.ti3 Parser / .3dlut Generator for madVR + .3dlut toolset - Page 3

nand chan · 24th August 2011, 15:37

Quote:

Originally Posted by cyberbeing

You probably forgot to specify a hardware device context. The following I assume are the relevant imported functions for what you want to do, since all lut loaders contain them.

User32.dll
GetMonitorInfoA
EnumDisplayMonitors
MonitorFromWindow

GDI32.dll
CreateDCA
GetICMProfileA
SetDeviceGammaRamp
GetDeviceGammaRamp
DeleteDC

Done.

leeperry · 24th August 2011, 15:47

Quote:

Originally Posted by nand chan

For the curious, a comparison of 16-bit LUT vs 8-bit LUT (to make madVR load it, I first converted it down to 8-bit and then converted that back up to 16-bit which is simply a bit-shift, the same way madVR converts levels). You can clearly see the advantage of greater-than-8-bit precision processing even on 8-bit devices. So if you want to use a .3dlut with avisynth be wary of this difference:

http://screenshotcomparison.com/comparison/75573

fair enough, but here's a real world banding test pattern going through LAV CUVID: Banding 720p.rec709.mkv

no post-processing in ffdshow, simply bypassing as 16-235 YV12, and mVR taking care of TV>PC conversion and dithering on top(no LUT enabled in there):

and now SmoothL() plus an 8bit YUY2 LUT in ffdshow, outputting 0-255 RGB32 to mVR, only taking care of dithering:

on my semi-10bit CRT, the latter looks much smoother.

nevcairiel · 24th August 2011, 15:51

Its an artificial pattern, how is that real world?

Anyhow, while the first shows very minor 1px bands (no severe macro bands), the second shows very strong dithering noise.
Personally, on this LCD screen, i prefer the first. But of course, that may just be a matter of taste, and/or different screen technologys.

Wile-E-Coyote · 24th August 2011, 15:52

Hello.

I just successfully calibrated my TV with dispcalGUI and wanted to create a 3Dlut file with it but when I open my ti3 file I get an error message saying the input string format is incorrect.

leeperry · 24th August 2011, 16:02

Quote:

Originally Posted by nevcairiel

Its an artificial pattern, how is that real world?

Anyhow, while the first shows very minor 1px bands (no severe macro bands), the second shows very strong dithering noise.
Personally, on this LCD screen, i prefer the first. But of course, that may just be a matter of taste, and/or different screen technologys.

Well, real world as in "going through the whole video pipeline".

I'll give you that I was using SuperSampling and Grainfactory4()....now w/ only SmoothL+YUY2 LUT:

any better on your LCD? Hopefully not a 6bit TN?

I could also disable dithering in mVR, but I like a grainy picture tbh:

nand chan · 24th August 2011, 16:09

Quote:

Originally Posted by leeperry

Well, real world as in "going through the whole video pipeline".

I'll give you that I was using SuperSampling and Grainfactory4()....now w/ only SmoothL+YUY2 LUT:

any better on your LCD? Hopefully not a 6bit TN?

I could also disable dithering in mVR, but I like a grainy picture tbh:

The third pattern looks identical to the first pattern on my monitor. The second pattern has awful dithering noise as mentioned.

10-bit H-IPS with 12 bit internal LUTs.

Quote:

Originally Posted by Wile-E-Coyote

Hello.

I just successfully calibrated my TV with dispcalGUI and wanted to create a 3Dlut file with it but when I open my ti3 file I get an error message saying the input string format is incorrect.

Upload the .ti3 you get errors with.

Wile-E-Coyote · 24th August 2011, 16:32

Here you go

nand chan · 24th August 2011, 16:33

Going to go ahead and release version 0.5 as-is because the software is still alpha either way and I don't want to delay the current release/feedback cycle too much.
Might still be some bugs, I haven't extensively tested every single tool after the code re-write. Everything should be a lot, lot cleaner though.

In version 0.6 I am going to be adding multi-threaded .3dlut processing to speed up execution time on multi-core processors.

As usual, see the first post for a changelog: http://www.mediafire.com/?wwbuu5vhau5x6jm

I'm still trying to think of a new name for the project.

Quote:

Originally Posted by Wile-E-Coyote

Here you go

Works fine for me in version 0.5, try it.

nevcairiel · 24th August 2011, 16:43

Quote:

Originally Posted by nand chan

The third pattern looks identical to the first pattern on my monitor. The second pattern has awful dithering noise as mentioned.

I agree, third and first look pretty much identical.
The 4th shows more banding then the others.

Also H-IPS 10bit with 12bit LUT (Dell U3011)

leeperry · 24th August 2011, 16:55

It's not dithering at all, it's grain added on purpose via the GrainFactory4() Avisynth script(I'm a movie grain whore

). But you can see that 10bit untouched mVR processing and SmoothL+8bit YUY2 LUT in ffdshow+mVR dithering look identical, which was my original point...so to me that still leaves the ffdshow LUT option very much open.

Wile-E-Coyote · 24th August 2011, 17:16

Quote:

Originally Posted by nand chan

Going to go ahead and release version 0.5 as-is because the software is still alpha either way and I don't want to delay the current release/feedback cycle too much.
Might still be some bugs, I haven't extensively tested every single tool after the code re-write. Everything should be a lot, lot cleaner though.

In version 0.6 I am going to be adding multi-threaded .3dlut processing to speed up execution time on multi-core processors.

As usual, see the first post for a changelog: http://www.mediafire.com/?wwbuu5vhau5x6jm

I'm still trying to think of a new name for the project.

Works fine for me in version 0.5, try it.

I did. Same error.

nand chan · 24th August 2011, 17:37

Quote:

Originally Posted by nevcairiel

I agree, third and first look pretty much identical.
The 4th shows more banding then the others.

Also H-IPS 10bit with 12bit LUT (Dell U3011)

Dell U2410 here

I've finished writing the multi-threading code now, applycal.exe completes /calculation/ in about 0.6 seconds. Saving/loading the .3dlut still takes about 1 second overall.

CPU usage spikes nice and high at 100%.

Quote:

Originally Posted by Wile-E-Coyote

I did. Same error.

I can't reproduce it. It works fine here, even using the GUI. Are you sure you're using the latest version?

Wile-E-Coyote · 24th August 2011, 17:42

Quote:

Originally Posted by nand chan

I can't reproduce it. It works fine here, even using the GUI. Are you sure you're using the latest version?

Yes, .NET framework is also up to date. I'll guess I have no choice but sit through another 2 hours of calibration to try again.

nand chan · 24th August 2011, 17:44

Quote:

Originally Posted by Wile-E-Coyote

Yes, .NET framework is also up to date. I'll guess I have no choice but sit through another 2 hours of calibration to try again.

Does running

Code:

genscript.exe -i profil.ti3 -o -

work?

It works here and it should output the following: http://pastebin.com/pmei93qL

You can go ahead and copy/paste that into a .txt file, run it using yCMS and it should give you a working .3dlut.
(You already have a copy of yCMS in %AppData%\TI3Parser\yCMS if you've been using this program)

Quote:

Originally Posted by leeperry

It's not dithering at all, it's grain added on purpose via the GrainFactory4() Avisynth script(I'm a movie grain whore

). But you can see that 10bit untouched mVR processing and SmoothL+8bit YUY2 LUT in ffdshow+mVR dithering look identical, which was my original point...so to me that still leaves the ffdshow LUT option very much open.

Fair enough. What is it exactly that you want to do? Apply calibration information from a .cal file to a YCbCr LUT?

If so, it should be a simple process of going through the YCbCr LUT, converting each entry to RGB, running the RGB through the Calibration LUT, then converting the result back into YCbCr.

I can have a program for this running by 0.6, but I can't test it myself. You'll have to check for bugs yourself.

Wile-E-Coyote · 24th August 2011, 17:49

Quote:

Originally Posted by nand chan

Does running

Code:

genscript.exe -i profil.ti3 -o -

work?

It works here and it should output the following: http://pastebin.com/pmei93qL

You can go ahead and copy/paste that into a .txt file, run it using yCMS and it should give you a working .3dlut.
(You already have a copy of yCMS in %AppData%\TI3Parser\yCMS if you've been using this program)

No running that command line gives me the same error message. I'll try your advice later. Thanks.

EDIT : yCMS gives me an error message as well

Quote:

Error! Invalid Command: '109.8925'.

nand chan · 24th August 2011, 18:15

Quote:

Originally Posted by Wile-E-Coyote

No running that command line gives me the same error message. I'll try your advice later. Thanks.

EDIT : yCMS gives me an error message as well

Very odd, I get:

Code:

The file '..\test.3dlut' was successfully created.
Running time: 4449ms; Computation time: 1765ms (39.7%)

When you're saving the file do you have some sort of word wrap enabled maybe? Also what operating system are you on? Maybe it's detecting the line breaks incorrectly, which would explain both issues.

@leeperry: Can you run inspect3dlut on a YCbCr .3dlut and upload the output somewhere? (Or simply tell me what it reports for the two color encoding fields)

Wile-E-Coyote · 24th August 2011, 19:36

Quote:

Originally Posted by nand chan

Very odd, I get:

Code:

The file '..\test.3dlut' was successfully created.
Running time: 4449ms; Computation time: 1765ms (39.7%)

When you're saving the file do you have some sort of word wrap enabled maybe? Also what operating system are you on? Maybe it's detecting the line breaks incorrectly, which would explain both issues.

Word wrap? Not that I know of. I pasted the text on notepad.

I'm running Win 7 Ultimate x64

nand chan · 24th August 2011, 19:39

Quote:

Originally Posted by Wile-E-Coyote

Word wrap? Not that I know of. I pasted the text on notepad.

I'm running Win 7 Ultimate x64

See “Format -> Word Wrap”.

Regardless, I don't know what else could be causing the problems.

Wile-E-Coyote · 24th August 2011, 19:43

Quote:

Originally Posted by nand chan

See “Format -> Word Wrap”.

Regardless, I don't know what else could be causing the problems.

Word wrap is deactivated, but I decided to open the file in Notepad++ and re-save it, and it worked, at least for yCMS.

I'm trying to use that method for the 3ti files but without success so far.

alph@ · 24th August 2011, 21:02

Quote:

Originally Posted by Wile-E-Coyote

Hello.

I just successfully calibrated my TV with dispcalGUI and wanted to create a 3Dlut file with it but when I open my ti3 file I get an error message saying the input string format is incorrect.

here too, same error message?i have the framwork 4,it's the last version of ti3 Parser
upload
http://www.mediafire.com/?rr2648w3ow9ftb4

24th August 2011, 15:51	#43 \| Link
nevcairiel Registered Developer Join Date: Mar 2010 Location: Hamburg/Germany Posts: 10,373	Its an artificial pattern, how is that real world? Anyhow, while the first shows very minor 1px bands (no severe macro bands), the second shows very strong dithering noise. Personally, on this LCD screen, i prefer the first. But of course, that may just be a matter of taste, and/or different screen technologys. __________________ LAV Filters - open source ffmpeg based media splitter and decoders Last edited by nevcairiel; 24th August 2011 at 15:54.

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

24th August 2011, 15:52	#44 \| Link
Wile-E-Coyote Registered User Join Date: Jun 2011 Posts: 61	Hello. I just successfully calibrated my TV with dispcalGUI and wanted to create a 3Dlut file with it but when I open my ti3 file I get an error message saying the input string format is incorrect.

24th August 2011, 16:32	#47 \| Link
Wile-E-Coyote Registered User Join Date: Jun 2011 Posts: 61	Here you go

24th August 2011, 16:55	#50 \| Link
leeperry Kid for Today Join Date: Aug 2004 Posts: 3,494	It's not dithering at all, it's grain added on purpose via the GrainFactory4() Avisynth script(I'm a movie grain whore ). But you can see that 10bit untouched mVR processing and SmoothL+8bit YUY2 LUT in ffdshow+mVR dithering look identical, which was my original point...so to me that still leaves the ffdshow LUT option very much open.