How to achieve calibration using the simulated Q60 graphic, example 3 with Canon EOS*DCS*3 Digital Camera.

Please read the introduction page first.

The Q60 target was photographed using the Canon EOD*DCS*3 digital camera under daylight simulating fluorescent tubes (Osram Dulux /12) and was acquired in the Linear Output Mode of the Kodak DCS Twain Driver.

In short the accurately simulated Q-60 CGI is compared against the acquired image of the target in difference mode using Photoshop 5.0.x. The Curves, ChannelMixer (and Levels) and possible Hue/Saturation adjustment layers are grouped to the layer that holds the acquired Q60 image and are then used to bring the composite difference image as close to black as possible. Since the comparison is done in difference mode black indicates no difference.  When the calibration is done the Curves, ChannelMixer (Levels) and Hue/Saturation adjustments are saved to the disk (into a suitable directory) to be used by an Action that calibrates the actual acquired images.

Layer arrangement for calibration. This camera is very accurate so only ChannelMixer and Levels adjustment are required for a good calibration. The first Levels for luminance match AdjustmentLayer is there for correcting possible overall luminance errors using the RGB composite layer only. It is not needed for Canon EOS*DCS*3.

So adter the calibration an Action is created using the saved adjustments with what the corrections can be easily applied over the subsequent acquired images in the workflow.

The calibration procedure

Curves adjustment

Curves adjustment layer is used general gray linearization only. 

• Evaluate the column 16 or and/or the horizontal gray scale at the bottom of the Q-60 and adjust the curves individually for gray neutrality. In other words add points to each of the the red, green and blue splines as necessary and adjust them so that the gray swatches are gray and appear very back in the difference view.

Canon is nicely linear so the curves is not used.

ChannelMixer Adjustment

• Use the Color Sampler Tool to add sampler point1 to the center of the D17 patch, sampler point2 to the center of the D18 patch and sampler point3 to the center of the D19 patch. 

 
• Have only the the Curves adjustment layer and the acquired image layer now in normal mode enabled and read and copy the sampler values from the Info palette to this Excel chart as "acquire values". 

 
• Disable all other layers than the Q60 CGI layer and the Levels for luminance match layer then read and copy the sampler values from the Info palette to the Excel chart as "Q60 values".

 
• Copy the calculated Channel Mixer values from the Excel chart into the Channel Mixer adjustment layer in Photoshop. 

 
Even if the ChannelMixer is based only on three RGB points it will help enormously with the calibration.

The ChannelMixer in Photoshop does not allow larger than 200% adjustment for any component, and it sometimes happens that this in not enough. In such cases the Excel chart calculates the colors in a reduced range and it says "Scale RGB white-Point to level xxx". If the xxx is other than 255 then you need to add such a Layers adjustment on top of the ChannelMixer layer. This scaling is not a problem since the actual work will be always done in the higher bit-depth.

Hue/Saturation Adjustment

Usually there will still be a little work to do with the Hue/Saturation that is added on top of the ChannelMixer and the possibly needed Levels adjustment layers.  Try not to make very narrow adjustments using the spectrum sliders since this tends to create a lot of artifacts.  Go through each of the colors and adjust/iterate the values so that the difference mode view goes as black as possible. This is not needed/ done in this example.

Writing the Action

Please remember to include the mode change into 15-bit as the first step of the Action that you write, this is important for quality. The steps and their order in the Action should be:
 
1. convert mode to 15-bit mode.
2. apply the Curves adjustment.
3. apply the Channel Mixer adjustment.
4. apply the possibly needed Levels adjustment.
5. apply the Hue/Saturation adjustment.

To calibrate the images just apply the Action right after the acquire. After that your task will then be to manually fine-tune or adjust the black-point and white-point of the image using the Levels dialog plus any subsequent image enhancement.

Canon EOS*DCS*3 Example

All the work was done in AIM RGB, the illustration below are scaled down in size, compensated by gamma 1/2.5 for viewing with an uncalibrated PC system.

	This is the Linear Acquire from Canon EOS*DCS*3 using the Kodak DCS Twain driver, in the Linear Output Mode. In the Linear Ooutput Mode the Kodak DCS Twain driver does not convert the color-space so it appears to be rather flat or undersaturated, this is in fact only good (the camera has a very large gamut).
	After the Channel Mixer with proper values from the Excel chart. If the Channel Mixer in Photoshop would allow 400% values then the next Levels adjustment step would would happen here, but it does not.
	Final Calibrated image after Channel Mixer + Levels dialog. There may be a little to do with the an additional Hue/Saturation adjustment but it is not done in this example.
	This is how the difference image looks like after the calibration is complete (ChannelMixer&Levels). It is not completely black, but if you use the on-top comparison at the end of this page you will see that it does not matter much, in fact this is a very good calibration. The remaining differences can be rather easily minimized by applying a Hue/Saturation adjustment as the final correction.	(The simulated graphic is the bottom most layer and the white borders (255,255,255) was removed from it so the face of the lady shows normally even in the difference mode) .
	And This is how the Q-60 R1:1997:04 simulated target looks like.

For easier comparison: