Back to www.coloraid.de

ICC Profiler and color scanner test data

Introduction

On this site I am providing data in order to test and help developing scanner color profiling software. I have prepared a number of measured test sets with 5 slides and 5 reflective prints. Users were asked helping by scanning the test images. The user made scans are collected and made available here on this site for free to anybody interested. The data can be used for a number of purposes. Starting with the better development of profilers, comparing the quality of scanner profiling software and more.

Test data

Each of the test set consists of five test images. A small manual on how to scan the test slides and provide the necessary info can be found here. One image is a standard IT8 target with 288 color patches that have been individually measured. This target can be used to generate profiles with standard IT8 compatible software:

Another four test images provide 1152 color test patches. 1000 patches basicly consist of 10x10x10 RGB grid in the used recorders device space. So they are not in sRGB nor did I had the time to adjust the used machines for a regular grid in CIE Lab space. But the number and spacing of patches should be more than enough to give a realistic view of the tested profiler and scanner. Another 152 test patches on the fourth slide mainly provide tricky dark, bright or highly saturated patches. These are the patches where we most likely find the biggest color faults of the scanner and profiler:

Image 1	Image 2
Image 3	Image 4

The slides were produced on Velvia 100F films. I simply choose that film because I was just busy producing targets for www.coloraid.de and had some space left on some Velvia 100F 8x10" films. I think it is also a good choice because Velvia film provide a rather large density range. Scanners and profilers able to handle these slides well will very unlikely run into color problems with other slide films.

Profiling

Hal Engel wrote a small manual on how to use the >1000 patches of all test slides for creating a profile using lprof >V1.11. You can download the manual here.

Problems and bad ideas

There are a number of issues you should be aware when using the data presented on this website. Here a list of them:

• This site is not made for reviewing/comparing scanner hardware. The data might indicate the color quality of a scanner, but you really should check other scanner hardware review articles.
• Scans are made by users. While I try to sort out faulty scans made by users, most user scans are far from optimal. Please report any problem found with the data and I will add a comment to the data for others.
• The colors of the test sets are not 100% identical (actualy, they are a lot worse compared to my normal IT8 productions). As a result, comparing scans/profiles made with different test sets should be done with care. In practice the differences should be too small to cause major differences in profile quality generated from the data.
• If your profile shows serious color faults for certain highly saturated colors, do not blame the profiling software without looking at the originals scan data. Check if the scanner has enough color resolution before blaming the profiling software. It might be scanner hard- or software that might be the cause of the problems.
• The test images have a very high density range. Clipping or other hardware limits often occur. Check the histogram of an image for clipping. Note that the DMin/DMax patches of the test scans usualy do not necessarily have the highest/lowest R,G,B value (for instance, check patch G8 in image 1). So do not only use a pipette on the dmin/dmax fields. Have a look at the histrogram in order to tell if there is clipping.
• The precission of the test sets is limited. Measurement faults, aging and more do affect the precission of the data. So please, do not try to report color faults with a precission that simply isn't there. Ask if you have certain questions regarding the precission of the test sets...

Download

TIFF source RGB image files

These are the (scaled) RGB image files used to produce the test slides and reflective prints. The RGB values of the images are in the used recorder device space. They are not in a widely known RGB space like sRGB.

Slide1.tif, Slide2.tif, Slide3.tif, Slide4.tif, Slide5.tif,
Print1.tif, Print2.tif, Print3.tif, Print4.tif, Print5.tif,

Reference files

These are the measured color reference files of each slide or print scanned. Note that there are several test sets number 1-15. Usualy only one set is shipped to the user for scanning. Each set contains 5 test images. Each slide/print has a charge wich consists of setno_slideno. The IT8 chart included in the test set has number 5. So the file testscan2_5 refers to the slide 5 in set 2.

Slide set 1: IT8 files, CGATS.5 spectral data files.
Slide set 2: IT8 files, CGATS.5 spectral data files.
Slide set 3: IT8 files, CGATS.5 spectral data files.
Slide set 4: IT8 files, CGATS.5 spectral data files.
Slide set 5: IT8 files, CGATS.5 spectral data files.
Slide set 9: IT8 files, CGATS.5 spectral data files.
Reflective set 11: IT8 files, CGATS.5 spectral data files.
Reflective set 12: IT8 files, CGATS.5 spectral data files.
Reflective set 13: IT8 files, CGATS.5 spectral data files.
Reflective set 14: IT8 files, CGATS.5 spectral data files.
Reflective set 15: IT8 files, CGATS.5 spectral data files.

Scans

Sorry, as some of the files are far too big for offering on the internet, I do publish the files in a scaled version. Those who need the full highres data should contact me and we work out a method how to get the data to you.

More scans are comming up....