Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion.

Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules.

 

Go Back   Doom9's Forum > Hardware & Software > Software players

Reply
 
Thread Tools Search this Thread Display Modes
Old 22nd March 2014, 22:47   #25241  |  Link
leeperry
Kid for Today
 
Join Date: Aug 2004
Posts: 3,477
IIRC this is the article that led me to believe that DXVA would play nicer in mVR over 3.0:
Quote:
going from the GPU to CPU there’s a nice 77% increase in bandwidth between PCI Express 3.0 x16 and 2.0 x16 (compared to 89% between 2.0 x16 and 1.0 x16)
Quote:
the OpenCL applications that we were able to test are for now far from being limited by memory bandwidth. In already offering almost 7 GB/s, PCI Express 2.0 x16 covers most usages, even if certain very specific pro applications will certainly be able to take advantage of the theoretical gains.
leeperry is offline   Reply With Quote
Old 22nd March 2014, 23:32   #25242  |  Link
kiccolsd
Registered User
 
Join Date: Mar 2010
Posts: 14
Quote:
Originally Posted by madshi View Post
Those version numbers don't exist.
Sorry I was talking about v0.87.4 vs v0.87.7
kiccolsd is offline   Reply With Quote
Old 23rd March 2014, 03:23   #25243  |  Link
seiyafan
Registered User
 
Join Date: Feb 2014
Posts: 162
Quote:
Originally Posted by e-t172 View Post
IF YOU ARE USING A 780 TI WITH MADVR, PLEASE MAKE YOURSELF KNOWN. Your input would be highly valuable.
My 780 runs fine without frame drops.
seiyafan is offline   Reply With Quote
Old 23rd March 2014, 03:55   #25244  |  Link
StinDaWg
Registered User
 
Join Date: Jan 2014
Posts: 216
Quote:
Originally Posted by e-t172 View Post
To be honest I don't really care about NNEDI3 or other OpenCL-based madVR features at this point. I'm only using "traditional" madVR features such as 3DLUTs, classic upscaling, smooth motion, and the (theoretical in my case) ability to achieve glitch-free silky-smooth playback. I want to fix that first and then maybe look into these shiny new features.
If you watch a lot of 720p content on a 1080p screen you really should be using nnedi. Upscaling quality is way better than Jinc. It's sharper than Lanczos with basically no ringing or aliasing. It's like using a sharpening filter with none of the side effects. The difference is pretty amazing to be honest. Even if your card can only get to Bicubic 75 for chroma and luma doubling at 16 neurons, you are 95% of the way there IMO.
StinDaWg is offline   Reply With Quote
Old 23rd March 2014, 07:05   #25245  |  Link
Qaq
AV heretic
 
Join Date: Nov 2009
Posts: 422
Some madVR users don't really care of its PQ and use madVR only for features.
Qaq is offline   Reply With Quote
Old 23rd March 2014, 08:17   #25246  |  Link
StinDaWg
Registered User
 
Join Date: Jan 2014
Posts: 216
Quote:
Originally Posted by Qaq View Post
Some madVR users don't really care of its PQ and use madVR only for features.
Which of its "features" are not directly related to picture quality?
StinDaWg is offline   Reply With Quote
Old 23rd March 2014, 08:38   #25247  |  Link
Asmodian
Registered User
 
Join Date: Feb 2002
Location: San Jose, California
Posts: 4,406
Quote:
Originally Posted by e-t172 View Post
Thanks. If that's not too much to ask, would you mind testing for frame drops/presentation glitches for 1 hour or so but with madVR defaults (except smooth motion on), and using 60Hz instead of 75? Just trying to make sure our results are comparable. If you don't want to have to reconfigure afterwards you can save your madVR settings from the registry to a .reg file and then restore it afterwards.
I gave this a shot, 1h 20min of anime bluray (24p), 0-0-0 drop-delay-glitch. I reset the stats after switching to full screen because with the default no playback delay some frames are dropped on open and when switching to full screen. Gets up to ~11ms rendering times due to throttling down.

Completely default madVR settings with madVR smooth motion enabled in Zoom Player. I also set my display to 60 Hz.

I don't see any odd usage spikes in GPU, CPU, or IO meters. SLI off.
Asmodian is online now   Reply With Quote
Old 23rd March 2014, 08:44   #25248  |  Link
Q-the-STORM
Registered User
 
Join Date: Sep 2012
Posts: 174
Quote:
Originally Posted by StinDaWg View Post
Which of its "features" are not directly related to picture quality?
display mode switching, IVTC, smooth motion and Exclusive Mode... that's pretty much it...
Q-the-STORM is offline   Reply With Quote
Old 23rd March 2014, 09:18   #25249  |  Link
Asmodian
Registered User
 
Join Date: Feb 2002
Location: San Jose, California
Posts: 4,406
Those all improve motion quality instead of picture quality which seems like a similar goal to me.

@madshi
Is this something easy to implement and actually worth it?

I like profiles a lot but I would like the ability to multiply in the if statement or maybe a single new variable instead, pixels-per-second.
This would help a lot with lower resolution interlaced material, I want to set higher neuron counts for 480p30 than 480p60 but I also don't want very complex or numerous if statements. I could use the pixels-per-second instead of width, height, and deintfps so 480p60 would automatically use my 720p30 profile and 720p60 would automatically use my 1080p30 profile.
Asmodian is online now   Reply With Quote
Old 23rd March 2014, 10:17   #25250  |  Link
StinDaWg
Registered User
 
Join Date: Jan 2014
Posts: 216
Quote:
Originally Posted by Q-the-STORM View Post
display mode switching, IVTC, smooth motion and Exclusive Mode... that's pretty much it...
First and last aren't exclusive to madVR. Pretty much any decent video player will have them, including MPC-HC out of the box. Smooth motion does more harm to the picture than good (IMO). I use IVTC all the time, very handy. Anyways...
StinDaWg is offline   Reply With Quote
Old 23rd March 2014, 15:41   #25251  |  Link
DigitalLF
Registered User
 
Join Date: May 2009
Posts: 120
i just upgraded my GPU from a 5750 to a 280x but my motherboard is a PCI-E 2.0 and when i tried NNEDI3 "64 neurons" it cant handle it but its ok at "32 neurons", material is 1080p24 High-Bitrate and ffdshow is set to output 16bit P016. Could that be that my PCI-E is to slow? Do i need to upgrade? I'm very bad at all this but i really try to learn..

lost frames seems to be way better with P016 then at RGB32 if that makes any sense..... but i get artifacts in windows when i use P016 and NNEDI13.
__________________
Gigabyte G1.Sniper Z87, I7 4770K (3.5ghz), Gigabyte Radeon 280x, Win8.1 x64, ATI 15.7.1, MPC-HC BE 1.4.6 1478, MadVR v0.90.21, LAV Filters 0.68.1,
XySubFilter 3.1.0.746
DigitalLF is offline   Reply With Quote
Old 23rd March 2014, 16:18   #25252  |  Link
StinDaWg
Registered User
 
Join Date: Jan 2014
Posts: 216
madshi, what are your thoughts on using sharpening filters in combination with nnedi doubling? nnedi gives a very nice, sharp, artifact and ringing/aliasing free picture, but I still prefer a bit more sharpness for 720p->1080p scaling. I'm using LumaSharpen at very low values pre-resize, just enough to take out the little bit of softness that is left over from upscaling, nothing more. I think the picture looks pretty much perfect, but I'm wondering if this works against nnedi at all or its just one of those things "do what you think looks best".
StinDaWg is offline   Reply With Quote
Old 23rd March 2014, 16:29   #25253  |  Link
iSunrise
Registered User
 
Join Date: Dec 2008
Posts: 496
@StinDaWg:
When you sharpen something, you basically add ringing. Other than that, it's all personal preference.

Personally, I always use luma doubling in combination with Lanczos 8 AR, which gives me an incredibly sharp and still mostly artifact-free result. I'm also using 1080p as an output device.
iSunrise is offline   Reply With Quote
Old 23rd March 2014, 16:50   #25254  |  Link
StinDaWg
Registered User
 
Join Date: Jan 2014
Posts: 216
Quote:
Originally Posted by iSunrise View Post
@StinDaWg:
When you sharpen something, you basically add ringing. Other than that, it's all personal preference.

Personally, I always use luma doubling in combination with Lanczos 8 AR, which gives me an incredibly sharp and still mostly artifact-free result. I'm also using 1080p as an output device.
So nnedi just uses whatever algorithm you choose and doubles it? I guess I was under the impression that nnedi was it's own upscaling algorithm itself. I'll try Lanczos 8 AR.
StinDaWg is offline   Reply With Quote
Old 23rd March 2014, 16:52   #25255  |  Link
huhn
Registered User
 
Join Date: Oct 2012
Posts: 7,903
the downscaler sharp it.
huhn is offline   Reply With Quote
Old 23rd March 2014, 17:02   #25256  |  Link
iSunrise
Registered User
 
Join Date: Dec 2008
Posts: 496
Quote:
Originally Posted by StinDaWg View Post
So nnedi just uses whatever algorithm you choose and doubles it? I guess I was under the impression that nnedi was it's own upscaling algorithm itself. I'll try Lanczos 8 AR.
NNEEDI is it's own algorithm, but in combination with Lanczos 8 AR, you get the sharpest results, because Lanczos 8 AR is very sharp to begin with. That's only for upscaling purposes though, I still use Catmull-Rom linear light for downscaling.
iSunrise is offline   Reply With Quote
Old 23rd March 2014, 17:12   #25257  |  Link
StinDaWg
Registered User
 
Join Date: Jan 2014
Posts: 216
I tried L8AR/CRAR and L8AR/L8AR and still prefer using LumaSharpen pre-resize. It takes out the softness better than Lanczos scaling alone. One thing is for certain though. nnedi is way better than Jinc when it comes to sharpness/ringing/artifacts.

I'm using
Code:
#define sharp_strength 0.5
#define sharp_clamp 0.05
#define pattern 3
StinDaWg is offline   Reply With Quote
Old 23rd March 2014, 17:22   #25258  |  Link
iSunrise
Registered User
 
Join Date: Dec 2008
Posts: 496
Quote:
Originally Posted by StinDaWg View Post
I tried L8AR/CRAR and L8AR/L8AR and still prefer using LumaSharpen pre-resize. It takes out the softness better than Lanczos scaling alone. One thing is for certain though. nnedi is way better than Jinc when it comes to sharpness/ringing/artifacts.

I'm using
Code:
#define sharp_strength 0.5
#define sharp_clamp 0.05
#define pattern 3
madshi still has some tricks up his sleeve. I expect madshi to provide an even more sophisticated (or rather, more optimized for good results) sharpener, once he gets to it. Same thing with an "add grain" algorithm.

With madshi's AR algorithm, possibly even NNEDI could be improved, which I suggested before. But we would need to test that.

And yes, I also prefer NNEDI (but NNEDI only works for powers of 2) > Lanczos AR > Jinc AR. But I expect that to be highly dependent on the content and your display setup. And of course, your personal preference.

Last edited by iSunrise; 23rd March 2014 at 17:30.
iSunrise is offline   Reply With Quote
Old 23rd March 2014, 22:01   #25259  |  Link
seiyafan
Registered User
 
Join Date: Feb 2014
Posts: 162
How does NNEDI in combo with Jinc3AR compare to NNEDI with Lanczos ?
seiyafan is offline   Reply With Quote
Old 23rd March 2014, 23:17   #25260  |  Link
dansrfe
Registered User
 
Join Date: Jan 2009
Posts: 1,210
Just to make sure, there's no current workaround for using NNEDI on nvidia other than downgrading the driver?

Last edited by dansrfe; 24th March 2014 at 01:16.
dansrfe is offline   Reply With Quote
Reply

Tags
direct compute, dithering, error diffusion, madvr, ngu, nnedi3, quality, renderer, scaling, uhd upscaling, upsampling

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 20:55.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.