ACR Calibrator Problems

[Detonn]

Hi

I'm trying to calibrate my Canon 40D with the Kodak q60 target and I
keep getting 100s or -100 which I've read is an error.

I photographed the target with a studio strobe.

I opened it with ACR and change it to 8 bits along with the smallest
file size and profoto. I had to convert it to an adobe dng file to
open it with ACR.

Then I used the pen tool and went from the top left corner to the
bottom left then over to the bottom right then to the top right.

Any ideas why?

[Steve Sprengel]

The 24-patch GMB/XRite Color Checker (and mini variation) is the only calibration target that the Fors and Tindemans scripts work with.

 

A Q60 has more colors in different positions so unless the target you are using has a small embedded CC-colored array like the CC SG and you use the path-tool to mark just that little area, it will not work at all.

 

I believe Rags has a variation of the script that will work with the Q60-type targets but you have to specify the three colors to use and I have never used it so don't know how easy or hard it is.

 

For reference a Color Checker looks like this:

http://www.babelcolor.com/main_level/ColorChecker.htm

 

And costs around $70 for the full-sized one, and a little less for the mini:

 

http://www.bhphotovideo.com/c/product/465286-REG/Xrite_MSCCC_Original_ColorChecker_Card.html

 

http://www.bhphotovideo.com/c/product/465287-REG/Xrite_M50111_Mini_ColorChecker_Card.html

[Detonn]

Thanks for the information. I'll buy the other color checker. Will
this script work on a dng file. I can't open the .cr2 file with CS2.

Gary

On Jun 4, 11:38 pm, "Steve Sprengel" <st...@sprengels.com> wrote:
> The 24-patch GMB/XRite Color Checker (and mini variation) is the only calibration target that the Fors and Tindemans scripts work with.
>
> A Q60 has more colors in different positions so unless the target you are using has a small embedded CC-colored array like the CC SG and you use the path-tool to mark just that little area, it will not work at all.
>
> I believe Rags has a variation of the script that will work with the Q60-type targets but you have to specify the three colors to use and I have never used it so don't know how easy or hard it is.
>
> For reference a Color Checker looks like this:http://www.babelcolor.com/main_level/ColorChecker.htm
>
> And costs around $70 for the full-sized one, and a little less for the mini:
>

> http://www.bhphotovideo.com/c/product/465286-REG/Xrite_MSCCC_Original...
>
> http://www.bhphotovideo.com/c/product/465287-REG/Xrite_M50111_Mini_Co...

>
>
>
> ----- Original Message -----
> From: Detonn
> To: AcrCalibrator
> Sent: Wednesday, June 04, 2008 8:30 PM
> Subject: [acrcal] ACR Calibrator Problems
>
> Hi
>
> I'm trying to calibrate my Canon 40D with the Kodak q60 target and I
> keep getting 100s or -100 which I've read is an error.
>
> I photographed the target with a studio strobe.
>
> I opened it with ACR and change it to 8 bits along with the smallest
> file size and profoto.  I had to convert it to an adobe dng file to
> open it with ACR.
>
> Then I used the pen tool and went from the top left corner to the
> bottom left then over to the bottom right then to the top right.
>

> Any ideas why?- Hide quoted text -
>
> - Show quoted text -

[Steve Sprengel]

The script works the same with DNGs as RAWs if the version of CSx's ACR version also supports the DNG's underlying RAW file format...in other words if you can open the RAW in Photoshop you can use a DNG or RAW interchangeably.

 

You are wanting to use the DNG from a new DNG-converter with an older ACR. 

 

From some experimentation when the 40D came out, I can tell you the script-computed calibration for a DNG with RAW format that is unsupported by ACR is different than the script-computed calibration for a DNG where the RAW format is also supported by the version of ACR. 

 

I would recommend keeping up-to-date with Photoshop versions instead of doing the new-DNG/old-ACR hack, but it may work ok, I don't know because I don't do things that way.

 

[Detonn]

Again Thanks for the info. I would have never know that CS2 could
not open my raw files if it wouldn't have been for attempting this
calibration because I always use Lightroom to process my raw files.
But it does open the dng file with ACR 3.7 so hopefully the script
will work correctly this time.

Gary

> > - Show quoted text -- Hide quoted text -

[Steve Sprengel]

Every new camera model has potentially a different light-response curve as well as potentially embedding new fields in the RAW file format; so ACR has to specifically be updated for each new camera model.  Adobe stops updating ACR when a new version of Photoshop comes out with its new flavor of ACR and your camera is just newer than the last version of ACR that came with CS2.

--

Presumably you are wanting to use your script-computed calibration with Lightroom?

 

When you create your new camera's DNG in the current free DNG converter that does understand the RAWs from your new camera and then open that DNG in CS2's ACR that does not understand your camera's RAWs, you will see that the profile is labelled "Embedded".

 

What I'm not sure is if that Embedded profile in the DNG IS the camera profile or whether it has been manipulated to make it more standard or generic.  If it IS your camera's profile then running the calibration-script in CS2 should produce the correct slider values.  However if it is something that isn't the same as your camera's profile, then running the script would not produce the values needed by Lightroom which is using a RAW engine that does know about your camera.

 

You can probably test to see if Lightroom and CS2's ACR both produced the same results from your DNG by opening the DNG in each, saving the result as a PSD or TIF (so there's no loss due to compression like there would be with a JPG) and then loading each of those images into Photoshop layers and set the blend-mode to difference--which should look almost completely black if they are similar, flatten the image so there is only one layer, then open Levels and drag the high-light slider down close to zero (the black end) and see if any detail starts to show up.  If not then your CS2 and LR DNG conversions are identical and you're probably ok in using CS2 to compute your calibration from the DNG.

 

In ACR and LR you need to turn off the sharpening and noise-suppression otherwise there will be differences just caused by those operations.

[Detonn]

Well I just recieved the new color checker and it didn't work with the
dng file because I got strange results and one of the numbers was
100. Next I'll test it out with my Canon 20D because that should open
in CS2. If that works I'll have to buy the next version of Photoshop.

Thanks
Gary

[steve sprengel]

If you want, you can send me the DNG and I'll test it with CS2...which I haven't deleted even though I have CS3.

[kermit]

Hi I'm new to the forum. Instead of clogging things up with another
thread I thought I'd add to this one as it is similar to the op - I
hope people do not mind.

I'm attempting to calibrate my D80 using the script. I have followed
the instructions on this site and watched the video on Martin Evenings
'CS3 for Photographers' DVD. Nevertheless, I have run the script 4
times this afternoon - each time on the same shot of an Xrite
ColourChecker. So far the results have varied wildly.

As for the background to the shot - shot in RAW under tungsten light
and converted to DNG in ACR.

What am I doing wrong?

[Steve Sprengel]

There is a video, somewhere, that tells the user to select the corners in an X shape, but it should be a U shape:  top-left, bottom-left, bottom-right, top-right.  Perhaps you're doing that, or maybe you are selecting the very edges of the squares instead of the centers.

 

HOWEVER, if you have CS3, with the release of ACR 4.5 and LR2.0 last week, you can use the new Adobe Labs DNG Profile Editor to create a custom Camera Profile that you select in place of the ACR 4.4 that you see in ACR and LR, now.

 

Using Adobe's DNG Profile Editor to create profiles for my camera, I see the same reduced color-error as when using Tom's script and it only takes a few seconds to run, as opposed to 10s of minutes for Tom's script.

 

http://labs.adobe.com/wiki/index.php/DNG_Profiles

 

Don't worry that it's called a DNG-something.  That's merely because the profiles you create are documented in the DNG 1.2 specification.  The profiles work the same as the Adobe-supplied profiles, for both RAWs and DNGs.

 

Besides the Profile Editor, Adobe Labs also lets you download a whole set of profiles for each camera.  Not sure what your Nikon has, but my Canon has profiles named:  Camera Faithful, Camera Neutral, Camera Landscape, Camera Portrait, Camera Standard.

 

There is a FAQ and a Tutorial about how to create and use the profiles.

 

 

From: kermit

Sent: Monday, August 04, 2008 1:14 AM

To: AcrCalibrator

Subject: [acrcal] Re: ACR Calibrator Problems

[kermit]

Many thanks for the link - have got it all working.

One observation though... The lecturer gave out a list of numbers -
and the numbers generated are not even close to those on the sheet.
Is this a problem witht he camera or what?

Kermit

[Steve Sprengel]

If you are talking about the recently obsolete Fors' script:  The numbers are the calibration and would be different for each model of camera, and can be entirely different for different camera models, quite a bit different for each white-balance situation, a little different from different cameras of the same model, and even slightly different from one run of the script to the next due to pen-placement differences.

 

You should be using the DNG Profile Editor to create a camera profile, with one color table for 2850K (Tungsten) and one color table for 6500K (Clouds/Shade) and use that for everything and leave the calibration-sliders at 0.