Hello all!

I've been taking digital photographs of fabrics for our online store at http://www.equilter.com for over a year now, and I've photographed more than 4000 of them to date. I currently use an Olympus C2500L, as I've tried many of the top names and models, and I find the results from the Olympus require less digital processing in Photoshop.

I have a photography setup using special fluorescent lighting: the Plume Scandles task lights at http://www.plumeltd.com/scandlestask.htm I've also calibrated the camera's white balance, and use the camera in manual mode, so that I can better control the results with my constant setup.

Since I digitally correct every photo, I have lots of experience with what needs correcting. Here are my general findings: 1. Purples always come out too blue. 2. Blues are frequently too light. 3. Pinks come out rather desaturated. 4. Reds are too dark. 5. Yellows are too light. 6. Oranges are too yellow. 7. All images generally require extra saturation.

I'm wondering if anyone has any suggestions for set-up or equipment changes that might help here.

Do the CCD's on these digital devices deteriorate with age?

Any suggestions would be greatly appreciated.

Sincerely, Paul Rubin prubin@equilter.com http://www.equilter.com

-- Paul Rubin (prubin@equilter.com), May 08, 2000

Answers

Paul, most of the hue-errors, especially with the cutted hues (orange, violet etc) are due to the incorrect transfer function that all the digicams have. You would benefit a lot if you purchace the Q-60 Reflective Color Input Target from Kodak (only 29USD) and create a calibration for your camera that you then apply using an Action in Photoshop. I have very detailed instructions for such calibration at: http://www.aim-dtp.net/aim/calibration/index.htm click the link: ---Kodak Q-60 usage--- I will gladly assist you if you choose to create the calibration, you will then see the objects terribly similarly on the CRT as they appear in the real-life.

Some of the hue-errors can be from the Scandles lighting, I do not believe they are 5300K full-spectrum as it says on their site. Please post the type markings of the fluorescent tubes they use then we can look up what they really are.

About the saturation: This is a relative issue, question of color mapping. It is good that an acquire device captures a wide color gamut since the visible gamut is so very large. Now when the wide captured gamut is shown on the CRT that has very limited color gamut compared to the visible gamut then we judge that the saturation of the image is too weak. Depending on how saturated colors there was in the original scene we can increase the saturation more or less, too heavy saturation increase will eventually wipe out image detail since those hues where the increased saturation exceeds the capability of the CRT will be mapped incorrectly. Now, if the acquire device would capture using a small color gamut, it would give more saturated images but then some objects with very strong saturation would appear just simiarly as they appear after too strong saturation increse in digical editing.

>Do the CCD's on these digital devices deteriorate with age?

They do not deteriorate in the common sense. Direct sunlight can burn the color-mask (alther it's hues). CCDs can fail similarly like what ever electrical component can fail, but that is very rare. Every electical component has a life-time but it is tens or hundreds of years.

Timo Autiokari http://www.aim-dtp.net

-- Timo Autiokari (timothy@clinet.fi), May 09, 2000.

I'd also suspect that any fluorescent lighting isn't going to give you a full continuous spectrum.

Another aspect that needs to be considered is the UV and Infra-red reflectivity of your samples. Although invisible to the eye, the camera will pick up IR and UV and display it as spurious colour.

I'm not sure that you aren't wasting your time trying to get the colours exactly matched, though. If your customers haven't all got calibrated monitors, they won't see anything like the original colour anyway.

-- Pete Andrews (p.l.andrews@bham.ac.uk), May 09, 2000.

Fluorescent lighting is always an approximation for daylight, but there are tubes that do give good approximation. The correlated color temperature (CCT) has to be over 5000K and the color rendering index (CRI) must be 98 or better. E.g. the Dulux /12 tubes are such. CRI=92 is not good at all.

Colors on uncalibrated CRT monitors actually are pretty accurate, most of the CRT tubes use same/similar phosphors (Trinitron or P-22 or their derivatives) and the gamma of the vacum tube is a pretty accurately 2.5, it comes from the constant of the physics. The only difference is the gamma that for PC is 2.5 and for Mac 1.72 because Mac OS by default applies a partial gamma correction to the data that is going to the monitor. When one publish calibrated images to Trinitron, D65, gamma 2.5 then there will be very good accuracy for large majority of visitors.

Timo Autiokari

-- Timo Autiokari (timothy@clinet.fi), May 10, 2000.