Kodak Q-60 and Agfa IT8 Color Input Targets These very useful Color Input Targets are  available from Kodak at about 30 USD, they are very useful for calibration. The data-files for the targets can be downloaded from Kodak FTP, there are data-files for each batch they have manufactured.  The only problem with this target is that it does not cover the highlights very well, but even so it is without doubt the best affordable target available. Accurate CGI simulations for these targets are provided below.

The IT8/7.2 specification defines manufacturing target values (Kodak calls them as aim values) for the 22 gray patches in the horizontal grayscale at the bottom of the Q-60 chart. These aim values are specified in Lab-space lightness (L*) so note: The step numbers are not densities! Just step numbers. Other patches are somewhat specific to the manufacturer or manufacturing process. The table on the right lists the reflectance and density values for the aim value of the 22 L* steps taken from the Kodak specification (in Acrobat PDF format). Note: The table is capture from Excel sheet and in Finland we use the comma "," in place of the dot "." so e.g. 70,01 means 70.01 in many other countries. In addition the table shows the average of Dmin patch (the leftmost patch in the horizontal gray swatch) and the average of the Dmax patch (the rightmost patch in the horizontal gray swatch). The averages were taken over 16 Kodak Q-60 reflective charts and min and max values shows the range of those averages. So the output range of a Q-60 chart is then 78.89/0.30 = 263:1. This is somewhat a larger range than the usual misinformation about the poor range of photographic paper (often said to be only about 40:1). Well, the target is manufactured on Kodak Professional paper. The main drawback of the Kodak Q-60R is that the Dmin patch is not very light, it only reflects about 79% so the chart does not cover the full range of scanners that are usually adjusted for 100% reflectance. In addition in some of the target batches the Dmin patch is not paper white.

Typical histogram of the Q-60 scan is shown below with the red arrows indicating the highlight range not covered by the chart. The photo on the left shows the Kodak Q-60 and the Q-60 simulation CGI that is on Agfa Professional photographic paper. As can be seen the media white (depending on the photographic material) can be considerably lighter than the Dim of the Q-60 chart so: Photographs can have image detail in the highlights that the Q-60 chart does not cover. Therefore if the calibration is constructed so that only the range on Q-60 is taken into account, highlight image detail will be lost.

Unfortunately many ICC profilers such as the MonacoEZcolor and IphotoMinus do just this, they confine the output range to that of the Q-60. One easy way to examine if the scanner profile cuts the highlight is simply to convert a chart of 256 grays using that profile and to examine the highlight. The CGI simulation assisted manual calibration explained below ensures that the available range scanner is not cut, and results equally accurate calibration.

Computer Generated Simulations ( CGIs) for Kodak Q-60 and Agfa IT8

It is rather odd that Kodak nor Agfa do not provide simulated graphics for the Q60 targets. Such an accurate computer generated graphic provides means to a very accurate input device calibration since the acquired image can be compared against the graphic very intuitively e.g. in Photoshop as outlined below. It can be used also for output device calibration as soon as one input device is calibrated. I had the need for such an accurate graphic so I wrote an Excel VBfA model that converts the XYZ values from the Q-60 data-sheet into a freely chosen RGB working space, by CIE tri-stimulus absolute colorimetry, and writes a Photoshop image out. Since I had the macro I then made it read the Q-60 data-files automatically so there are now an accurate CGI graphic for each q60 batch that I found the data-file for.

Example of the CGI simulation of Kodak Q-60R target

Example of the CGI simulation of Agfa IT8 target

The simulated Q-60 graphics are:

• Higher bit-depth Photoshop RGB images in a linear wide-gamut color-space (700/525/450nm -primaries, 5000K white-point and gamma 1.0), download the profile if needed, unzip and save it into: C:\WINDOWS\SYSTEM\COLOR\  -directory. They have the the profile embedded.
• Photoshop PSD files in a zip compressed archive.
• In miniature size of 26 x 17 pixels. In order to use them properly you need to resize the image size using Nearest Neighbor interpolation method, e.g. in Photoshop choose the menu command Image/ImageSize and select the following sizes depending on the scan resolution:
• actual Q60R1 chart scanned at 150DPI, change CGI width to 1000 pixels.
• actual Q60R1 chart scanned at 300DPI, change CGI width to 2000 pixels.
• actual Q60R1 chart scanned at 600DPI, change CGI width to 4000 pixels.

  Let the height scale by constrain proportion.

The file-space of the CGIs is: Primaries=700/525/450nm, white-point=D5000 and gamma=1.0),
download the profile if needed, (The other name for the "700/525/450nm" primaries is "Wide-Gamut" -primaries).

Note that I will not automatically update the CGIs, please e-mail me in case yours is not yet listed.

Download CGI simulation:

Download Kodak Q-60 CGI simulation (opens in a new browser window)

Download Agfa IT8 CGI simulation (opens in a new browser window)

Please note that you can use the XLProfiler to write CGI simulation for any Kodak or Agfa target directly to user selectable working-space, and it does a pretty good job in calibrating too!

Accuracy of the Simulated Graphics

The graphs are accurate per the CIE_XYZ and the L*a*b* values from the Q60 data-files down to the 15th bit (that is all that Photoshop has). When you compare the L*a*b* values as reported by Photoshop from the CGI against the values in the Q60 data-file you will notice some errors, these errors are due to the inaccuracy of the L*a*b* space of Photoshop. The simulated graphs were calculated in double-precision floating-point in Excel and converted according to CIE, by the book. The Lab mode in Photoshop is only 8-bit and it calculates the conversions by some optimized look-up-tables. I have evaluated the color mode conversion errors of Photoshop 4.0.x and there are similar but slightly smaller errors in v 5.x.

Color Converting the Simulated Q-60 Graphic for use

The simulations are in a wide-gamut color-space because the Q-60 target has a little larger gamut than what is the Trinitron gamut; taking the Q-60 simulation into Trinitron will clip some very saturated hues in cyan and yellow. This normally does not have any perceivable effect when systems are correctly calibrated but printing the Q-60 simulation that has been converted to Trinitron using a profile driven system that is poorly calibrated can make the clipping visible. The wide-gamut space holds all the Q-60 data, but because you probably are not using that gamut as the working space you must convert the Q-60 CGI simulation. Fortunately this is easy, simply allow Photoshop to convert the colors into you working space when opening the CGI, the simulations do have the embedded profile. You must use Relative Colorimetry and have the Black-Point compensation Off in the Profile to Profile conversion -dialog.

After the conversion you will need to convert the 15-bit graphic to 8-bit mode because Photoshop does not allow to use layers in 15-bit mode.

There will be errors to the dark-end if you let Photoshop convert the linear CGI simulation into a gamma space. The errors are due to the "feature" called slope-limiting in the color-management engine of Photoshop. But creating an accurate color calibration in a non-linear space is not possible so there is no reason to do this.

Q60 Computer Generated Graphic Assisted Manual Calibration

There are many ways to calibrate, the basic rule is "use the appropriate tools that your image editing software provides in order to bring the acquire device into calibration". It is very important to use (as much as possible) tools that apply the effect continuously without lightness or hue based discrimination or restrictions.

Basically what is needed to bring a tri-chromatic acquire device into calibration are are

1. Curves adjustment for linearization and color temperature correction.
2. Channel Mixer for the color-space conversion.

But the internal software of the acquire device can be messing around with the data so quite often a hue-selective Hue/Saturation step is needed as the last step.

Continue to the Calibration Example (Umax PowerLook III scanner calibrated to WideGamut).

Please be sure to look at the XLProfiler page, it is a great improvement over the manual CGI assited calibration, Microsoft Excel does it all for you, very accurately too.

Old examples (will be removed soon):

1. Original version using Curves, Hue/Saturation and Levels. This can be very tedious.
2. Greatly improved version, Umax Powerlook III calibration using Curves, ChannelMixer and Hue/Saturation Microsoft Excel sheet is provided for calculating the ChannelMixer values (simple but very effective).
   
3. Latest one with Canon EOS*DCS*3 digital camera with Kodak DCS Twain in linear output mode