What's the REAL color spectrum of the CF daylight bulbs

[michaelbsc]

Hey guys, I need to make some color copies of a few things, and I'm beginning to talk my self out of using the now prolific "daylight" 5000K compact fluorescent bulbs that I see everywhere. Normally I don't care about getting the color "right" but this time I do.

Here's my thinking, and if someone can either confirm that I'm right or point out that I'm out to lunch I would appreciate it.

Almost everything I do is B&W, and the few occasions I do drag out and put up a copy stand I use the CF curly bulbs. They work great and don't heat up my smallish place too much. But some of the family want copies of some departed parents and grandparents, so I'm going to make color negatives.

Now, my question is that all the color film I have is daylight, and my CF bulbs are "daylight" bulbs as well. But thinking about how fluorescent bulbs work, i.e. stimulated electron decay from a specific material, I get the distinct impression that the color spectrum might "look like daylight" to the eye, but might not look all that close to daylight to a piece of film. Specifically, I can imagine that the CF bulbs have giant gaps in the spectral coverage, gaps that a hot glowing tungsten filament would not have.

Obviously, I care about the color for these particular pictures far more than I do about most of the stuff I shoot. Usually I just stick in the film and let it go, and no one complains about the color balance, they're just happy to have the pictures. But I want these to be as right as I can get without spending a fortune.

So, will the CF bulbs really do the trick with daylight film, or should I spring for a brick of real tungsten film?

Somewhere I've got a couple of old movie camera light bars with the massive photo-flood lamps on them. Should I dig that out, too? Or will regular grocery store bulbs work? If I should use the light bars, I know some of the lamps are dead, so what lamps should I use to get as close to tungsten film without filters as possible?

And I have seen some really, seriously blue incandescent bulbs at a camera shop about an hour from here. My assumption was that these are "daylight" tungsten bulbs. Will those work with the daylight film I have? Or do they cost more than tungsten film?

What's the best "dedicated enthusiast" way to do this vs. the prohibitively expensive professional way. (I know, some wag is going to say take them in the back yard and use real daylight. But let's be realistic.)

Thanks,
Michael

 

[Bob-D659]

Well you can/could get 3200k or 3400k tungsten film, and you will need matching bulbs. If you are shooting daylight color neg film, you could try cfls, but get ones with a real high CI number, closer to 100% is better. I haven't seen a spectrograph of any newer bulbs, but I imagine they are pretty peaky. Halogen bulbs would be your best bet, and if you have a variac transformer, you could overvoltage them and get a real nice high colour temp, at the expense of running life.

 

[PanaDP]

I run into fluorescents on film every day at work as a cameraman on movies and TV. The answer is that they don't have a color temperature. Color temperature is a term reserved for incandescent light sources. Fluoros do have a "correlated color temperature", though which is the same thing in practice. Because they aren't a continuous source o light, there are some pretty big spikes in the spectra of light they emit. The worst of these is green.

Get the highest CRI (color rendering index) that you can. It's a percentile scale. Also, try and borrow or rent a color meter and have some minus green gel on hand. Fluoros are rarely green/magenta neutral, even high CRI ones, and will probably need some filtering. If you can't get a hold of a color meter, I would filter the bulbs with 1/4 or 1/2 minus green, that's what is usually needed. Higher CRI will probably need 1/4 and lower will need 1/2. If the CRI is REALLY low (think cheapest fluoro tubes you can buy) they might even need full minus green. You'd have to meter each bulb and filter them separately, since no two are really alike in that respect.

In all honesty, I would use something else. Fluoros are a pain on film, with the exceptions of kinos and others that are designed for it.

 

[JBrunner]

What PanaDP said. We call them a discontinuous spectrum light. The predominant spectrum is what you "see" and that is what they use to determine the "color" of the flo. What you don't readily see until you get used to looking at it are the green, purple and magenta spikes. The exact color and effect of these varies from brand to brand and even with age in the same family. They are not the choice for shooting when color is critical. Negs can be very hard to correct because it is difficult to isolate the effect of the spikes, because they are so narrow, and can include spikes of different colors in the same output.

 

[nworth]

I don't have a daylight CF lamp to look at, but I am looking at an ordinary incandescent substitute with my trusty hand spectroscope. I see a violet line (probably a mercury line), a blue band, a green band, rather separated from the blue band, and a red band that includes some orange and yellow. There is only a small gap between the red band and the green band. The gaps in the spectrum cause aberrations. The light gives a good imitation of a incandescent bulb for non-critical uses, but the spectrum is certainly not continuous. This sometimes become apparent when matching colors from dyes or pigments. The bands are fairly broad, and photographs should work with them in most cases. But some colors may not reproduce as you would see them in daylight due to the blank spots in the spectrum. The light also probably has a different general distribution of power over the spectrum than black body radiation, as you get from an incandescent source - one band may have more or less power than it should. This could also cause some colors to look "off." I suspect the daylight variety of CF uses similar phosphors in a different mix, and it would have similar problems. Someone should look at one with a spectroscope and report. There are phosphors available that do a pretty good job of giving a continuous "daylight" spectrum. Bulbs that use these phosphors usually have some blurb in the name or on the package about "true sunshine" or "natural color", and they usually state a color fidelity index rating. The good ones have a rating above 94. If these are used, we may have a chance.

 

[Heinz_Anderle]

There are fluorescent tubes from General Electric specially suited for motion picture lighting (either for daylight- or tungsten-type film) as well as compact fluorescent lamps e. g. from Osram-Sylvania (Biolux?) or from Narva (Bio-Light) with a color rendering index (CRI) of 960 (= at least 90 %, 6000 K).

I have obtained good results with the Narva lamps on a copy stand with Agfa slide film (but it has to be tested with every type of film).

CRI 827-lamps turn out like mustard or curry; such cheap compact fluorescent lamps are a nightmare even with color print films such as the Fujicolor Superias with a cyan-sensitive color layer.

 

[PanaDP]

There are fluorescent tubes from General Electric specially suited for motion picture lighting (either for daylight- or tungsten-type film) as well as compact fluorescent lamps e. g. from Osram-Sylvania (Biolux?) or from Narva (Bio-Light) with a color rendering index (CRI) of 960 (= at least 90 %, 6000 K).

I have obtained good results with the Narva lamps on a copy stand with Agfa slide film (but it has to be tested with every type of film).

CRI 827-lamps turn out like mustard or curry; such cheap compact fluorescent lamps are a nightmare even with color print films such as the Fujicolor Superias with a cyan-sensitive color layer.

If you want to get perfect color with those, you still need to test them with a color meter. The CRI is assuming a properly working ballast, which many household and industrial fixtures don't have. If the ballast isn't working right, the first sign will be color shifts in the tubes.

 

[michaelbsc]

Sounds to me like the easiest/cheapest method is just buy a brick of Tungsten film and do it the old way. Unless, of course, I can't find the old style photofloods. I know there were available in several "types" that had slightly varying temps.

Can anyone point out which ones were the closest match to Tungsten film without a correcting filter?

MB

 

[David A. Goldfarb]

Some excellent answers above. Buy, rent or borrow a color meter, take some notes, and you'll really get a handle on this issue. Helen Bach was nice enough to lend me a couple of her color meters for a few months a while back, and generally, I found that a household "daylight" fluorescent bulb needed about 30M to get close to photographic daylight, and a household fluorescent bulb that was designed to be close to tungsten also needed about 30M to match tungsten approximately, and then with either, you might need to do some fine tuning in the red-blue spectrum once the magenta-green was sorted out.

"Reveal" incandescent bulbs (the blue ones) aren't that close to daylight. Verilux incandescents, though, aren't so far off, but still might need a little correction to be accurate on film.

I agree that with what you have, Photofloods and tungsten film are probably the way to go. There are 3400K Photofloods (Type A) and 3200K Photofloods (Type B--closer to household bulbs, but household bulbs are usually warmer and less consistent). There are also blue 4800K Photofloods, closer to daylight, but if you've got to filter them anyway to match daylight, it seems easier just to do everything with normal tungsten bulbs. If you're buying new bulbs and new film anyway, then look at the data sheet for the film you want, and buy the bulbs that match.

 

[Tim Gray]

I just whipped this up.

Note, this is not a daylight balanced bulb, its what was in the fixtures of my lab (Philips TL830, 3000 degrees, CRI 85). CRI 85 isn't great, but its not bad. I've read for good color balance, you want to shoot for over 90. 95 is even better and the very expensive bulbs used in movie production are 98 I think. Anyway, I've shot video using 90-93 CRI bulbs and they look pretty dang good. You can even get away with 85. Its the 70's that give a really sickly green cast.

First graph is a spectrometer reading of the above mentioned bulb. The second is a reading from a hand flashlight with a tungsten bulb. Daylight would look similar but shifted to the left.

 

[michaelbsc]

First graph is a spectrometer reading of the above mentioned bulb. The second is a reading from a hand flashlight with a tungsten bulb. Daylight would look similar but shifted to the left.

Yea, these graphs are about what I had envisioned. I'm not going to give old Max Planck a run for his money on black body radiator theory, but I knew fluorescing was a far different mechanism to produce visible light.

For now I'm going to start looking at tungsten film and floodlamp datasheets to see what my cost would be. I have a suspicion that will be the least expensive route for me, and I will have them in the closet next time something comes up. I suppose if I tried to get into this professionally I'd probably spring for a much more progressive set up, but for occasional hobby use tungsten is probably the easiest way to go. And for B&W in the future I can still use my grocery store CFs.

Michael

 

[Tim Gray]

Having used CHEAP fluorescent bulbs in video production, I can tell you that you can get remarkably close with them. I would recommend going to Lowes or Home Depot (try both) and looking at the small T5 bulbs (I think). Hunt around and find the bulb with the highest CRI you can find and a 'color temp' of 5600. You should be able to find something around 95. You can get them small, like 18 or 24" long. Buy two or four, and then go into the light section and find a cheap plastic under-the-cabinet fixture. I found mine for about $20. For very little money, you can get something that is pretty dang close to daylight balanced. Whether or not this would give you the ultimate in color accuracy, I can't say, but for what I think of when I think if white balance, this will get you there.

Something like this fixture:
home depot light

Check out this page for a bit more info:
DIY light

 

[Bruce Watson]

What you want maybe is a Solux bulb. Their light fixtures suck, but the bulbs are first rate. They are for all practical purposes artificial sunlight. The 4700K bulbs are perhaps the best bet, because the 5000K bulbs are more than twice as expensive. Both have a CRI of 100 IIRC. Takes a 12VDC fixture, and of course you need exactly 12VDC to get the color temperature that's advertised, and most low voltage lighting transformers are somewhat downrated (less than 12VDC) to improve lamp life at the expense of color temperature.

I've used these for color proofing for four or five years now. They do and excellent job for me.

 

[PanaDP]

Yeah, Tim. I think the kino tubes I use working are 98 CRI. Film needs a very high CRI like that. If you're shooting a digital format or video you can get away with somewhat lower CRI.

 

[Tim Gray]

Yeah, Tim. I think the kino tubes I use working are 98 CRI. Film needs a very high CRI like that. If you're shooting a digital format or video you can get away with somewhat lower CRI.

Probably true. I always chalked it up to the fact that if you are paying your actors millions of dollars, what's a couple extra thousand in light bulbs?

That's why I always get a chuckle when I hear people talking about digital in Hollywood films. Not that its not making inroads, but seriously, the cost of film in a Hollywood film is a only a small portion of the budget. It's all about efficiency...

 

[PanaDP]

Very true. There is a lot of digital production now but, for the most part, anything over 4 or 5 million dollars tends strongly to be shot on film. Kodak and Fuji still very strongly support film for motion picture production and there are some pretty amazing stocks available. I wish I could get some of those emulsions in sheet film.