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 14th April 2009, 21:32   #281  |  Link
yesgrey
Registered User
 
Join Date: Sep 2004
Posts: 1,295
Quote:
Originally Posted by yesgrey3 View Post
I will look into this...
Quote:
Originally Posted by leeperry View Post
that'd be pretty cool to get perfect colorimetry in mVR, hope you can work it out!
I've been looking into this and there is not any problem with cr3dlut or with madVR.
I've compared leeperry's settings with cr3dlut 8bit and 16bit output 3DLUT files using t3dlut and madVR (only the 16bit), and with ddcc 3D LUT files with the same settings, and the images all look the same.
In fact, there is a small difference, between some of the images, but only because those images use 3D LUTs created with different settings. These images have the suffix "leeperry".
You can compare the results here.

So, what was the problem?
In his original settings, he was using a custom gamma curve for removing and reapplying gamma, for the color gamut correction, and with the new settings, he was using the standard curves. So, the color gamut correction is slightly different, but the more accurate should be the the newest.
yesgrey is offline   Reply With Quote
Old 14th April 2009, 21:38   #282  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,137
Quote:
Originally Posted by chros View Post
- support for the built-in subtitle engine (like Haali's Renderer) (important !!!)
- if possible:built-in shader support
Already on the to do list.

Quote:
Originally Posted by chros View Post
question: isn't dxva support available at all by design ?
It's not available at all due to design.

Quote:
Originally Posted by Egh View Post
Seems though when enabling fullscreen it first shows it as entire screen, and only then applies aspect ratio correction.
Yes, not sure if I can improve that. That's probably just the way the events come in...

Quote:
Originally Posted by yesgrey3 View Post
I've been looking into this and there is not any problem with cr3dlut or with madVR.
I've compared leeperry's settings with cr3dlut 8bit and 16bit output 3DLUT files using t3dlut and madVR (only the 16bit), and with ddcc 3D LUT files with the same settings, and the images all look the same.
Thanks! I'm relieved...
madshi is offline   Reply With Quote
Old 14th April 2009, 21:47   #283  |  Link
leeperry
Kid for Today
 
Join Date: Aug 2004
Posts: 3,463
Quote:
Originally Posted by yesgrey3 View Post
I've been looking into this and there is not any problem with cr3dlut or with madVR.
I've compared leeperry's settings with cr3dlut 8bit and 16bit output 3DLUT files using t3dlut and madVR (only the 16bit), and with ddcc 3D LUT files with the same settings, and the images all look the same.
In fact, there is a small difference, between some of the images, but only because those images use 3D LUTs created with different settings. These images have the suffix "leeperry".
You can compare the results here.

So, what was the problem?
In his original settings, he was using a custom gamma curve for removing and reapplying gamma, for the color gamut correction, and with the new settings, he was using the standard curves. So, the color gamut correction is slightly different, but the more accurate should be the the newest.
complaining that it doesn't work in public makes it sound like I'm b*tching, I was simply wondering if anyone else tried to set custom primaries...nothing else.

I've made compares w/ rec709.mkv and just dropped you a PM.

for your answer!

Last edited by leeperry; 14th April 2009 at 23:01.
leeperry is offline   Reply With Quote
Old 14th April 2009, 23:01   #284  |  Link
cyberbeing
Broadband Junkie
 
Join Date: Oct 2005
Posts: 1,859
Quote:
Originally Posted by leeperry View Post
I was simply wondering if anyone else tried to set custom primaries...nothing else.
I have been ever since madVR was first released and haven't noticed any problems. madVR = t3dlut = ddcc.
cyberbeing is offline   Reply With Quote
Old 14th April 2009, 23:22   #285  |  Link
leeperry
Kid for Today
 
Join Date: Aug 2004
Posts: 3,463
Quote:
Originally Posted by cyberbeing View Post
I have been ever since madVR was first released and haven't noticed any problems. madVR = t3dlut = ddcc.
well mVR=t3dlut YES! I messed up and put ConvertToRGB32() instead of ConvertToYUY2()...my bad

but I don't get the same colors from the original PS script and rgb3dlut()

I just want plain 1/0.45 primaries correction...I will try ddcc in realtime now, this gave me the same colors as the PS script a while ago

PS: ddcc in realtime = PS script...exactly what I want! but when I go LUT it goes really wrong I think

Last edited by leeperry; 14th April 2009 at 23:33.
leeperry is offline   Reply With Quote
Old 14th April 2009, 23:43   #286  |  Link
flanger216
Registered User
 
Join Date: Mar 2004
Posts: 140
Quote:
Originally Posted by leeperry View Post
I've always been convinced that tearing was due to some bad BIOS design/user misconfiguration...mostly the video data is not copied fast enough from the northbridge I think.

anyway, I've tried numerous mobos/graphic cards...never had any tearing on XP or Vista.
I used to own a Samsung monitor that teared on everything - video games, movies, even the Windows desktop. The only way I fixed it was to fire up a tearing test, load Powerstrip and mess with the refresh rate at hundreths of a hertz. Eventually I found that tearing went away when I changed the refresh rate from ~59.9hz to ~59.8hz. Weird, huh?

Quote:
Originally Posted by Egh View Post
Wheeeeeeee! So I'm not the only person in the world who thinks that "tearing" etc is most overrated issue in video ever I had 7600 and 7900, on completely different systems, on old Athlon and much newer E8500, on XP and on XP64, and yet to see any tearing on video whatsoever
Gee, so you think a problem you've never had is "overrated." And, of course, your sample size of two completely different systems is astonishingly comprehensive. Personally, I think the bubonic plague is 'overrated,' based on the sole fact that I've never had it.
flanger216 is offline   Reply With Quote
Old 14th April 2009, 23:47   #287  |  Link
leeperry
Kid for Today
 
Join Date: Aug 2004
Posts: 3,463
that's the original untouched file(RGB32HQ 0-255 in VMR9) :



that's the PS script on top of it :




that's ddcc 1.9 in realtime (identical to the PS script!) :



Code:
colorYUV(levels="tv->pc")
ConvertToRGB32(matrix="PC.709")
ddcc(chr_i=3,gam_i=5,ofile="C:\HC3100.txt",threads=4,opt=1)
then I created a LUT with
Code:
ddcc(chr_i=3,gam_i=5,ofile="C:\HC3100.txt",lutfile="Z:\lut.txt",threads=4,opt=1)
then played it w/ :
Code:
colorYUV(levels="tv->pc")
ConvertToRGB32(matrix="PC.709")
rgb3dlut(lutfile="z:\lut.txt",threads=4)
and I get this :



so maybe I'm doing something really wrong...or rgb3dlut is mixing YUY2 and RGB LUT's, or well I dunno...I'm off to watch a movie w/ ddcc in realtime in HR is what I know

I'm getting the same sort of completely burned up green w/ yesgrey's app/t3dlut, considering they're using the same code..

Last edited by leeperry; 15th April 2009 at 00:19.
leeperry is offline   Reply With Quote
Old 14th April 2009, 23:59   #288  |  Link
leeperry
Kid for Today
 
Join Date: Aug 2004
Posts: 3,463
Quote:
Originally Posted by flanger216 View Post
I used to own a Samsung monitor that teared on everything - video games, movies, even the Windows desktop. The only way I fixed it was to fire up a tearing test, load Powerstrip and mess with the refresh rate at hundreths of a hertz. Eventually I found that tearing went away when I changed the refresh rate from ~59.9hz to ~59.8hz. Weird, huh?
hah, I don't have ANY experience w/ LCD screens..can't comment on this, I was mostly talking about tearing created by the graphic card itself(and I tried several generations of ati/nvidia on many different mobos)

DX can't measure the refresh rate accurately, so maybe it was only meant to work at 59.940000....but DX couldn't measure it accurately enough

Kazuya's Sanyo Z4 stutters w/ 50.000 in Reclock, but works fine if he uses pstrip's search engine to find the perfect "theoritical" 50.000000000Hz(that happens to give 50.002 in Reclock )...if your display is not multisync AT ALL, I guess that feature in pstrip is unmatched

Last edited by leeperry; 15th April 2009 at 00:12.
leeperry is offline   Reply With Quote
Old 15th April 2009, 00:05   #289  |  Link
racerxnet
Registered User
 
Join Date: Jul 2004
Location: ILLINIOS
Posts: 50
Quote:
Originally Posted by ericd View Post
please can anybody post a kind of tutorial to use these video renders ,cause some people like me are newbies ,thx a lot....

eric

Here is my short attempt for a guide for SD DVD playback as the bulk of movies (700+) are this format:

1 Download DVD43 which is a on the fly decrypter similar to anydvd. You will need to rip the protection or MadVR will not start. I tried to add the decrypter pack in external filters but it did not work. http://www.soft82.com/download/windows/dvd43/

2 download MPC-hc: http://www.xvidvideo.ru/content/category/1/1/2/

3 Download FFdshow: http://ffdshow-tryout.sourceforge.net/

4 Download MadVR from this site.


Install DVD 43 so that you can decrypt on the fly as Anydvd does. This is only for SDdvd's and not BL or HDdvd.

Install MPC-HC

Install MadVR and click on install.bat

Install FFdshow: select codecs on the left column and scroll to the bottom. Select Raw video and set to all supported. Scroll up to the mpeg2 setting and select libmpeg2.

Back in the left hand column select Resize and Aspect: check the resize button. Highlight the radio button marked specify horizontal size, no aspect ratio control and resize always. Check the process pixel aspect ratio internally.

Move down to the Overlay/output and check YV12, YUY2, YVYU,UYVY,NV12, select closest matching colorspace,set pixel aspect ratio in output media type.

In RGB conversion select BT601, Input levels standard ( if you are calibrated to 16-235) output : TV/projector.

InMPC_HC uncheck all internal filters. Add FFdshow audio decoder, ffdshow video decoder, mpeg2 splitter, and madVR and set all to prefer. Insert DVD and select the VOB to start the movie. You will get a "Macrovision failure" if you try to start from the "Open disk" bar in MPC-HC.

Hope this helps some newbies and a big thanks to Madishi for this great renderer.


MAK

The most important settings are the ones in resize and aspect. If you get no video you have done something wrong.


I hope that madishi can incorporate or explain how to decrypt without other software for SDdvd's. DeCSS has been out for some time and could be added to the code.
racerxnet is offline   Reply With Quote
Old 15th April 2009, 00:16   #290  |  Link
noee
Registered User
 
Join Date: Jan 2007
Posts: 530
If you're going to use FFdshow as your decoder, choose YV12 as the output and ignore the settings for RGB. In addition, I would *not* use FFDshow resize if using madVR. That's really part of the whole point of this renderer.
noee is offline   Reply With Quote
Old 15th April 2009, 01:38   #291  |  Link
flanger216
Registered User
 
Join Date: Mar 2004
Posts: 140
Quote:
Originally Posted by leeperry View Post
hah, I don't have ANY experience w/ LCD screens..can't comment on this, I was mostly talking about tearing created by the graphic card itself(and I tried several generations of ati/nvidia on many different mobos)

DX can't measure the refresh rate accurately, so maybe it was only meant to work at 59.940000....but DX couldn't measure it accurately enough

Kazuya's Sanyo Z4 stutters w/ 50.000 in Reclock, but works fine if he uses pstrip's search engine to find the perfect "theoritical" 50.000000000Hz(that happens to give 50.002 in Reclock )...if your display is not multisync AT ALL, I guess that feature in pstrip is unmatched
Seriously... I nearly leapt for joy when pstrip finally fixed it. But yeah, I agree - probably a rounding error introduced somewhere in the chain, whereas the LCD processor might've needed an exact sync. Of course, I've also come across a few LCD monitors (including Samsung) that actually had bogus EDID info, so they'd provide the graphics adapter with incorrect native refresh-rates

Blech. Consumers LCD screens are just disasters in motion. Speaking as an HTCP enthusiast, I find them unsatisfactory; speaking as a photographer, I find them loathsome
flanger216 is offline   Reply With Quote
Old 15th April 2009, 01:52   #292  |  Link
leeperry
Kid for Today
 
Join Date: Aug 2004
Posts: 3,463
Quote:
Originally Posted by flanger216 View Post
Seriously... I nearly leapt for joy when pstrip finally fixed it. But yeah, I agree - probably a rounding error introduced somewhere in the chain, whereas the LCD processor might've needed an exact sync. Of course, I've also come across a few LCD monitors (including Samsung) that actually had bogus EDID info, so they'd provide the graphics adapter with incorrect native refresh-rates

Blech. Consumers LCD screens are just disasters in motion. Speaking as an HTCP enthusiast, I find them unsatisfactory; speaking as a photographer, I find them loathsome
anything that doesn't support a multiple of 24p is not worth using for a HTPC.
yeah Seb.26 really had to persuade Rik Wang to add this feature, because it asks the GPU about the refresh rate...not windows

if mastering studios still stick to CRT, there's a very good reason...best CR, multisync, etc etc

bogus EDID was not too hard in my experience, simply had to check "allow unsupported resolutions"...reboot a few times and tada! the Sanyo Z2 EDID says "max res: XGA" so sending 720p is kinda awkward

I'll tell tritical know about the LUT issue in ddcc()

Last edited by leeperry; 15th April 2009 at 02:00.
leeperry is offline   Reply With Quote
Old 15th April 2009, 02:17   #293  |  Link
yesgrey
Registered User
 
Join Date: Sep 2004
Posts: 1,295
Quote:
Originally Posted by leeperry View Post
PS: ddcc in realtime = PS script...exactly what I want! but when I go LUT it goes really wrong I think
leeperry is right. The 3D LUT is giving different results compared with the PS script.
I have already found the cause, and changed cr3dlut to give the same results as the PS script, but I'm still investigating it to see which is the more accurate...

EDIT: This only affects the color gamut correction. So, the majority should be ok with the current version. I will try to release a fix tomorrow...

Last edited by yesgrey; 15th April 2009 at 02:30.
yesgrey is offline   Reply With Quote
Old 15th April 2009, 02:32   #294  |  Link
leeperry
Kid for Today
 
Join Date: Aug 2004
Posts: 3,463
Quote:
Originally Posted by yesgrey3 View Post
leeperry is right. The 3D LUT is giving different results compared with the PS script.
I have already found the cause, and changed cr3dlut to give the same results as the PS script, but I'm still investigating it to see which is the more accurate...
ok thanks for the confirmation, and the new cr3dlut
well I thought you checked the PS script code w/ matlab and stuff? it looks great, and not GREEN...me like
leeperry is offline   Reply With Quote
Old 15th April 2009, 04:56   #295  |  Link
racerxnet
Registered User
 
Join Date: Jul 2004
Location: ILLINIOS
Posts: 50
May be a Bug:

Unless the movie is ripped with the main audio stream, disc playback includes the directors comments. I have not been able to find a way around this with the original disk yet for SDdvd.

Thanks,

MAK
racerxnet is offline   Reply With Quote
Old 15th April 2009, 07:41   #296  |  Link
Snowknight26
Registered User
 
Join Date: Aug 2007
Posts: 1,394
Quote:
Originally Posted by racerxnet View Post
Unless the movie is ripped with the main audio stream, disc playback includes the directors comments. I have not been able to find a way around this with the original disk yet for SDdvd.
Sounds like a splitter issue.

madshi, I'm not sure if this would be classified as a MPC-HC issue or a madVR issue, but whenever you try to take a screenshot while using madVR (only tested it with the non-special build), an error appears about GetCurrentImage failing with error code 80070057.
Snowknight26 is offline   Reply With Quote
Old 15th April 2009, 08:07   #297  |  Link
ericd
Registered User
 
Join Date: Jan 2009
Posts: 25
Quote:
Originally Posted by racerxnet View Post
Here is my short attempt for a guide for SD DVD playback as the bulk of movies (700+) are this format:

1 Download DVD43 which is a on the fly decrypter similar to anydvd. You will need to rip the protection or MadVR will not start. I tried to add the decrypter pack in external filters but it did not work. http://www.soft82.com/download/windows/dvd43/

2 download MPC-hc: http://www.xvidvideo.ru/content/category/1/1/2/

3 Download FFdshow: http://ffdshow-tryout.sourceforge.net/

4 Download MadVR from this site.


Install DVD 43 so that you can decrypt on the fly as Anydvd does. This is only for SDdvd's and not BL or HDdvd.

Install MPC-HC

Install MadVR and click on install.bat

Install FFdshow: select codecs on the left column and scroll to the bottom. Select Raw video and set to all supported. Scroll up to the mpeg2 setting and select libmpeg2.

Back in the left hand column select Resize and Aspect: check the resize button. Highlight the radio button marked specify horizontal size, no aspect ratio control and resize always. Check the process pixel aspect ratio internally.

Move down to the Overlay/output and check YV12, YUY2, YVYU,UYVY,NV12, select closest matching colorspace,set pixel aspect ratio in output media type.

In RGB conversion select BT601, Input levels standard ( if you are calibrated to 16-235) output : TV/projector.

InMPC_HC uncheck all internal filters. Add FFdshow audio decoder, ffdshow video decoder, mpeg2 splitter, and madVR and set all to prefer. Insert DVD and select the VOB to start the movie. You will get a "Macrovision failure" if you try to start from the "Open disk" bar in MPC-HC.

Hope this helps some newbies and a big thanks to Madishi for this great renderer.


MAK

The most important settings are the ones in resize and aspect. If you get no video you have done something wrong.


I hope that madishi can incorporate or explain how to decrypt without other software for SDdvd's. DeCSS has been out for some time and could be added to the code.



thank you very much sir
ericd is offline   Reply With Quote
Old 15th April 2009, 10:03   #298  |  Link
chros
Registered User
 
chros's Avatar
 
Join Date: Mar 2002
Posts: 1,507
Quote:
Originally Posted by madshi View Post
Already on the to do list.
Thanks, we're waiting for this ...

Bugfix request: multiple monitor support ...

Thank you for your work ...!
__________________
Ryzen 5 2600,Asus Prime b450-Plus,16GB,MSI GTX 1060 Gaming X 6GB(v385.28),Win10 LTSB 1607,MPC-BEx64+LAV+MadVR,Yamaha RX-A870,LG OLED65B8(2160p@23/24/25/29/30/50/59/60Hz)
chros is offline   Reply With Quote
Old 15th April 2009, 10:33   #299  |  Link
ericgur
Registered User
 
Join Date: Dec 2001
Location: Israel
Posts: 34
Color Space Conversion & gamma correction

Hi madshi,
I was wondering why you chose to use 3D LUTs for CSC and gamma correction.

Performing CSC using 3D LUTs is a major overkill and should be significantly faster by performing the calculations in a shader.
A simple CSC (YCbCr BT601/709 to R'G'B') involves 7 multiplications, 5 additions/subtractions and a 3 clip/saturate operations. These do not involve expensive random memory reads that are performed in large 3D LUTs where the GPU caches are almost useless.

R'G'B' gamma corrections can be performed by 3 1D LUTs (one for each color component) and performed by a shader only when requested (most users will probably not use it as the display pipeline should be calibrated to do this in first place via an ICC profile).

3D LUTs are normally used for complex Gamut correction (mostly for printing purposes) and are (almost) never used in professional video processing HW.

Using 3 1D LUTs for RGB gamma correction gives the following advantages:
* Use less GPU resources - tables should be small enough to fit in the GPU caches (3x16bitx16bit or 3x10bitx16bit or 3x10bitx8bit).
* GPU memory traffic is reduced.
* Real time manipulation of LUTs (with GUI).
* Quick load time - very important when viewing several video files sequentially.

Writing shaders for CSC will also allow adding features like contract brightness, saturation and hue modifications with a small complication to the shader code - perform full 3x3 matrix multiplication (9 multiplications, 12 additions, 3 clips). All of these can be easily modified in realtime.
ericgur is offline   Reply With Quote
Old 15th April 2009, 10:44   #300  |  Link
tetsuo55
MPC-HC Project Manager
 
Join Date: Mar 2007
Posts: 2,317
Quote:
Originally Posted by ericgur View Post
Hi madshi,
I was wondering why you chose to use 3D LUTs for CSC and gamma correction.

Performing CSC using 3D LUTs is a major overkill and should be significantly faster by performing the calculations in a shader.
A simple CSC (YCbCr BT601/709 to R'G'B') involves 7 multiplications, 5 additions/subtractions and a 3 clip/saturate operations. These do not involve expensive random memory reads that are performed in large 3D LUTs where the GPU caches are almost useless.

R'G'B' gamma corrections can be performed by 3 1D LUTs (one for each color component) and performed by a shader only when requested (most users will probably not use it as the display pipeline should be calibrated to do this in first place via an ICC profile).

3D LUTs are normally used for complex Gamut correction (mostly for printing purposes) and are (almost) never used in professional video processing HW.

Using 3 1D LUTs for RGB gamma correction gives the following advantages:
* Use less GPU resources - tables should be small enough to fit in the GPU caches (3x16bitx16bit or 3x10bitx16bit or 3x10bitx8bit).
* GPU memory traffic is reduced.
* Real time manipulation of LUTs (with GUI).
* Quick load time - very important when viewing several video files sequentially.

Writing shaders for CSC will also allow adding features like contract brightness, saturation and hue modifications with a small complication to the shader code - perform full 3x3 matrix multiplication (9 multiplications, 12 additions, 3 clips). All of these can be easily modified in realtime.
Would the conversion still be 64bit float?
tetsuo55 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 09:26.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2019, vBulletin Solutions Inc.