Post Script images are pictures created from sampled image data. A file of sampled image data, called a bitmap, is simply a pattern of bits. Bitmaps contain the image matrix, the image size, and a procedure that delivers image data to the image operator. Here we will focus on the image matrix.
The image matrix is made up of a series of square pixels that are mapped to some user coordinate system. In this project, our pixels are mapped to printer paper. Printers represent pixels by a series of ink dots. Thus, we have round ink dots representing square pixels. This configuration results in a higher ink coverage than originally indicated by the image matrix. The purpose of printer calibration is to reconcile the difference between printer ink dispursement and the stored pixel data. This is accomplished by creating a tone reproduction curve.
The tone reproduction curve relates color reflectivity and ink density. The curve allows us to correctly map desired reflectivity to the amount of ink applied to the paper.
Printer Calibration is a two step process. The first step involves changing an images XYZ values to respective CMY values. The second step involves changing the CMY values to new CMY values based on the halftoning scheme that will be applied to the image before printing.
In this experiment, the conversion from XYZ to CMY was based upon a matlab code created by Rick Koehler. This code does not directly transform XYZ to CMY because CMY colors are "subtractive colors". The transform matrix created by this program is the transformation from ink levels to the XYZ values of the amount of light "absorbed" by the inks. The image matrix is calculated by subtracting a 3x3 matrix whose colums contain the XYZ values for the white point of the printer from a matrix whose columns contain XYZ valus of Cyan, Magenta, and Yellow.
T = [whiteXYZ whiteXYZ whiteXYZ] - [CyanXYZ MagentaXYZ YellowXYZ]
This gives us the CMY values for the monitor.Multiplying the monitors CMY values by the inverse T matrix will give us the printer's CMY.
Now that the printer's CMY values have been calculated, we further adjust them with the aide of tone reproduction curves. These curves relate the amount of ink applied to the paper to the halftoning scheme used. Multiplying the CMY values by the functions obtained from the tone reproduction curve, prime the printer values for use by a specific halftoning scheme.
This image is the result of directly applying the printer calibration methods mentioned above. The white in this picture looks more like lime-green, or a very pale cyan. To insure that white looks white, we chose to scale the values of X , Y, and Z so that the white point of the monitor matches the white point of the printer. I.E. , the new_X = original_X * (WhitepointX_printer) / (WhitepointX_monitor). Scaling the X, Y, and Z values produce this image. As you can see, this image is superior to the unscaled image.To view the inputted image file, click here.
The calibrated file shows a fairly reasonable match between the following colors: black, magenta, cyan, white, and blue. The red is passable because generally speaking, creating red is extremely difficult in the printer industry. In this case, we're putting down more cyan than is needed.
View the following locations to see a comparison of the famous photo of Lena. Click here for the original Image To see the printer calibrated image click here.
The photo shown here shows the original image of a Monarch butterfly. The photo shown here shows the printer calibrated image of the monarch butterfly.
For a printer calibrated image of peppers click here. To see the original image, click here.
