Create a synthetic test chart from several individual image files
Composite Chart can be used for two essential purposes.
- You may want to analyze colors in a microscopic image that has a field too small to capture an entire test chart (for example, a replica of the X-Rite Colorchecker, which is available in 35mm film size (24x36mm).
- You may want to combine several images to obtain tonal response and noise, following recommendations in certain ISO standards.
Composite Chart start window
First, select the chart type.
Options for chart types are
- Construct a 24-patch Colorchecker*.
- Construct an IT8.7 (rows A-L, columns 1-19: 228 patches). The first two were requested by a customer with a microscopic system where the field of view was too small to capture the entire chart. We don’t generally recommend the IT8.7 because has too many patches to be practical (it’s easy to make sequence errors).
- Construct a 1xN linear step chart*. This combines several images for certain ISO standard calculations.
- Construct a 4×12 grid chart. This was a special customer request. We can easily add additional charts on request.
*These are the most popular charts: they are recommend unless you have a specific need for a different chart.
The second dropdown menu has two selections:
- Small chart: small squares with averages of crops (not for noise measurements) This is a good choice when colors and tones are to be analyzed (but not noise). patches are always square.
- Large chart: Use full crops (good for noise measurements, except for IT8.7) Patches maintain the aspect ratio of the crop (below, after images are read).
Second, press to choose the files to combine. The order will be based on a sort of the file name: we recommend that the names contain a numeric sequence. Use the usual techniques (shift-click, control-click; wildcard characters if needed) for selecting multiple files.
Third, click .
You’ll be asked to crop the chart with the normal sequence of windows. If a crop selection has been saved, you’ll be asked if you want to repeat it. If not, you’ll be asked to enter a new crop. Here is a typical fine crop window.
The crop size is not critical, but if noise statistics are required the crop should be at least 100×100 pixels; between 200×200 and 400×400 is optimum. Very large crops (over 800×800) are unnecessary and may slow down the analysis of the results. Remember the final image will be larger than any of the crops, so for example, a 1×12 linear chart with a 400×400 pixel crop will have a total size of 400×4800 pixels.
Once the crop has been selected, the chart will be displayed.
If the chart image looks correct, click and follow instructions. ,
What to do with the composite image
Most composite image files can be analyzed by Color/Tone Interactive or Color/tone Auto (these were called Multicharts and Multitest in earlier versions of Imatest). Legacy modules Colorcheck can also be used for the 24-patch Colorchecker and Stepchart can be used for grayscale charts, but we don’t recommend these for new work. The new modules have superior noise and dynamic range analysis and can be used for calculating Color Correction Matrices (CCMs).
We recommend running Color/Tone Interactive before Auto — to veryfy that settings are correct. Once the settings look good in Interactive, it’s OK to run auto, which can analyze batches of images.
Before running Color/Tone Interactive, the chart must be selected in the Color/Tone Interactive window.
Selections for Colorchecker-24 and linear grayscale charts are shown.
For linear grayscale charts (stepcharts), you will need to select the number of patches and the density reference (which can be steps of 0.1, 0.15, or 0.3 density units, or (more commonly) a density reference file.
Further instructions are in Using Color/Tone Interactive and Using Color/Tone Auto.
One more detail: some tiny charts are now available that were not available when Composite Chart was originally released. They may allow direct image capture without the need for Composite Chart. The most interesting among these is the ColorGauge, which comes in several sizes.