|  IP: Logged

>>> Phil

Mr. 3D!

When we try to perfectly match two projectors at Kodak (for critical side-by-side comparisons of new films), we start with reflectors from the same batch, and xenon lamps of the same batch and age (hours). Even lenses of the same brand may have a slight color difference, so we try to simply switch the lenses between the two projectors to get the best color match. The final "tweek" is to use a SINGLE gelatin filter (usually a Wratten color correcting filter) in a filter holder in front of each lens. Note, using a filter to get "perfect" color match will only work on a smaller screen, as a really large lamp will fade or even warp the filter. Needless to say, the filter should be free of scratches or dirt, and monitored for any fading.

Standard SMPTE 196M specifies the color temperature for theatre projection. For theatres, the light reflected from the screen should be 5400 Kelvin, with a tolerance of -200 and +600. In other words, between 5200K and 6000K. Natural daylight has a color temperature of about 5500K.

------------------
John P. Pytlak, Senior Technical Specialist
Worldwide Technical Services, Entertainment Imaging
Research Labs, Building 69, Room 7525A
Rochester, New York, 14650-1922 USA
Tel: 585-477-5325 Cell: 716-781-4036 Fax: 585-722-7243
E-Mail: john.pytlak@kodak.com
Web site: http://www.kodak.com/go/motion

I'll have to disagree on this. The different color IS apparent at first, but within a few minutes the brain compensates and the cones balance. I work with full color anaglyphs all the time, which have about as strong a variation in color as is possible. Since my eyes have become trained, normalizing occurs within five minutes or less. After a session, for the first few minutes after I remove the red/blue glasses, one eye sees objects with a pinkish tint, the other blue.

Color temperature tweaking can almost be called a subjective science, since about 9% of the population has color deficiency and a large portion of the population has other vision defects, such as night blindness, which can affect color perception at lower light levels.

Jerry (member of NSA, and certified 3-D nut since age 6.)

While color matching for 3-D is a nicety, and provides a good initial impression, many other factors are more important.

Lets clarify "color temperature" and "color matching", where color matching is getting two (or more) projection sources to have the same color temperature on screen within a desired range. SMPTE 196M that John mentions specifies a color match in terms of "chromaticity", the x- & y-coordinates of the color spectrum. I don't have 196M handy to reference, but I don't remember a +/- range in Kelvin for matching in the latest document. Not all that long ago, I think it was +/- 200K, but I wouldn't bet the house on it. (John, do you remember or have a copy handy?)

There are a lot of factors involved in getting the desired color temperature on screen, from reflector (to heat filter) to lamp to lens to port glass and yes, even the screen itself. It's amazing the people that do color temperature measurements without considering ambient light from things like exit signs. A few years ago I had a guy that wanted to trade out all his reflectors -- until we determined he was doing his measurements with the auditorium cleaning lights on. Oi.

In my experience, the xenon lamp is fairly constant in color temperature, the main change (if any) due to age rather than variations between manufacturers, operating amperage, etc. The reflector does play perhaps the biggest part, since it's generally dichroic coated, but they are pretty consistent within batches. Some big differences come from port glas and the screen, something that's very oftem overlooked in items that affect color temperature.

Pat

D5500 has x=0.332 and y=0.347

Not that I'm running out to get a 3-color temperature meter, anyway. I checked the price (admittedly, this was 10 years ago) for a thing made by HP, and it was about $20,000.

There were two things that surprised me somewhat when I got in this business mumblemumble years ago. One was how much footage that you would think would be MOS (since the shot contained no dialog) in fact was actually shot sync sound. And secondly, that dailies were timed, not one light. When you consider that the dailies will become the workprint and get shown to the big shots and to test audiences this makes good sense.

 |  IP: Logged

I worked out the MIRED shift values using the tables listed in the ASC Cinematographers Handbook and ended up using a 1/4 CTO gel on the bluer machine. I just taped it to the back of the shutter housing instead of projecting through it. After that I eyeballed the difference between the machines during setup each day and ended up changing the gel about once a week or so--we were cranking about six hours of dailies a day on that production, plus other special-request stuff on top of that. Lots-o-hours for four months of principal photography!

Last time I checked, that Minolta CS-110 listed for around $5800 US. As is, it reads out in x,y chromaticity coordinates and foot Lamberts, and comes with a chart for converting x,y to correlated color temp. For another $1700, Minolta supplies a little processor that connects to the meter and provides the correlated color temp directly plus printer capability. Kinda pricey!

Paul
SMPTE Hollywood Section
Unemployed mercenary film/video projectionist/engineer
"Otaku wa tsurai yo" <-- Yeah I know, bad pun. Sue me.
It's tough being a fan!
-28

Likewise a digital camera could do it, with auto white balance disabled and set to expose the white screen as a middle gray. Then just read the RGB value of a pixel in Photoshop. Once again, calibrating is the problem.

Note, an 18% gray card illuminated by sunlight on a clear blue sky midday will have a luminance of about 1400 footlamberts. Your screen should have 16 footlamberts. So you will need to adjust camera exposure by about 6-1/2 camera stops. But the color temperature should be nearly identical.