I'm an architectural photographer and would like to just use Fuji chrome film because I like it better for exteriors. My problem is I have discovered that Fuji seems to have inherantly more green so that when shooting under fluorescent lighting my color meter is not accurate. If the meter says I should add 20 magenta to an interior exposeure I really have to add 35 or 40 magenta if I am using Fuji chrome. I know Fuji has technical sheets telling how to filter depending on the particular type of fluorescent bulb in question but then I have to go through the hassle of determinging what the bulbs are. If they are not all the same I am in trouble. Has anyone out there run into this problem and come up with any ideas or tables that are helpful. Thanks

-- Bob Freund (SPDutton@msn.com), September 17, 1999

Answers

Which Fujichrome? Velvia daylight flourescent: 40R+10M; Cool White:40M+5R; White: 40M+10B; Warm white: #80C+25M

RDPIII:(same sequence): 30R+10M; 30M; 15M+20B; #80B;

RAP: (same sequence): 35R; 30M; 15M+15B; #80B

These are according to the Fuji Pocket guide.

You might want to have your color temp meter checked out.

-- Ellis Vener (evphoto@insync.net), September 17, 1999.

Ellis is right-on regarding the filtration. However I will add that these are _starting points_ for filtration, and for critical work, the photographer should do some field tests. This is important to remember because (for America) the main lamp manufacturers (General Electric, Philips, and Osram Sylvania) do not have a single lamp color standard that the all conform to. This means that a 3500K lamp by one manufacturer will have a different plot from the next, in any color space you plot it into. That's an engineering way of saying that the lamps from one manufacturer may look more or less the same to the human eye, but you will find that films will "see" them differently.

I'm certain this is the case everywhere else, also.

This is independant of the individual film characteristics, which are noted by the different filtration guidelines printed on the tech sheets, and kindly stated by Ellis.

---Michael

-- Michael Mutmansky (psu4ever@ix.netcom.com), September 17, 1999.

re: "I have discovered that Fuji seems to have inherantly more green so that when shooting under fluorescent lighting my color meter is not accurate."

It was my understand that color temp meters measure the relative color temp of sources based on heated filaments (Quartz Halogen, the Sun, Strobe) with some accuracy and aren't that accurate with fluorescent or the various non-full spectrum discharge type lights.

-- Larry Huppert (Larry.Huppert@mail.com), September 18, 1999.

Well, for what it's worth, during the past two weeks I photographed separate industrial and architectural assignments on location using Fuji Astia 120 size in my Hasselblad. Photographing under flourescent, Tungsten and Metal Halide light sources, my Minolta Color Meter 4 was absolutely accurate regarding the filter corrections it recommended.

-- Craig Maxwell (cm@craigmaxwellphoto.com), September 20, 1999.

Hi Bob;

With regard to filtration for specific films and situations, I'll defer to others (such as Ellis) with more experience.

With regard to your color meter problems, the short answer is that color meters only yield accurate results only for blackbody (broad-spectrum) light sources such as the Sun, skylight, tungsten bulbs, etc. Color meters DO NOT yield accurate or useful results when confronted with a line-emission source like a fluorescent bulb or soium-discharge lamp.

The long answer is that color meters work by measuring relative intensities over a few (typically 3) broad bands of the spectrum. They then infer the behavior of the light source from this small set of measurements, based on the ASSUMPTION that the light source exhibits the characteristically broad spectral distribution of a blackbody radiator. The problem is that fluorescent lamps and vapor-discharge lamps emit light in a series of narrow bands, and therefore don't look anything like a blackbody radiator.

On the measurement side, the solution to the problem is a nifty piece of hardware called a photospectrometer. They aren't cheap (the Gretag unit I use here at work costs something like $7000), but they have the ability to directly measure spectral distributions with high enough precision to accurately characterize the individual peaks which make up the emission spectrum of a line source such as a fluorescent bulb.

Even with a photospectrometer, though, your problems don't end: The basic design of color film itself also incorporates the assumption that whatever it images is illuminated by a broad-spectrum source. Films are just like color meters in the sense they also attempt to represent the full gamut of colors as a linear combination of 3 broad spectra (the dyes), which are stimulated via the 3 color-sensitive emulsion layers. When you are dealing with a line-spectrum source such as a fluorescent bulb, very small changes location of the bulb's emission peaks can dramatically change the ways in which they stimulate (or don't stimulate) the different film emulsion layers, and therefore the color balance of the final image. In order to predict these changes, you need to know both the spectral sensitivity distributions of the individual layers, and the spectral density distributions of the corresponding dyes. This is why the filtration numbers suggested by Ellis vary so dramatically from film to film, even when the light source is held constant.

Coming back to reality, the bottom line is that you really have 3 choices here:

1. Get a photospectrometer and the spectral sensitivity and spectral dye-density curves of all of the films you want to use, and learn enough about color reproduction that you can actually do something useful (i.e. out in the field) with all of the resulting information.

2. Go to the site of your photograph in advance, shoot a test roll (or 2, or 3, or 10), and figure out what filtration you need.

3. Use print film, do the best filtering you can on-site, and fix the rest in printing. ;-)

-- Patrick

-- Patrick Chase (patrick@sdd.hp.com), September 21, 1999.