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 6th December 2012, 10:43   #16021  |  Link
dansrfe
Registered User
 
Join Date: Jan 2009
Posts: 1,210
While DXVA2 is selected as the decoder in LAV Video and a file is playing, if MPC-HC's window is dragged to another screen then the video size resizes to what MPC-HC originally started with. This doesn't happen when software decoding is selected though.
dansrfe is offline   Reply With Quote
Old 6th December 2012, 11:26   #16022  |  Link
Toku
Registered User
 
Join Date: Jan 2012
Posts: 14
Just noticed that It's pretty difficult to see progress in the FSE seekbar when you're a bit back from the monitor unless the screen is pretty dark. Is it possible to get the progress color darkened or something? Obviously isn't anything major, and I seem to remember you said that you wouldn't rethink the presentation until v 1.0, but just thought I'd mention it.

Last edited by Toku; 6th December 2012 at 11:54.
Toku is offline   Reply With Quote
Old 6th December 2012, 11:26   #16023  |  Link
THX-UltraII
Registered User
 
Join Date: Aug 2008
Location: the Netherlands
Posts: 851
Since that madVR is very mature already I think it might be usefull that a WIKI / manual will be created for madVR. I do not have the knowledge for this unfortunately.
THX-UltraII is offline   Reply With Quote
Old 6th December 2012, 11:46   #16024  |  Link
pandy
Registered User
 
Join Date: Mar 2006
Posts: 1,049
I have small idea - perhaps not very smart - is there any chance to add display support for YCbCr data without conversion to RGB on RGB output - this can provide capability to change RGB analog output to YPbPr output - very simple adapter (perhaps some buffer only, perhaps only some caps + sync combiner) should be enough to connect any graphic card to YPbPr input of TV (one one typical video modes need to be set anyway) - in normal conditions people need to buy/build ( VGA RGB to YPbPr converter ) special converter from RGB to YPbPr. With such settings - i.e. RGB used as YPbPr output only sync combiner will be required.
pandy is offline   Reply With Quote
Old 6th December 2012, 12:00   #16025  |  Link
petri234
Registered User
 
Join Date: Nov 2012
Posts: 2
Quote:
Originally Posted by madshi View Post


Does this occur with all 23.976p movies for you, or just with some? Does it e.g. depend on the codec? Have you updated to the latest LAV version? I've not heard about such instability by anyone else yet, so I wonder why it only seems to affect you. Maybe it's just with some specific video files?

I checked, and it did indeed occur with every movie that is 1080/23.976p. Updating to the latest LAV filters did the trick though, so it must have been on their end. Thanks!
petri234 is offline   Reply With Quote
Old 6th December 2012, 12:31   #16026  |  Link
hannes69
Registered User
 
Join Date: Nov 2012
Posts: 99
Quote:
Can you guys please test the following:

(1) Use software decoding and e.g. Bilinear scaling.
(2) Use native DXVA2 decoding and e.g. Bilinear scaling.
(3) Use software decoding and DXVA2 scaling.
(4) Use native DXVA2 decoding and DXVA2 scaling.
I use AMD HD4550 with Windows7/64bit.
Colors in (2), (3), (4) are the same but different from (1). Especially green is much too oversaturated. (1) stays the same by using copy-back-decoding instead of software.
DXVA-Deinterlacing works as intended, no change in colors.
Interesting: using native decoding instead of copy-back lowers cpu load as expected but as well gpu load! Furthermore dxva2 upscaling has about the same gpu load as Mitchell-Netravali.
The quote refers to 0.85.1

Redone the test with 0.85.2:
Now (1), (3) and (4) are the same, (2) is different (in the same way with oversaturated green like in 0.85.1). So using native decoding without dxva scaling is still broken (at least for my setup).
With 0.85.2 GPU load with native DXVA decoding is about 0 - 30% HIGHER than copyback decoding (depending on scaling algorithm!)

Last edited by hannes69; 6th December 2012 at 13:15.
hannes69 is offline   Reply With Quote
Old 6th December 2012, 14:02   #16027  |  Link
truexfan81
Registered User
 
truexfan81's Avatar
 
Join Date: Nov 2012
Posts: 138
Quote:
That is very weird, I've no explanation for that. Do you have an SSE4.1 capable CPU? Are you sure that cuvid deinterlacing is done with highest possible quality? You could double check by using the cheese slices test video. Most important of all: When you get that slideshow, which of the queues are empty and which are full? You can see that in the madVR debug OSD (Ctrl+J).
Madshi
this is my cpu
and now that i have complained about it, it appears to be acting like when you take a car to the mechanic, it just won't make that same sound or in this case, a slide-show. i will report back later if it starts doing it again.

truexfan81 is offline   Reply With Quote
Old 6th December 2012, 14:20   #16028  |  Link
nevcairiel
Registered Developer
 
Join Date: Mar 2010
Location: Hamburg/Germany
Posts: 10,342
AMD CPUs only support SSE 4.1 in Bulldozer and above, so no, your CPU does not qualify.
__________________
LAV Filters - open source ffmpeg based media splitter and decoders
nevcairiel is offline   Reply With Quote
Old 6th December 2012, 14:31   #16029  |  Link
leeperry
Kid for Today
 
Join Date: Aug 2004
Posts: 3,477
Quote:
Originally Posted by madshi View Post
Both CUDA and OpenCL (1.2) have interop functionality to Direct3D9 which in theory allows accessing D3D textures. However, it only half works at the current time and is very slow with current AMD drivers. But there's hope that this might be improved in the future.
Does that imply that Jinc might be less demanding on nvidia boards someday in the future?
leeperry is offline   Reply With Quote
Old 6th December 2012, 14:57   #16030  |  Link
phoenixxl
Registered User
 
Join Date: Aug 2011
Posts: 21
Language

When Madvr crashed earlier today , I noticed this:

http://i.imgur.com/aA090.png

However , the correct value should be this :

http://i.imgur.com/y8zVD.png


Please consider revisiting the code that determines said value.

Friendly regards,

Phoenixxl.

Last edited by phoenixxl; 6th December 2012 at 15:02.
phoenixxl is offline   Reply With Quote
Old 6th December 2012, 15:44   #16031  |  Link
aufkrawall
Registered User
 
Join Date: Dec 2011
Posts: 1,812
Quote:
Originally Posted by madshi View Post
Well, I guess it would be possible. But I somehow doubt the usefulness of using DXVA chroma upsampling with non-DXVA luma upscaling. So I wonder whether I should invest the development time to make this combination work.
What if you watch a movie in native resolution? Then it will use the selected chroma scaling method (by default bilinear), right?.

If true, that's why I think an option to select DXVA chroma scaling for such situations would be a good idea since it looks better than bilinear and doesn't require (m)any ressources.
aufkrawall is offline   Reply With Quote
Old 6th December 2012, 15:51   #16032  |  Link
sneaker_ger
Registered User
 
Join Date: Dec 2002
Posts: 5,565
Quote:
Originally Posted by madshi View Post
That's bad. I'll try to reproduce this on my development PC. You're using win7, right?
Yes, Win 7 x64.

The question is, why am I the only user to report this? I can't be the only ATI user on Win 7. So chances are you won't be able to reproduce it, I guess.
sneaker_ger is offline   Reply With Quote
Old 6th December 2012, 15:57   #16033  |  Link
ajp_anton
Registered User
 
ajp_anton's Avatar
 
Join Date: Aug 2006
Location: Stockholm/Helsinki
Posts: 805
Quote:
Originally Posted by madshi View Post
DXVA scaling is only active if the video needs to be scaled. If DXVA scaling is active, chroma upsampling is being done by DXVA, too. If the video doesn't need to be scaled, chroma is still upsampled by the madVR algorithms.
Why? Isn't chroma upscaled all the way up to full resolution (and converted to RGB) by DXVA? If the uses chooses DXVA, why change it when the luma scaling factor happens to be 1?
ajp_anton is offline   Reply With Quote
Old 6th December 2012, 16:13   #16034  |  Link
6233638
Registered User
 
Join Date: Apr 2009
Posts: 1,019
Madshi, I've just seen a new bug introduced in 0.85.2

If you are downscaling with DXVA2 and switch to anything else, it results in a green image. I have to switch to a non-DXVA2 scaling algorithm a second time to restore the correct colours.


Quote:
Originally Posted by aufkrawall View Post
What if you watch a movie in native resolution? Then it will use the selected chroma scaling method (by default bilinear), right?.

If true, that's why I think an option to select DXVA chroma scaling for such situations would be a good idea since it looks better than bilinear and doesn't require (m)any ressources.
If you have an Nvidia card, DXVA2 scaling is just bilinear with a sharpening filter applied, and it comes at a significant performance hit compared to using just about anything else. Selecting "bilinear" in madVR runs about 10x faster, and looks better. (no ringing or other artefacts introduced)

To get a basic idea of what kind of scaling DXVA2 is using on your system, right click and save this image

If you open it in MadVR and switch between DXVA2 and Bilinear scaling on an Nvidia card, there is no difference between the two, and a clear difference when using anything else, for example.


As a side note, there are some interesting things that happen when using the anti-ringing filter and/or linear light scaling with this test, and I wonder if something like this could possibly be used to help fine-tune the algorithms further. (it was also interesting to use a multi-coloured one to see how they were blended together) Then again, it's a completely artificial thing that may have no bearing on real-world content at all, so maybe it's pointless.

Still, it's an easy way to get an idea of what scaling algorithm DXVA2 is using, through comparisons.
6233638 is offline   Reply With Quote
Old 6th December 2012, 16:50   #16035  |  Link
DragonQ
Registered User
 
Join Date: Mar 2007
Posts: 934
Hmm it appears my new laptop comes with an nVidia NVS 5200M, as well as the Intel HD 4000 IGP. There's very little information about this GPU out there so I guess I'll have to investigate which one is best for MadVR!
__________________
TV Setup: LG OLED55B7V; Onkyo TX-NR515; ODroid N2+; CoreElec 9.2.7
DragonQ is offline   Reply With Quote
Old 6th December 2012, 17:02   #16036  |  Link
wanezhiling
Registered User
 
Join Date: Apr 2011
Posts: 1,184
NVS 5200M is based on GT435M(GF108), a weak GPU.
wanezhiling is offline   Reply With Quote
Old 6th December 2012, 17:35   #16037  |  Link
DragonQ
Registered User
 
Join Date: Mar 2007
Posts: 934
Yeah but the Intel HD 4000 isn't exactly a beast either.
__________________
TV Setup: LG OLED55B7V; Onkyo TX-NR515; ODroid N2+; CoreElec 9.2.7
DragonQ is offline   Reply With Quote
Old 6th December 2012, 22:34   #16038  |  Link
DragonQ
Registered User
 
Join Date: Mar 2007
Posts: 934
Madshi, how do I submit a freeze report? I'm getting freezes using the NVS 5200M in my laptop with MadVR (no issues using the Intel HD 4000 with MadVR, or either card with EVR).
__________________
TV Setup: LG OLED55B7V; Onkyo TX-NR515; ODroid N2+; CoreElec 9.2.7
DragonQ is offline   Reply With Quote
Old 7th December 2012, 00:21   #16039  |  Link
DarkSpace
Registered User
 
Join Date: Oct 2011
Posts: 204
Quote:
Originally Posted by 6233638 View Post
To get a basic idea of what kind of scaling DXVA2 is using on your system, right click and save this image
Quote:
Originally Posted by 6233638 View Post
As a side note, there are some interesting things that happen when using the anti-ringing filter and/or linear light scaling with this test, and I wonder if something like this could possibly be used to help fine-tune the algorithms further. (it was also interesting to use a multi-coloured one to see how they were blended together) Then again, it's a completely artificial thing that may have no bearing on real-world content at all, so maybe it's pointless.
Those images also show these "gremlins" that madshi noticed in the Jinc AR algorithms, though they're more pronounced when scaling in Linear Light and using 8 taps. Also, am I the only one who thinks that Linear Light upscaling with Jinc (AR or not) looks worse* on these images?

* I don't really know, but I was expecting at least some kind of darker grey, Linear Light scaling makes it look as if there's hardly and black in the image, at least for the black-and-white checkerboard pattern. I may be wrong, though.

Edit: Looks like I'm getting Bilinear upscaling on AMD Radeon HD 6970M when scaling with DXVA. However, I didn't pay too close attention to things like additional sharpening algorithms.

Last edited by DarkSpace; 7th December 2012 at 00:24.
DarkSpace is offline   Reply With Quote
Old 7th December 2012, 00:33   #16040  |  Link
vivan
/人 ◕ ‿‿ ◕ 人\
 
Join Date: May 2011
Location: Russia
Posts: 643
DragonQ, try to untick "use a separate device for presentation" in madVR settings. This happens (image freezes after some time, not madVR itself) on all Optimus systems, when using dGPU.
vivan 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 13:18.


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