in motion than not compared to NGU AA.
The addition of denoising may be the variable here which is causing you and I to produce different results in our testing.

File: [jackoneill] Armitage III 01 [C13202FE].mkv_snapshot_01.05_[2017.04.03_16.48.34].jpg (233KB, 704x480px) Image search: [Google]

233KB, 704x480px

What I should use for old anime that doesn't have a bd release?

>>60762212
Jinc/ewa_lanczossharp if you want to preserve grain, NGU AA if you do not.

>>60762064
The post you are replying to did not state an opinion.

>>60762256
Who doesn't want to preserve grain?

>>60761977
>>60762127
Another variable which may be causing you and I to produce different results in our testing could be that I've been testing with stronger sharpening settings than you have, as a matter of personal preference.

File: 0be.png (365KB, 673x680px) Image search: [Google]

365KB, 673x680px

>>60762351
>grain
>mfw
I bet you like tape/vinyl hiss too.

>>60762212
Nothing. Stretched and blurry, because that's how we watched dvds back in the day. Or are you trying to say you deserve better than dvd watchers? that's racist

>>60762401
false equivalence

>>60762401
I already have grainy vision so grain doesn't affect me

>>60762435
It's not at all. Grain and tape/vinyl hiss are both distortions resulting from the use of analog technology.

What is wrong with old codecs? It will still play anime, will it not?

File: 1472242390903.jpg (133KB, 1920x1040px) Image search: [Google]

133KB, 1920x1040px

>>60762401

>>60760015

The MadVR upscale actually looks better than the waifu2x upscale. Huh.

File: 93233266.jpg (17KB, 405x289px) Image search: [Google]

17KB, 405x289px

>>60761866
>This is incorrect. IIRC NGU AA was developed precisely because of the strange artifacts I was talking about.
Go tell it to madshi and he will probably spit in your face.
>>60761866
>MPV WHICH I AM ON THE DEVTEAM OF
Isnt it obvious that i use madvr myself?
>>60762009
>I should add that the artifacts I'm talking about, IIRC, were primarily the result of quadrupling when the original NGU (it was just called NGU) algorithm was in MadVR. These artifacts still occur with NGU Sharp but not with NGU AA, and these artifacts generally don't occur when only doubling with NGU Sharp.
What you see is NOT ARTIFACTS. You might call it artifacts if you insist but its just a garbage in the image. Its already there. NGU Sharp just reveals it. Again, if you think im worng go tell it to madshi if youre not a pussy. Pic related is him listening to you.

File: 134351305.jpg (856KB, 2880x2160px) Image search: [Google]

856KB, 2880x2160px

>>60762212
Old version of CCCP that has fddshow filters

>>60762401
Grain is love. Grain is life

>>60762212
Literally doesnt matter. I just tested with Jinc, NGU AA and nnedi3. They all look like shit and very similar. NGU looks very slightly smoother. You'd probably not notice any difference without zooming into the picture.

File: 1386796587895.png (7KB, 463x313px) Image search: [Google]

7KB, 463x313px

Haven't updated my MPC-HC + MadVR setup in years. Do these settings still hold up for cartoons or do I need to search for something more modern?

>>60764135
yes

>>60764135
Update MPC-HC and MadVR. Use NGU AA for image upscaling.

File: 534534535.png (32KB, 1920x1040px) Image search: [Google]

32KB, 1920x1040px

>>60748169
>he doesn't use VLC with Lanczos scaling

>>60756678
You fucked up the smoothness of the lines. Look at those misshapen lines of her chin when it meets her hair.

>>60764352
>software scaler

>>60764135
What is your videocard?

1.)Chroma upscaling: NGU Anti-Alias, medium.
2.)Image upscaling: NGU Anti-Alias (luma doubling: medium; chroma: normal) Everything else "let madVR decide"
3.)Image downscaling: Bicubic 150 AR LL.

Enable "reduce banding artifacts" medium/high and tick "reduce ringing artifacts" in processing->artifact removal.

Tick smooth motion in rendering->smooth motion.

Untick everything under rendering->trade quality for performance.

>>60764462
>You fucked up the smoothness of the lines.
That's the least of his worries. He's fucking the artistic decision making of the anime's creators in the ass by thinning the lineart like that.

>>60756678
I think i know who you are.

>>60764497
Exactly what this guy said.

>>60764135
>>60764497
And start cooking eggs on your gpu

>>60764580
If you have a toaster. Any non-onboard GPU made within the last few years that wasn't absolute bottom-tier three years ago should be able to handle >>60764497 just fine.

>>60764580
NGU Medium is very fast. Only ancient onboard cards might have problems with it.

>>60762687
It's literally a fucking bullshit because NGU is a proprietary algorithm.

>>60760753
Yes but the last updates are mostly things like changing copyright 2016 to copyright 2017. That's why people keep saying it's dead and they are right. :(

>>60764887
Read the patchnotes.

>>60761977
But NGU-AA has a terrible issue, it crops picture a little bit just like NNEDI3.

>>60762212
Raspberry pi + RCA or S-video cable + old CRT TV + LibreElec
I'm serious, best way to watch evangelion tbqh

>>60764918
Reading is not enough you have to understand too.
https://github.com/mpc-hc/mpc-hc/commits/develop

>>60764497
>>60764347

Does MPC come with xysubfilter/xy-VSFilter/whatever sub renderer by default or do I need to update that too?

>>60759724
>hqdn3d
Where do I find it?
Also, what dithering do I use?

>>60765069
You need to update it.

>>60765070
Are you an mpv or MadVR user?

>>60765328
madVR

>>60765344
Default dithering options have always worked well for me, meaning ordered dithering.
For hqdn3d install ffdshow, set ffdshow video decoder to prefer under external filters, enable denoise3d in ffdshow settings under Blur & NR, check the HQ box, turn time slider to 0 to prevent trailing ghost images in motion.

>>60765344
>>60765441
Also, in ffdshow settings, set directshow control merit to normal and under codecs set all the decoders to libavcodec or similar.

>>60750575
IRC is great for downloading books.

File: ffdshow.png (2MB, 1920x1080px) Image search: [Google]

2MB, 1920x1080px

>>60765070
I should illustrate the one small drawback to denoise3d to you. This is denoise3d.

>>60764934
You mean a half pixel shift. It's just offset. Not an issue.

>>60765528
Was grass blurry without it? Or is it dogshit?

File: amd.png (2MB, 1920x1080px) Image search: [Google]

2MB, 1920x1080px

>>60765528
This is AMD's denoising with my R9 270x. Note the grass in the lower left. denoise3d has an unfortunate tendency to blur out of focus content like that, but it's typically not something you'll ever notice if you're not looking for it, and only occurs on a handful of scenes. Still, if your GPU is capable of doing denoising, that's preferable to denoise3d.

So where can I find good configurations and tips for different players compatible with linux? I'm currently using a 3200x1800 screen with no configuration and I want my shit to look good.

>>60765528
>>60760015
Dude your pictures look disgusting. Holy shit. Those fucking thick oily lines,

Tfw I can't use those settings in madvr because my GPU ramps up and the fan noise gets annoying

>>60765677
Which settings? Whats your GPU?

>>60765676
You're replying to an image upscaled with http://waifu2x.udp.jp/ you silly goose.

>>60765668
PS I'm running on a laptop with Intel HD Graphics 520

>>60765676
>>60765704
Also, I wouldn't normally use such high sharpening settings with a poor source like >>60765528
That image was only done to compare denoising algorithms.

>>60758275
make sure you have youtube-dl.exe in your mpv.exe folder
https://mpv.srsfckn.biz/
then you should be able to just drag twitch links into mpv window to stream them. or get the firefox addon to right click and open links in mpv

File: Capture.png (119KB, 566x538px) Image search: [Google]

119KB, 566x538px

>>60764497
>66C just to watch chinese cartoons

>>60765703
These settings >>60759724
7970 vapor-x. It's noisy as hell anyway with very little load, so no surprise here.

>>60765704
Quoted wrong post.
>>60765528
>>60765566
>>60750404
ABSOLUTELY DISGUSTING.

>>60765677
>tfw when fully passive
Feels good
(unless you are a gayymer)

>>60765754
My cuck 970 used to go a little hotter back when I used nnedi on madVR. The room got very toasty at times.

>>60765677
Use headphones? Turn your volume up? This is one of the sillier complaints I've read.
>>60765769
>how would you like your anime
>JUST BLUR MY SHIT UP SENPAI

>>60765754
66C is literally nothing.
>>60765763
Don't use these, use >>60764497 instead.

>>60765707
mpv with profile=opengl-hq in mpv.conf

>>60765858
Many thanks, any other suggestions?

>>60765828
>>how would you like your anime
>>JUST OIL PAINT MY SHIT UP
>>JUST OVERSHARP MY SHIT UP
Pick both.

>>60765921
>how would you like your /g/ thread
>JUST MPVSHILL MY SHIT UP

>>60762212
smdegrain on the chroma planes

>>60765921
Oversharpening is when it produces artifacting and/or noise. >>60750404 does not exhibit either.

>>60765946
Idiot, i use madvr and NGU AA. NGU AA looks much better without all of the shit you put on top of it.

>>60765975
hideously warped lineart is artifacting

>>60765980
>NGU AA looks much better without all of the shit you put on top of it.
That's your personal opinion. What isn't an opinion is that nobody has managed to complete the >>60750404 challenge because MadVR > mpv.

File: NGUpixartVeryHigh.png (279KB, 512x447px) Image search: [Google]

279KB, 512x447px

>madvrfags think this looks good

>>60766010
>lineart that isn't blurry is artifacting

>>60766014
>What isn't an opinion is that nobody has managed to complete the >>60750404 challenge
The challenge is to make your picture look like overprocessed shit? For what purpose???

>>60766062
These mpvshill excuses are getting amusing.

>>60766043
video scaling and pixel art scaling aren't comparable

File: comparison - Copy.png (336KB, 1045x653px) Image search: [Google]

336KB, 1045x653px

>>60766043
Its just made for example.

>>60766075
You ask people to make a picture to look like shit and they cant, what a surprise.

>I think Flash’s strong points are “smooth movement and clean lines.” So I think it’s suited for portraying small objects growing larger and larger without losing their shape, or for things that smoothly change shape like water. You see, with traditional animation, you invariably end up with some shaking. It’s possible to make lines in Flash look hand-drawn, but I prefer these clean and simple lines. However, when I say this to Flash specialists, they tell me, “I don’t like that because it’s so Flash-like.” (laughs)
This is a reminder that anime visionary Maasaki Yuasa likes crisp, clean lineart. Blurfags BTFO.

File: muhtestpatterns - Copy.png (1MB, 1807x749px) Image search: [Google]

1MB, 1807x749px

>>60766043

>>60748169
>not using mpv
>using windows unironically
your life must be suffering

>>60766107
>mpvfags actually think left looks better

>>60766137
mpv forever rekt.

>>60766190
>screenshots made by some autist in madvr
>mpv rekt
you madvr fags are forever retrded

>>60766136
clean lines in the source are not the same as artificially smoothed lines by some dipshit on the internet

>>60766222
>artificially smoothed
What's next, forced animation? You can't point out any objectively definable artifacting in that image such as ringing, bloating, etc., so you resort to /v/-tier buzzwords?

>>60766136
Literally who? What does it have to do with blur? Also lines drawn in flash style have nothing to do with your shitty oil painting lines. Main point is youre ruining your source and authors intent in favor of your disgusting preferences. Sad.

>>60766254
Things cant look like shit without artifacts? Thats news for me.

>>60766269
>Literally who?
You don't belong in any discussion even remotely involving anime if you posted this.
>Also lines drawn in flash style have nothing to do with your shitty oil painting lines.
Flash lines are crisp and clear, like the lines in >>60750404
>Main point is youre ruining your source and authors intent
Yeah, I'm sure that anime creators' intent is for their content to be blurry as shit at non-native resolutions. That makes perfect sense.

>>60766311
>blurry as shit
everything looks like blurry shit to you, sounds like you need a glasses

>>60766311
>Flash lines are crisp and clear, like the lines in >>60750404
hahaha, no, flash lines are vectors, they look nothing like your very obviously postprocessed garbage

>>60766406
I get it. You're an image purity fag. You're the picture quality equivalent to retarded audiophiles who refuse to EQ.

>>60766311
>You don't belong in any discussion even remotely involving anime if you posted this.
Name a single good anime he made i dare you.
>Flash lines are crisp and clear, like the lines in
Google flash lines or something, i dunno.
>Yeah, I'm sure that anime creators' intent is for their content to be blurry as shit at non-native resolutions. That makes perfect sense.
Studios release what they release. If they wanted their anime to looks so fucking sharp they'd run some sharpeners over it. And again its not about the blur but mostly your retarded oil painting lines.

>>60766395
This makes sense actually.

>>60766481
>If they wanted their anime to looks so fucking sharp they'd run some sharpeners over it.
You're fundamentally misinterpreting the necessary logic to understand this conversation. AT NATIVE RESOLUTION, meaning viewed on a 720p display, >>60750437 will appear crisp and sharp and require no sharpening. However, viewers viewing anime at non-native resolution is completely out of the anime studios' control. Their artistic intent has nothing to do with how people upscale anime.
The Tatami Galaxy and Ping Pong are two very highly regarded anime by Yuasa, regardless of your personal opinion of them.

>>60762256
Well, to be fair, while NGU AA might smooth over grain a little bit, ewa_lanczossharp sharpens the grain to a certain extent. I don't think either classifies as a perfect grain treatment.

>>60762127
Noise or grain? The key question here is not which image is more noisy, but which image comes closer to the original 1080p Blu-Ray. And NGU Sharp should achieve that, given a high quality source.

Noise by itself isn't automatically bad. What we want is a faithful upscale, not necessary a noise-free upscale. Of course some people prefer noise-free images. In that case it's absolutely ok to run noise reduction algos. But it's not necessarily a good thing for an upscaling algo to remove noise, unless the algo manages to differ between true artifact noise (e.g. compression mosquito noise or block artifacts) and grain.

>>60766556
you trolled yourself
>>60742741
>>60743195
2nd pic wasn't sharpened, only upscaled and downscaled again in mpv

>>60766556
Compare your upscaled images to 1080p BD.
>>60766556
>Ping Pong
Normalfag anime. Kek, i bet you've seen less than 50 anime titles.

>>60766679
You cant argue with that guy using logic.
>SHARP GOOD
>THICC LINES GOOD
>NOISE BAD
>UNGA BUNGA

>>60766694
>Compare your upscaled images to 1080p BD.
anibin has YrYr seasons 1 and 2 listed as 720p lineart. 1080p BD rip will be a typically poor anime studio bilinear/bicubic upscale, meaning blurry as shit.

>>60766759
Thats why you get another anime for a comparison.

>>60766679
>The key question here is not which image is more noisy, but which image comes closer to the original 1080p Blu-Ray.
If anibin lists the show as being 720p lineart, then a typically poor anime studio bilinear/bicubic upscale to 1080p is not exactly a good frame of reference for quality. The whole point of downloading 720p in this case is because MadVR and mpv can do a better job of upscaling than the anime studios.

Why is 'reduce ringing artifacts' so costly when used on old releases ? I mean enabling it makes rendering times go from ~30ms to ~60ms and I start dropping frames.
Though GPU is shit, 1050Ti.

File: Daria S01E01 - Esteemers - restored V2 [JewRye].avi_snapshot_07.37_[2017.06.05_17.11.43].png (430KB, 512x384px) Image search: [Google]

430KB, 512x384px

It makes so little difference, you should only be concerned if you're watching in a 4K monitor or playing DVD to 1080p/4K.
If you're upscaling 720p to 1080p you're better off using image enhancements like luma sharpen, etc. Even 720p to 1440p isn't worth using image doublers.

http://screenshotcomparison.com/comparison/211907

>>60766679
>Noise or grain?
Noise. The testing I was referring to was done with high quality BD rips of digital age anime.

>>60766873
>>60766873
>The testing I was referring to was done with high quality BD rips of digital age anime.
Lol. Explain your testing and tell anime's name.

>>60766855
http://screenshotcomparison.com/comparison/211908
nnedi128 vs NGU AA High.

>>60766855
>It makes so little difference
Some of us believe, by principle, that even small improvements are worth it, and that voluntarily passing on small improvements is logically flawed. You might call it autism; I call it being logically correct.

>>60766916
Flawed logic, if you're that concerned about quality then get the original release bluray/DVD/ISO/remux.

>>60766855
It makes a huge difference for autists apparently. All they do is comparing pixels all day.

>>60766907
>Explain your testing
Already did:
>when I make NGU Sharp and NGU AA equivalently sharp with sharpening settings and video card denoising enabled, NGU Sharp still contains significant noise which is much more apparent in motion than not compared to NGU AA.
>tell anime's name.
[Doki]_Yuru_Yuri_-_01_(1280x720_h264_BD_AAC)_[5B7C4C02]
[Doki]_Nichijou_-_01_(1280x720_Hi10P_BD_AAC)_[C8F9BE9D]
[Chihiro]_K-On!!_-_01_[1280x720_Blu-ray_FLAC][B7458A48]
[mudabone] Hidamari Sketch x Honeycomb - 01 [BD 720p Hi10P H264-AAC] [AEAABE13]
All well-encoded BD rips.

>>60766942
As a consumer I have the right to fundamentally reject absurdly high anime BD prices which are designed to fleece obsessed otaku.

>>60767016
you also have the right to shut the fuck up.

>>60766985
how about some sources that aren't moeshit

File: staythere.jpg (109KB, 924x664px) Image search: [Google]

109KB, 924x664px

>>60767103

>>60766942
If it was made in 720p their upscale will probably be worse, studios aren't autistic with scaling.

>>60767246
>probably
I've never seen a good anime studio upscale. They're either blurry bilinear/bicubic or godawful Q-TEC.
>people think MadVR upscales are oilpainty
No, Q-TEC is oilpainty.

File: [SallySubs] Aria The Animation - 01 [BD 720p AAC] [87D98A49].mkv_snapshot_14.37_[2017.06.05_15.55.46].png (2MB, 960x720px) Image search: [Google]

2MB, 960x720px

>>60767365
>Q-TEC
Why is this allowed?

File: me.png (1012KB, 720x960px) Image search: [Google]

1012KB, 720x960px

>>60767398
kill me

>>60748186
Wait when did vlc stop being the best?

Im from 2009