New project: camera color profile calculation - page 24 - RAW Shooting and Processing - CHDK Forum

New project: camera color profile calculation

  • 480 Replies
  • 349008 Views
Re: New project: camera color profile calculation
« Reply #230 on: 25 / June / 2008, 03:15:28 »
Advertisements
Sess, A650 is supported. DNG4PS-2 "knows" about dimensions of A650's RAW and A650's color profile

OK, that's why I saw no differences between DNG4PS-2 and DNG4PS-2matrix output.

Do you know why the a650 isn't listed in a 12Mpx combo like other cameras ?

aymeric

*

Offline ArtDen

  • ***
  • 175
    • dng4ps2
Re: New project: camera color profile calculation
« Reply #231 on: 25 / June / 2008, 03:39:46 »
aymeric106, A650's RAW file has unigue file size so it is unnecessarily to include it into options

*

Offline Sess

  • *
  • 28
Re: New project: camera color profile calculation
« Reply #232 on: 25 / June / 2008, 15:37:05 »
Sess, A650 is supported. DNG4PS-2 "knows" about dimensions of A650's RAW and A650's color profile

Ah okay. I took a series of test photos with my A650. I took RAW+JPG pictures outside in daylight, and then inside at night with tungsten lighting. I noticed that the white balance chosen by ACR after conversion with DNG4PS-2 was skewed for some images.

Almost all of the outside daylight photos looked great. There were just a few pictures that had a pool in the image where the white balance was poorly chosen, however when I manually set those pictures to the "Daylight" setting in ACR, then they looked perfect.

For the tungsten lighting, however, the colors were all highly skewed. ACR would choose a white balance and the resulting images were very red. Not only that, but the white balance chosen by ACR was very close to the manual "Tungsten" setting, which when set did not look any better.

So to reiterate, in both daylight and tungsten lighting, ACR will choose a white balance that is nearly the same as the daylight or tungsten manual setting (the actual lighting conditions). However only the colors of the daylight photos look right; the tungsten photos look very red.

What do you think could be causing this problem?

*

Offline tb

  • *
  • 1
Re: New project: camera color profile calculation
« Reply #233 on: 28 / June / 2008, 10:59:33 »
a630 test pair:
zSHARE - a630-rawjpg.zip

a630-simple-pictures:
zSHARE - a630-simple-pairs.zip

*

Offline ArtDen

  • ***
  • 175
    • dng4ps2
Re: New project: camera color profile calculation
« Reply #234 on: 29 / June / 2008, 11:25:04 »
tb, thanks! I will look on your files later

*

Offline yujin

  • *
  • 15
Re: New project: camera color profile calculation
« Reply #235 on: 01 / July / 2008, 09:45:08 »
yujin, you can test:
zSHARE - dng4ps2-0.2.2matrix.exe

Hi ArtDen,

I found the previous palette & simple pictures are mistakenly set to "-1" saturation during shooting. Will this affect your calculation?

If yes, I have uploaded the corrected palette & simple pictures (with no saturation adjustment) for you. Sorry for the inconvenience caused :-

S2 palette : zSHARE - s2-palette-new.zip
S2 simple pictures : zSHARE - s2-simple-pictures-new.zip

If no,  I found that
1. converted DNG are brighter than JPG.
2. DNG Photos taken at noon are little less reddish than JPG.
3. DNG photos taken in late afternoon are little more reddish than JPG.
« Last Edit: 01 / July / 2008, 09:59:20 by yujin »

*

Offline cyril42e

  • ***
  • 111
  • SD1000/Ixus70 1.02a
    • CR-TEKnologies
Re: New project: camera color profile calculation
« Reply #236 on: 03 / July / 2008, 13:37:26 »
Hi,

I'm having some trouble with color profile for my SD1000. quietschi has provided test images and the matrix is included in DNG4PS2. So I converted my CRW file, opened it in rawtherapee with camera white balance, but noticed that the resulting jpeg was a lot more greenish than the original jpg.

So my first question: what's the relationship between color profile and white balance? Is it possible to correct a wrong color profile only with white balance?

Anyway I tried to calibrate my camera, had to fight a lot, but a the end managed to get a matrix with low enough error (had to remove the last black zone, set camera's white balance custom on the white square, only select a small area inside each square, and at the end all zones are detected with error < 10%, but 19 "white areas" detected...).

The result looked ok in dng4ps2 result preview, but when I apply the matrix on the DNG file with dngsetmatrix and then open it in rawtherapee, the result is different. I reduced brightness so that the colors are closer, but it is still different (see pictures). Do you know what could be the problem?

Thanks

*

Offline ArtDen

  • ***
  • 175
    • dng4ps2
Re: New project: camera color profile calculation
« Reply #237 on: 05 / July / 2008, 13:56:55 »
cyril42e, colors are different as it is not possible to restore whole camera profile from jpeg and raw files. Dng4ps2 calculates only color matrix without calculating linearization curves.

*

Offline ArtDen

  • ***
  • 175
    • dng4ps2
Re: New project: camera color profile calculation
« Reply #238 on: 05 / July / 2008, 15:06:27 »
zSHARE - dng4ps2-0.2.2matrix.exe - dng4ps2 with new test S2 matrix and IXUS950 matrix from dcraw.

tb
I couldn't get good results for your images

*

Offline yujin

  • *
  • 15
Re: New project: camera color profile calculation
« Reply #239 on: 06 / July / 2008, 01:53:23 »
zSHARE - dng4ps2-0.2.2matrix.exe - dng4ps2 with new test S2 matrix and IXUS950 matrix from dcraw.

tb
I couldn't get good results for your images

Hi ArtDen,

I tried the new test S2 matrix. The colors of the converted DNG  are almost the same as the JPG now, the JPG is a little bit brighter than DNG.

To me the new S2 matrix is excellent. I have started using it to convert all S2 RAW files. Thanks!

 

Related Topics


SimplePortal © 2008-2014, SimplePortal