Color Temperature

I need to figure out what is going on here.

 

I have a constant lighting source at 5500K.

 

I am trying to reduce the intensity by using an iris on the light

source. I thought the color temperature would stay constant as I

decreased intensity that way.

 

What I find is that as I decrease the size of the opening, the color

temperature falls in proportion to the light intensity. I thought

that the color temperature would remain constant. After all, if

color temperature fell with intensity (as in changing the f stop on

your camera), color photography would be impossible.

 

Is this some artifact?

 

Any suggestions out there?

 

Thanks.

 

Ron Mowrey

Ron; please explain some more. <BR><BR>You have an iris on a tungsten or Halogen source? Is the iris neutral in tone? or a goofy 60's tie die hippy teeshirt? :)<BR><BR>How are you measuring the color temp? with a meter; film; digital? <BR><BR>Ok I and just shooting in dark here!.<BR><BR>Here I still use 3200K Halogens alot;

Kelly;

 

It is a halogen lamp at 5500 degrees measured with a Gossen Sixticolor meter. Voltage regulated, etc...

 

The iris is standard blackened metal. I have a series of 3.

 

As I change iris, effectively going from an open spot down by 3 'stops' if you will, the Gossen meter measures decreasing temperature in steps / stop and finally ends at about 4000K.

 

I don't have any reason to add blue filtration to bring up the temp, as I have no real reason to believe that the temp is dropping in the first place.

 

Ron Mowrey

I forgot to add..

 

If I move the meter closer to the source, the temp goes up again.

 

I tested that after the first post here, and that bothers me. The Gossen meter checks out very nicely and has been operating well in the past. But something isn't right here.

 

Ron Mowrey

I suspect this has to do with the meter's sensitivity. Have you shot some images to see if

there is an actual change?

This may be stupid but that doesn't stop me-) Could the iris be heating up? Becoming a radiator? Would that skew things?

Ellis;

 

I agree with you.

 

Since that might be the problem, I don't trust the original reading of 5500K, so I have not shot anything yet. If it is the meter, who is to say that the first reading is correct?

 

I'm going to run some additional tests using filters over the meter to see if that is the problem. I've been toying with several ideas while the shot is still sitting there waiting for me.

 

Ron Mowrey

There is obviously no "shift" occuring that would redden the scene,so it has to be the meter.As you stated,an opening alters all waves evenly.Can you borrow or rent a 2nd meter?(I suggest the Minolta color meter).Best way to experiment with color temp and meter issues,is by using flash as a light source.Check to see the temp of your flash first,you can then add filters etc to raise or lower temp,with a continous temp light source.

Ellis and Steve confirm my suspicions.

 

The iris is not heating up. That I checked. Sorry I didn't include that in the posts above. The lamp housing has a blower on it that keeps it pretty cool.

 

Thanks for the replies.

 

Ron Mowrey

A phenomena of diffraction?

The Gossen Sixticolor is a wonky thing.

 

If the light intensity drops too much, it seems to become less reliable.

 

I use it mostly outdoors to determine whether I need, say, an 81A, B, or C, and less so indoors.

 

The light also needs to strike the diffuser evenly, because there are two cells in there--one measuring blue and one measuring red. If you direct the diffuser toward an open sky, for instance, and cover half the diffusion disk with your hand, the needle will skew to one side, so be sure the meter isn't so close to the light source that the diffusion disk isn't evenly illuminated.

 

It's also possible that you have some other light source in the room (tungsten or daylight, for instance), and when you reduce the intensity of the halogen with the iris, you are reading more of the other light source.

RON; My Gossen Sixticolor reads Halogens as 3.2K; open blue sky only K; household 60 to 75 W bulbs as about 2.8K; it reads GE chroma-50 fluorescent bulbs as 5.0K; direct sun about 6K or less; with about 3k at sundown. The meter I think has two cells; one with a one with a red filter; one with a blue filter; so the ratio to the red to blue give the color temp; for a black body source. There might be linearity problems at low light levels; mine is well behaved in daylight brightness levels; and it more squirelly at lower levels. This of course is a self powered meter; so at some low level it has to stop working; ? The incident surface accepts a broad angle of light; so backing away from a spot light allows other lights to compete with the light on the meter. My meter here reads 2.8K when 1/2 foot from a 75watt household lamp; or 6 ft away. BUT sometimes it is squirelly; and gives an off reading. Im not sure if this an old age of the meter problem; meter needle imbalance problem; etc. It seems to be more sluggish at lower levels too. Maybe the movement is gunked up with crud; and there is less DC current to move the meter movement at low light levels; and or low Kelvin reading? ie low ebd of the scale...<BR><BR>The redsensitive cell is one the left of the white round window; below the "S"; ie below the <b>"S</b>ixticolor. The BLUE sensitive cell is under the "r"; ie "Sixticolo<b>r</b>.<BR><BR> One can block either the right or left side of the white incicent window; and get ANY reading from 2.6K to 20K!

Does the temp. vary as you DECREASE the meter to source distance?

Disregard, i reread the posts.

Are you metering in a 'dark' room -- with only your 5500K lights on?

 

 

 

 

My guess is the room lighting is getting mixed into the equation and the closer you move the lights, the meter is getting a more 'intense' blast of light. Moving the lights away will spread the light out....and your light meter will measure some existing room light in with your 5500K lights.

Metering is done in a dark room.

 

I went ahead and did some B&W exposures to test it out.

 

I'm pretty sure that the meter is 'going wonky' as said above.

 

Thanks agin.

 

Ron Mowrey

Ron; my meter is serial 3A1669. When it is 0 to 1 ft away from a dual set of F15T8C50 GE 15" 15W C50 Chroma fluorescent tubes; it reads 5.0K. <BR>It reads at 2 ft 4.5K<BR> It reads at 3 ft 4.0K<BR> It reads 3.7/3.8K at 6 ft.<BR> It reads 3.5K at 12 ft.<BR> It reads 3.0K at 15ft.<BR><BR>It reads 2.8K from 1 to 8 feet away from a 75W Edison Household bulb. <BR><BR>It appears the blue cell versus red cell response varies with light level; and the "blue" cell poops out before the red cell on mine. On mine the meter has the "Kelvin Droop" at high bluer Kelvins; than the lower redder Kelvins; as illumination is dropped. Mine works well with bright lights; and has this linearity/droop problem; in lower illumination levels. I had one of these in then 1960's; then got another awhile back used.<BR><BR>Once in the 1960's we placed it on a color TV screen; and had to have the repair guy bring a degauser; to fix the blob on the screen; due to the magnet in the meter movement. Their super High buck TV had no degauser.

<I>Since that might be the problem, I don't trust the original reading of 5500K,

so I have not shot anything yet. If it is the meter, who is to say that the first

reading is correct? </I><P>

 

What you would look for is just a color shift as you use the irises, not whether

the first (no iris reading is right. That is a different matter. The only color

meters I have found to be worth a damn are the Minolta Color Meter Ii and

even better, the Minolta Color Meter IIIF.

 

I suspect that a test with B&W will tell you very littleabouta color shift or only

show a subtle set of differences (blues darker & reds/yellows lighter) and that

would be if the test target is something that shows a range of known color

values like a GretagMacbeth Color Checker.

The B&W exposures were to check out the linearity of the aperture based exposure shifts, not color. They were correct, so the aperture is working properly and the exposures turned out correct in B&W. Now it is time to commit some color to the project.

 

My serial number is 10788. Very old instrument. And, I never noticed the dropoff before because I always used it in relatively high light conditions.

 

I guess I have to get another meter.

 

Using two sensors makes me feel that the system is subject to seriously false readings under some conditions. Ah well, I guess nothing is perfect.

 

Thanks again guys.

 

Ron Mowrey

Ron,

 

I don't know if I'm adding much here, but I also do not think anything about the iris should change color temp, unless possibly it affected temp inside of the lamp housing.

 

I have a couple of ideas about testing your meter, though. David Goldfarb indicated that there are two (color sensitive) cells behind the diffuser, and it is possible to direct the light more to one of the cells. I don't know how your light source optics are working, but possibly the meter illumination becomes uneven on stopping down. To test if this is a factor, you might take one reading, then invert the meter (using some sort of holding jig) and see if the reading stays the same. If no change, it seems likely (as others have suggested), that the meter goes into error with low light levels.

 

I'm not sure how one would get 5500 K out of a tungsten filament. I expect you know this, but just in case, tungsten melts ~3400 deg C, I believe. So clearly there needs to be some color filtration to mimic 5500 K light. Possibly there is some sort of mismatch between the Gossen sensors and the light source filtration such that the light seen be one of the cells is drastically reduced compared to the other.

 

It might be interesting to see what the meter reads from the bare tungsten source, including at various distances.

 

PS: for the casual reader, Ron said it's a halogen source; this normally means there is a tungsten filament in a small bulb filled with a halogen gas. For various reasons, this allows one to operate the filament at a slightly higher temperature than with a conventional tungsten bulb.

Bill;

 

You are right. Due to the aperture, I am able to use a 'gel' to correct the color temperature to 5500K. That is one reason that the light level is so low to start with.

 

I did try rotating the meter, and it made no difference. The light is uniform in density and temperature over the field it covers.

 

I checked 4 light meters also, two Seconic spot meters, a Gossen, and a GE meter (very old). I got 4 different readings from them in daylight and under the low light conditions. The daylight readings were within about 1/2 stop but in all directions, and the low level readings were only within about 1 - 2 stops. This indicates that all meters fall off in capability with light level. Something I should have remembered from my early days in photography.

 

Taking the average of the low light readings, I was still within 1 stop of the actual value. The average was f8 at 0.2 seconds for a 200 speed film not compensating for reciprocity. I ended up with a value between f8 and f11 for the 200 speed film and 0.5 seconds compensating for reciprocity. This was my test last night with the B&W film.

 

If all goes well, I'll be using two exposures tonight for 160 speed color, bracketing those above. Maybe 4 exposures total to make sure I hit it, not sure yet, I'm still thinking it over.

 

Ron Mowrey

The main problem is that the self powered Gossen Sixticolor meter was used at too low a light level. Other more expensive color temperature meters drop in accuracy; at much lower light levels. The Gossen Sixticolor meter here is well behaved at high light levels; where it normally is designed to be used. <BR><BR>With alot of studio lighting; mine reads type B Halogens as 3.2K; old Photofloods as 3.4K (Type A); the blue sky as 10K; direct sun today was 6K; 75w bulb as 2.8K; fluorescent GE chroma 50's as 5.0K. <BR><BR>If you use the meter with decent amounts of light; mine always is accurate. With dim lighting; it has a droop in Kelvins. <BR><BR>Instead of an iris; try a ND filter over the flat incident face. <BR><BR>Bill C; Re <i>It might be interesting to see what the meter reads from the bare tungsten source, including at various distances. </i><BR><BR>As mentioned above; tungsten response of the Gossen Sixticolor meter here doesnt change on mine from <I> "It reads 2.8K from 1 to 8 feet away from a 75W Edison Household bulb."</i> The meter of mine works well with lower levels of tungsten light; but droops in response quicker as a High Daylight Kelvin source is moved away from the meter. The other data set I posted was for a pseudo "Daylight source"; it shows howe the meter droops when moved away from a "daylight source".<BR><BR>I first used one of these meters in the 1960's; and have lost the specs versus light level sheet that came with the meter. When using the meter below the rated minimum light levels; one gets worse results. My current Gossen Sixticolor meter has a lower footcandle minimum for low tungsten Kelvins; than high daylight tungstens. I dont consider it a bad meter; when it reads low; when used below the light level that I know it droops. <BR><BR>At some point of much lower illumination; the Minolta's probably will act goofy too.<BR><BR>

Well, the negatives are drying, but it looks like I goofed. The reciprocity failure of the color neg was less than the B&W test shots from last night so they look kinda dark.

 

I won't know about the balance until I can either measure the grey scales included in the scene, or print the negatives. I will probably wait until I get a better exposure.

 

Thanks.

 

Ron Mowrey