Macbethcal Reflectances

Dear Group,

I am using macbethcal/pcomb to determine the color and reflectance of various scanned paint swatches. One thing that has come to my attention is that the overall reflectance seems out of range. For example, a creamy white color once calibrated has an overall reflectance (determined by ximage and grey(r,g,b)) equal to .93. From RWR I think it was mentioned that a reflectance of greater than 80% may be unrealistic. Although I have had some colors out of gamut, particularly Cyan,(noticed by looking at the debug.pic from macbethcal), there seems to have been no apparent issues with the white, neutral 8, neutral 6.5, neutral 5, neutral 3.5, black (I am assuming that these patches are used to calibrate the reflectance) during the generation of the calibration file. Has anyone else encountered this issue?

Regards,

Marcus

Hi Marcus,

You must ensure that there are no lighting or gain changes between the scanning or capture of the Macbeth chart and the swatches, otherwise you will have an absolute calibration error. If you are using a flatbed scanner, set the exposure to manual and fix it at a single value. If you are using a camera, set it on manual mode and use a constant illumination with the chart in the same position as the swatches in your two photographs. I often shoot outdoors to assure even illumination.

This is the most likely culprit.

-Greg

···

From: "Marcus Jacobs" <[email protected]>
Date: October 18, 2005 12:20:13 PM PDT

Dear Group,

I am using macbethcal/pcomb to determine the color and reflectance of various scanned paint swatches. One thing that has come to my attention is that the overall reflectance seems out of range. For example, a creamy white color once calibrated has an overall reflectance (determined by ximage and grey(r,g,b)) equal to .93. From RWR I think it was mentioned that a reflectance of greater than 80% may be unrealistic. Although I have had some colors out of gamut, particularly Cyan,(noticed by looking at the debug.pic from macbethcal), there seems to have been no apparent issues with the white, neutral 8, neutral 6.5, neutral 5, neutral 3.5, black (I am assuming that these patches are used to calibrate the reflectance) during the generation of the calibration file. Has anyone else encountered this issue?

Regards,

Marcus

Greg,

I decided to go back and try to set up another calibration file. I used the same steps that I used before. The software associated with the flatbed scanner has no direct exposure controls, however there is an option (which I select) to do no color correction. This, I believe, prevents the enactment of the auto exposure feature. Here are the results (debug.pic) from my latest scan Macbeth color chart scan:

As you can see, many colors were out of gamut however the white and neutrals were able to be calibrated. After running pcomb with my color swatch, the reflectances were still high (0.80 < x < 0.95 on a scale from 0 - 1). Here is the paint color swatch before calibration:

and after calibration

I usually take the sample and the center of each. Do you believe that the overall reflectances are valid or out of range?

Thanks

Marcus

Hi Marcus,

Without seeing your swatches or having a reflectance measurement to compare to, it's difficult for me to guess whether the calibration is working or not. Certainly, the debug output from Macbethcal indicates that it was able to do a reasonable job of inverting your scanner's response. However, it's still possible that some kind of auto-exposure is changing the results from one scan to the next, so it may be that your absolute levels are off.

-Greg

···

From: "Marcus Jacobs" <[email protected]>
Date: December 23, 2005 10:44:09 AM PST

Greg,

I decided to go back and try to set up another calibration file. I used the same steps that I used before. The software associated with the flatbed scanner has no direct exposure controls, however there is an option (which I select) to do no color correction. This, I believe, prevents the enactment of the auto exposure feature. Here are the results (debug.pic) from my latest scan Macbeth color chart scan:

http://i35.photobucket.com/albums/d178/marcdevon/MacbethColorChart_12212005002_templ.jpg

As you can see, many colors were out of gamut however the white and neutrals were able to be calibrated. After running pcomb with my color swatch, the reflectances were still high (0.80 < x < 0.95 on a scale from 0 - 1). Here is the paint color swatch before calibration:

http://i35.photobucket.com/albums/d178/marcdevon/Behr_White_before.jpg

and after calibration

http://i35.photobucket.com/albums/d178/marcdevon/Behr_White_after.jpg

I usually take the sample and the center of each. Do you believe that the overall reflectances are valid or out of range?

Thanks

Marcus