I can't believe my eyes! A Shutterbug contributor and forum moderator singing the praises of SoLux! Maybe there is hope after all. If only you would now calibrate you monitor to 5000K to match your lighting or am I asking too much? Anyway, Larry, welcome to the light side.

I can't imagine setting my monitor to 5000K. My entire business revolves around preparing accurate sRGB files for digital jurying, which is 6500K. What I see on my monitor matches exactly the images I project with my digital projector set to sRGB.

I just need a way to accurately view prints. According to Andrew Rodney, the difference in color temperature shouldn't matter.

Larry,

Print viewing booths with D65 lighting are available. If I recall either B&H or Adorama lists them. However because they are not an item sold in large quantities they tend to be pricey. Very likely a Google search would reveal a source for D65 cold cathode tubes as they are what is used for the backlight in LCD displays, and are produced to ANSI standards for that purpose.

Most of the color management industry experts I have corresponded with agree with the point Andrew Rodney makes, Including Steve Upton President of Chromix.com. Human perception compensates in the brain almost instantly for changes in color temperature, this was cited in Richard Zakia's book on perception for photographers published originally many years ago by RIT Press and is still available in a newer edition.

Larry,

I'm afraid the "experts" may not be totally correct, to use a euphemism. You are not going to have a good monitor-to-print match if your monitor is calibrated to 6500K and you illuminate your prints with 5000K. True, the brain compensates for color changes, also known as color constancy, but when you have your print viewing area right next to your monitor (the ideal solution in my opinion) then your brain cannot compensate, because both the monitor and print are in close proximity. I've worked on this issue intensively and done many tests with many different light sources and monitor calibrations. The main complaint you hear about setting your monitor to 5000K is that it looks too yellowing/reddish, but your brain adjusts, compliments of the same color constancy.

If you develop images for the web, where most monitors are set to 6500K, regrettably, than by all means set your monitor to 6500K as well. If you print images and want to best possible monitor-to-print match, use the best possible lighting, SoLux 5000K and calibrate your monitor to 5000K.

And stay away from any lighting or viewing booth that uses fluorescent bulbs, even the ones specifically designed/promoted for color-critical work; they all, even the best ones, have horrendous spikes in the color spectrum.

Frans,

Are you equating cold cathode tubes with household-type fluorescents? If so then your condemnation has to include the light sources of most scanners and all LCD displays. Do they all exhibit the spikes you are so incensed about?

I don't doubt your claim of studying this subject of print viewing illumination intensely. However inherent to what you say and have said on the subject in the past, is the possibility you have not used everything that is available nor studied and tested all of the currently available light sources.

David,

All fluorescent tubes - whether equipped with a heated filament or cold cathode, whether designed for household use or color-critical applications like viewing booths, scanners or LCD backlighting - all have enormous spikes in their color spectra. If you don't believe me then get from manufacturers like Philips, Sylvania, Gretag MacBeth, etc. power density curves for those tubes designed specifically for color-critical applications. And yes, those lightsources have severe implications for viewing booths, scanners, LCD monitors and the like. And yes, LCD manufacturers like NEC and Eizo offer LED backlit LCD monitors to get rid of the spikes and improve the gamut range. By using spiked lightsources to view your prints in addition to the use of spiked lightsources in your monitor or scanner, you aggravate the situation.

Rather than questioning if I have used and studied "everything that is available", why not recommend exactly what kind of digital darkroom lighting you recommend for Shutterbug forum participants to use and what your rationale for those recommendation are.

I think for the two or three times a year I have to visually judge print color, the desk lamp and D50 bulb I ordered from Solux will work just fine.

Now back to the entertainment.

David,

I would appreciate it if you would react to my post of December 9 on this subject.

I asked David Brooks on December 9: "Rather than questioning if I have used and studied "everything that is available", why not recommend exactly what kind of digital darkroom lighting you recommend for Shutterbug forum participants to use and what your rationale for those recommendation are."

Well, David told me in a private email that he is not going to answer that question. To quote him verbatim: "You are just full of a lot of hot air, and think you know a lot more than you do." So there you have it. Expect no recommendation for digital darkroom lighting from David. I'm not too terribly surprised because during the 5 years or so that he and I have exchanged ideas on this subject, to say it mildly, he has never come up with any useful recommendation, nada, zilch. The only product he mentioned was a fluorescent tube for which he had no technical specifications and that is not available in the US. Too bad. It looks like my recommendation to use SoLux 5000K bulbs remains unchallenged.

High quality fluorescent tubes all have Colour Rendering Index (CRI) ratings. A rating of 100 is the equivalent to continuous spectrum sources - which may never be achieved with fluorescents. While the spectrum when analyzed will still show spikes, with a high CRI the spikes are well enough distributed that they will not cause significant colour shifts for either viewing or shooting. Manufacturers are now supplying a wide variety of fluorescent lighting units for photography, motion picture and network TV studios where colour is critical. (Arri, Bowens, Lowell, Photogenic, Sunpak...)

Tubes in the 90+ range are generally considered more than accurate enough for viewing or as a source for shooting with transparency film, though some have written that a CRI of 85 is the threshold for critical colour. Since tubes in the 95 to 98 CRI range are available and inexpensive, I see little reason to go with lower rated tubes. The lower one moves on the CRI scale, the more unpredictable are the results.

When shooting digital RAW, the CRI can be considerably lower before they become unusable with current sensors. Common household and commercial tubes tend to be in the 60 CRI range, if rated at all. I suspect many are much worse. While my first digital camera had great difficulty with extremely cheap fluorescent lights, my current camera gets fine skin tones under the exact same lights. While they eye compensates for the discontinuous spectrum of fluorescent light, cameras - specially when shooting chromes - do not. Slide film is totally unforgiving and fluorescent filters only work well with some sources.

For shooting or viewing, tubes with a CRI in the 90s are not significantly more expensive and have an accurate equivalent to colour temperature. An example would be
32W T8 FLUORESCENT 5000K 98CRI

Larry,

Fluorescent tubes may work well for you and many others, but there are significant drawbacks to these products as compared to low-voltage tungsten-hologen lamps like those from SoLux.

An independent field study measured and analyzed 61 critical viewing booths at 14 different companies using fluorescent lamps from 6 manufacturers and low-voltage tungsten-halogen lamps from SoLux. This study found the following:
- the fluorescent lamps showed Delta E color errors in the range of 0.7 to 3.19
- the SoLux lamps showed Delta E color errors in the range of 0.16 to 0.53
A Delta E of 1 or more causes colors to be seen as different. The study further concluded that the higher color error of the fluorescent lamps is caused by their spiky spectral energy distribution and that fluorescent lamps have sufficiently different rendering properties to produce visually different color appearances for as many as 50% of the colors.

Remember that the fluorescent lamps involved in this study are specifically designed and specified for color-critical applications. A SoLux fixture including the bulb can be had for as little as $75, so this seems to me an ideal solution for the photography enthousiast and professional for print viewing.

Frans Waterlander