ColorChecker
|
|
|
|
Looking for information on the ColorChecker? Here you will find:
March 2007: Added visual comparisons between measured data and the average chart. June 2006: New averaged data (Excel file) and updated images (now also in 16-bit format). The "RGB coordinates of the ColorChecker" PDF was updated and augmented; it now presents a detailed comparative analysis of the latest (2005) reference data, provided by GretagMacbeth, against the results derived from our spectral average. Click on any line of the following list to go directly to the section of interest:
Attention ! Warning ! Important ! Read-me !This page is not really related to BabelColor. Well, it is in the sense that the ColorChecker includes six neutral patches, from very dark to almost white, and that BabelColor also sells a white target (in fact, a whiter target!). It is not related in the sense that the ColorChecker is the main subject of the page, that many products from other companies are mentioned, and because of the less informal tone used in the presentation. In fact, this page, on-line since 2004, simply exists because of my longtime interest in the ColorChecker. But before we continue, just a small formality: ColorChecker is a Registered Trademark of GretagMacbeth (purchased by X-Rite), and GretagMacbeth is a Trademark. OK, back to the ColorChecker! Comments welcomed! Danny Pascale
A brief historyThe ColorChecker Color Rendition Chart, this very well known chart with an array of 4 x 6 color patches, is an icon of the imaging industry. It was presented in a 1976 article (this is more than 30 years ago!) by C. S. McCamy and his colleagues from the Macbeth Company, a Division of Kollmorgen Corporation at the time:
Mr. McCamy is photographed with the ColorChecker
during a visit at the Munsell Color Science Laboratory, in 2002: The ColorChecker consists of a series of six gray patches, plus typical additive (Red-Green-Blue) and subtractive (Cyan-Magenta-Yellow) primaries, plus other "natural" colors such as light and dark skin, sky-blue, foliage, etc. The color pigments were selected for optimum color constancy when comparing pictures of the chart with pictures of the natural colors... as reproduced on color film! Optimizing the human visual match was thus not the first priority; still, it was shown, by the chart designers, that the degree of metamerism was also very small when directly comparing the chart to the natural colors. Expressed otherwise, the perceived colors of the ColorChecker vary in the same way as the natural colors they represent when the light source changes*, either when imaged on film or compared directly.
Traditionally seen in photography magazines, in articles dealing with color film rendition, the chart has found a renewed interest with the advent of digital cameras. In view of what was mentioned above relative to the ColorChecker being optimized for film based reproduction, it could be argued that this optimization would not be ideal for the sensors of these new cameras. However, because the imaging sub-systems of all digital cameras are specifically designed to mimic the response of the human visual system, we see no major reason why the ColorChecker's natural color patches should not be trusted as valid substitutes for their natural counterparts in various lighting conditions**. And what about the other non-natural colors, the additive and subtractive primaries, and the gray scale patches, which are not designed to be equivalent to a particular real-life color? They were valid then, and they still are. In reality, the patches of a color chart do not need to be surrogates of natural objects to make them useful, but the fact that some of the ColorCheker patches can be used in such a way is a definite plus.
As many of us know, the various brands of digital cameras are like film; each has its own color rendering characteristics, which is constantly modified by automatic white balance software within the camera. Basically, the accuracy of the colors at the output is a moving target. The problem is such that many high end cameras provide a way to get the "raw" measurements extracted from the camera sensor, without any processing; the file format is not surprisingly called "RAW" (RAW as in non-gamma-corrected, non-color-corrected, high resolution - much more than 3 x 8-bit RGB - data from the camera sensor). The user can then manipulate the data and try to extract accurate color information over a wide dynamic range. Where we just had to take the picture and let the photo finisher color balance our print, we now have to process the colors ourselves; such is progress.... One way of helping the process of obtaining color fidelity is to
use the ColorChecker as a target within the scene. Other charts with more
patches have been devised, but this simple chart satisfies the needs of many
since you can make a rapid judgment by just looking at it. A smaller, more
portable, version of the ColorChecker is also available. And to illustrate how
much we "know" this chart, here are three different layouts of its patches: the
first one shows patches sorted in a* (the a* of L*a*b*, from green to red, and
from top to bottom, then left to right); the middle one is the usual chart
layout, and the third one shows patches sorted in hue (the h* of L*C*h*). The
first and third layouts do not look as familiar! Where to buy?Before the Internet, it was quite difficult to find where to buy the chart. With the wide spread of the Web plus the renewed interest due to digital cameras, it can be found at many places. You can use the "Dealer Locator" on the X-Rite Web site. Here are a few of them (use these links at your own risk, and please send me an e-mail if you have a bad experience!):
ColorChecker dataHere is a tutorial-type article which presents the RGB coordinates the ColorChecker patches in 4 common RGB spaces (Apple, Adobe (1998), ProPhoto and sRGB); additional data for these and 15 other RGB spaces can be found in the accompanying spreadsheet. The coordinates were derived from the new GretagMacbeth (2005) L*a*b* D50 reference data and from our averaged spectral measurements (now from 20 charts); the two data sets are compared against one another and against other references, such as the 1976 article. For those who want to dig deeper, the document also describes the data conversion procedures and presents all the required equations.
Most ColorChecker owners will have a need for the actual RGB values of the patches at one time or another. This used to be hard to find information since, until late 2005, the RGB values provided with the chart corresponded to no known RGB space. The only reliable reference data was xyY coordinates determined with Illuminant-C, as they appeared in the 1976 article. It was, and still is, possible to find, on the public Web and in scientific sources, published data obtained by people who measured their chart. While OK in principle, this is not enough (for me at least). How old was the chart? Is the chart representative of the production average? How precise was the measuring instrument? Was the instrument calibrated? How was the data processed afterwards? Should I say more? These and other factors make the results of a single measurement unreliable. This was the justification for writing the first version of the "RGB coordinates of the Macbeth ColorChecker" article presented above. This was also the reason behind the project to average spectral data provided by chart owners. Why is spectral data important? Because spectral information gives you the capability to determine the color coordinates as seen under any illuminant, and thus any RGB space, with more precision. Analysis of the spectral data also gives some hints that various pigments are/were/could-have-been used to manufacture the charts. This figure shows two types of spectrums seen for the light-skin patch; these two types are seen with both the Standard and Mini charts, with either new or somewhat older (3 years +) charts. Other patches show a similar behavior. It is important to note that these different spectrums do not necessarily result in distinctly different colors (or color coordinates) and other factors contribute to make the spectral data more fuzzy, such as storage, frequency of use/light exposure, and chart edition (manufacturing variability). One of the nice surprises in this analysis is that the charts seem to age very slowly when kept in dark storage and used on an occasional basis, which is the case for most serious amateurs and pros. A 12 years old chart (well kept, I insist) has 23 patches out of 24 within the standard deviation! When this project began, we could compare only the data derived from the original 1976 article to a handful of average measurements. Because experimental measurements with a few samples cannot be considered reliable, I placed, at the time, more importance on the derived data than on the measured data. As the number of charts measured grew, it became obvious that the measured data was getting stable but still different from the original reference values. While the accumulating evidence was that user measurements were a more accurate assessment of the "average" chart than the data in the 1976 article, RGB coordinates and images (synthetic charts) determined using both approaches were still made available on this site so that users could select the ones they preferred. Around October 2005, GretagMacbeth updated its reference data on the ColorChecker and made it publicly available (to see it, click on the Colorimetric Data tab of this link). This data includes L*a*b* D50 and sRGB values. However, they did not include RGB values for Adobe(1998), ProPhoto or any other RGB space. The new GretagMacbeth L*a*b* D50 data matches very well our average measurements. On the other hand, both our average measurements and the new GretagMacbeth data do not match that well what was derived from the xyY coordinates of the 1976 article. These comparisons are discussed further in the revised pdf. As a result, we now consider the 1976 data as obsolete; nonetheless, for completeness, the coordinates derived from the 1976 data can still be found in the Excel spreadsheet presented above. For the updated article, we used the GretagMacbeth L*a*b* D50 data to derive the RGB coordinates for Adobe(1998), Apple, ProPhoto, and sRGB (even if coordinates for this last space are proposed by GretagMacbeth). Of course, we also provide the RGB coordinates obtained from our averaged measurements for comparison purposes. In addition, the RGB data is now presented in both 8-bit and 16-bit format, the later in full precision, not just a multiple of the 8-bit format, i.e. not just 8-bit values x 256. Finally, we saw that while the match between the RGB coordinates derived from the GretagMacbeth L*a*b* D50 data and the RGB coordinates determined from our average spectral measurements was quite good, the match between the sRGB values given by GretagMacbeth and the sRGB coordinates determined from average measurements was not as good. The numbers we obtain show it is likely that the GretagMacbeth sRGB data was derived from another data set than their L*a*b* data; much more details on this analysis are given in the pdf. We thus prefer the sRGB coordinates we derived from L*a*b* D50 to the ones they published. As for which data we think is better, the data from our average measurements or the new GretagMacbeth values, here are suggestions:
Visual comparisons between measured data and the average chart (March 2007)
The three PatchTool screenshots above are examples of
individual charts compared to the current average of 20 charts. The first
example is from the chart whose average error is the smallest relative to the
average chart, i.e. the "best". The second screenshot compares a typical "middle
of the pack" chart to the average, i.e. the "median". The third one compares the
chart which is the most different from the average, i.e. the "worst". In these
screenshots, each square is made of two triangles such as this one: For many patches, especially the ones of the "best" and the "median" charts, the triangles cannot be distinguished since the difference is very small; enlarging the chart so that it fills the screen will help, as well as taking the time to look for differences (5 to 10 seconds for each patch). To help classifying the patches, colored borders are added when color difference thresholds are exceeded. These borders are yellow and red in the above examples, with the yellow border applied to all patches whose difference is larger than the value assigned to the lower threshold but smaller or equal to the higher threshold, and the red border applied when the color difference exceeds the higher threshold. In practice, 50% of the charts should be similar to, or better than, the median chart, with essentially one or two color differences which are more easily noticeable. The next three screenshots respectively show the standard deviation in Lightness (L*), Chroma (C*), and hue (h*) computed from the 20 charts. The top triangle shows the average value minus the standard deviation, and the bottom triangle shows the average value plus the standard deviation. For a given screenshot, only one parameter is different, either L*, C*, or h*, and the other two are the same. Expressed in terms of statistics, there is 68% probability that your chart patch is between the top and bottom triangle for this parameter.
As we see in the screenshots, the Lightness variation is the most important factor, followed by Chroma, then hue. The Chroma and hue differences are particularly small and difficult to notice. In the third screenshot, all neutral patches are assigned a red border, corresponding to an apparently large standard deviation in hue higher than 20, even though the difference is not visually noticeable; this is expected for very neutral tones, near the illuminant, for which the Chroma (saturation) is very low. Finally, to get an idea of the maximum color difference that you could encounter between your ColorChecker and the average values, we used the CIEDE2000 color difference formula and, for each patch, compared the average value (click here for the average file) to the individual data of the 20 charts. We then selected the "worst case" for each patch, and built a composite color list (click here for the "worst case" file); this list is made of patches from 12 of the 20 charts (i.e. all the patches of the other 8 charts are better than these worst cases). The top triangle is the "BabelColor Average" patch, and the bottom one is the "worst case" patch. Please note that the "worst case" patches are NOT representative of the statistical "standard deviation", which is much less, as can be seen in the three screenshots just above.
You will (hopefully!) never obtain a comparison as in this worst case screenshot when comparing your chart to the average, since a typical chart may contain between zero and four similar patches, and finding between 0 and 2 is more likely. Also, you will note that six patches have a CIEDE2000 color difference of less than one, which is excellent, especially considering these are the "worst cases"! ColorChecker imagesClick on one of the following ColorChecker icons to see the image, or right-click (control + click on a Mac) to download the linked file or the target. The QuickTime plug-in may be required to view TIF images in a browser, which will not show the colors correctly anyway since these images have NO embedded profiles. In addition, 16-bit TIF files are not well supported in Web browsers. It is thus strongly recommended to first download these images and open them in a graphics editing program afterwards:
The "Derived from GretagMacbeth L*a*b* D50 data" images were obtained by starting with the recently published (2005) GretagMacbeth data, i.e. L*a*b* D50, which was converted to D65 with the Bradford Chromatic adaptation matrix when required, and finally converted to RGB. Because, as we mentioned, these images have NO embedded profiles, each of them should be assigned its respective RGB space profile (the Adobe RGB profile to the Adobe file, etc.) using Photoshop or any other program which can assign a color profile to an image. RGB coordinates tables for these and fifteen other spaces are also available in an Excel spreadsheet. The "Derived from the spectral average of 20 charts" images were obtained by averaging the data sent to me by YOU! This data was extracted from the spectrums of Standard and Mini ColorCheckers charts. RGB tables for these and fifteen other spaces are also contained in the Excel spreadsheet. Not decided on which image is better? Read this. These images and spreadsheet will be updated as I get more data. Thanks in advance! If you are not satisfied by the looks of the chart image on your display, you may be interested in an article by Charles Maurer in Tidbits, an online publication for the Macintosh internet community. His article, Colour & Computers, has a few paragraphs on his experience of the various renderings of the ColorChecker chart obtained with different flavors of the sRGB profile and different applications. Do you want to provide some data?You can contribute to this study if you have a ColorChecker chart (Standard or Mini), and if you can generate a spectral data file from a spectrophotometer. I will compile the results and make them available on this page. If you have a colorimeter with which you can measure only the color coordinates, without spectral data, well, send your data also! Suggested file formats: *.txt (plain ASCII format used by Measure Tool,
part of
ProfileMaker), *.cxf (XML format used by Eye-One Share and other GretagMacbeth
software), *.xls (standard Microsoft Excel format that can be exported by
Eye-One Share), and any other standard
text based file format from any other software/spectrophotometer. Please send the file at the following address (click
the button to send an e-mail):
Using the chart to auto-correct images or make ICC profilesBecause the ColorChecker patches cover quite a large color gamut, not with a lot of samples mind you, one can think of using them as a reference to correct color casts (or bad lighting). This is what is done with:
inCamera is a Photoshop plug-in designed for the sole function of generating ICC profiles for digital cameras and scanners from captured images of color charts. These ICC profiles can be applied to similar images taken in the same lighting conditions. Apart from the standard ColorChecker, it supports most commercial charts, such as the ColorChecker SG, a chart which comprises 140 semi-gloss patches which should, and does, provide better quality profiles, and industry standard IT8 targets. Picture Window Pro has a "Match Reference" feature which generates a color transformation based on the captured image of the ColorChecker chart. Once the transform is generated, you can rapidly correct images taken in the same lighting conditions. This feature is very similar to inCamera except that the correction is a command instead of an ICC profile. Picture Window Pro has many other useful tools and features for the digital photographer, such as perspective correction, selective color control, a complete set of mask tools, and 48 bit (3 x 16 bit) files support. ProfileMaker Pro is a very complete high-end measuring and profiling package. The Digital Camera module, available in the PhotoStudio bundle, can generate an ICC profile for a camera using the ColorChecker as well as the ColorChecker SG or the older ColorChecker DC (at least in version 5.0.5b and 5.0.8). Please note that the ColorChecker is supported even if only the ColorChecker SG is mentioned in online specs; however, we can confirm it is supported in the Profile Maker 5 multi-lingual Quick Start Guide (see for instance page 29 in the English section). For a lower cost solution that combines software with the measuring instrument, have a look at the i1XTreme bundle, which will make profiles from the ColorChecker SG (purchased separately), but not the ColorChecker. The Monaco PROFILER can generate an input device ICC profile from many chart types, including the standard ColorChecker chart, even if some spec sheet do not mention this chart explicitly. Of course, the software can do much more, and it includes many tools to create high quality profiles for monitors and color output devices (printers and presses), but its price is in consequence. The Capture Studio software from Kodak is a Mac only product dedicated to the RAW file format used in the Professional DCS Pro Back line of cameras (which is now obsolete but support is still available). One of its features is ICC profile generation from a ColorChecker. This free software can be downloaded from the DCS Cameras support page (look for Software Downloads/Camera Software, or click this link ); registration is required. If you own Photoshop and have a digital camera that outputs RAW files, you should try ACR Calibrator, a free Photoshop script written by Thomas Fors, which automates the process of using the standard 24 patches ColorChecker for calibrating RAW file via the Adobe Camera Raw plug-in (known as ACR). With all the above software, you generate your color correction transform or profile by incorporating the ColorChecker card in your photographed scene, and by analyzing the chart patches in the resulting image (this also works for calibrating a scanner). They are best used when you can manually adjust and freeze your digital camera white balance. If your digital camera does not enable you to control white balance, then you may still obtain good results if the lighting AND the color content of your scenes do not vary too much. These tools can also be used with traditional film if the lighting conditions do not change during the photo session (the white balance does not change in a traditional film, it is just not balanced!). Of course, you need a new profile each time you change the lighting conditions, and some may feel that generating and managing many ICC profiles--lighting does vary a lot--could become cumbersome. Having a set of transforms, as with Picture Window Pro. or an automated script, such as ACR Calibrator, are two alternatives to eliminate that burden. There is also a simple method of achieving white balance using a single neutral patch, one of the ColorChecker neutral patches for instance (!), with the dedicated white balancing tool of programs such as Aperture, Bibble, Capture One (or C1), and Lightroom, and Photoshop ACR. Once color balance is achieved by using the target in a first image, the settings can be saved and applied on subsequent images taken in the same conditions.
Of course you do not absolutely need a ColorChecker to calibrate RAW files, as you can manually tweak the RAW import parameters. This method, however, can be more demanding than the ones presented above, and it is important to note that RAW files are NOT standardized, with each digital camera company proposing its own variant, some of them proprietary (see the OpenRAW web site for more info on this potential problem). Adobe has proposed the "public" Digital Negative (DNG) archive format to address the issue. In any case, using the ColorChecker within your images, as part of the RAW file manipulation process, or with any other file format, will definitely provide helpful visual feedback. Finally, even with all these tools, there are pictures which remain difficult to correct. Cal McCamy's picture shown above is one such case. Taken under fluorescent lighting and saved under a JPG format, it cannot be corrected with a simple click. The old adage which says that it is preferable to place a filter on the lens than to correct the picture afterwards applies here; however, such filters are not universal to all types of fluorescent lamps (and there are many) and are simply not part of most photographers' accessory bag. Using the chart to analyze your camera's performanceWould you like to measure your camera's color quality, noise, and tonal response? Then you can use the Colorcheck module of the Imatest program offered by Norman Koren. The Colorcheck program uses images of the ColorChecker chart to perform its analysis. Use this program to check how the automatic white balance feature of your camera is working. The other Imatest modules are dedicated to measuring the sharpness and gray response using other types of targets (you can print some of the targets). This program is becoming a reference for digital camera analysis, as can be seen on the imaging-resource.com site. If you are interested in photography in general, and in its technical aspects in particular, you have to go on Norman's namesake site, www.normankoren.com, where you will find some of his images, which can be purchased, and a tremendous amount of free technical info on digital photography. The problem with the ColorChecker SGThe ColorChecker SG is a new chart offered by GretagMacbeth and dedicated to digital photography. It has a Semi-Gloss finish with highly saturated colors. On the chart Web page we read that it "Includes standard ColorChecker chart colors," and the corresponding patches are laid out on the SG chart in the same configuration as in the standard or Mini ColorCheckers. Many would assume, by the layout and these words, re-affirmed in the ProfileMaker Pro manual, that they are the same, but this is not the case. Except for the bluish-green, white and four grey patches, all the other patches are relatively far from the original chart colors (with an average DeltaE*ab difference of about 7 with the standard ColorChecker). This has been confirmed from measurements on two charts (the measured values of these two charts are very close to one another however, which is good since it shows manufacturing consistency). Because of this, it is not recommended to use the spectrums and RGB values given in this Web site for comparison with the "equivalent" patches of the SG chart. Note: these comments have nothing to do with the quality or utility of the SG card per se. | |||||||||||||||||||||||||
|
, with the
top triangle assigned to the average.
