Lowest Safe Current 2k

[Steve Kraus]

What is the lowest current you'd run a 2K bulb at for long term health and stability? It's an Ushio DXL-20SN3 which is a 2K digital bulb. Specs say 50-85 A but the projector (NEC NC1200C) won't allow me to go below about 60 A. The iris thing is in place.

I used to use 2K film bulbs but went digital this time. As a screening room the bulbs last for years. Each round I consider going with a digital bulb and dropping to a 1600 W (pulling the iris thing) but no one can guarantee I'd still make brightness and continue to do so over bulb life. We are common height so Scope's stupid cropped format takes a LOT more light than 1.85. I have backed off Z mechanical adjustment to get level. Wouldn't mind dropping current a bit if I could.

Next question: I always set level with the white found in the test charts on DCC. Is that right? The reason I ask is this. The other day I made a white field in Photoshop. RGB values FFFFFF. Made that into a test DCP. (Later noticed I already had the same thing among my test DCPs.) But on testing it's not as bright as the white I get via DCC. Seems to me that FFFFFF is as bright as it can get in a real movie which is what should be 14 fL. Not a huge difference but def not the same. What do you use for setting level?

[Leo Enticknap]

My guess is that the lower limit that the software allows you to set is the lowest that you can run the lamp at without the risk of an unstable flame and flicker. In situations where we have not been able to go down far enough to be within DCI (a common problem in many residence theaters and some post house screening rooms), we would install a neutral density filter of a rating that puts us in the sweet spot with the lamp at roughly 50% power (if the screen is common width, as close as possible to that in flat and scope).

As for test patterns, an uncorrected one (as in, produced by the formatters without the MCGD color calibration file applied) will be a bit brighter, and obviously a different color, than corrected. I can't remember if the NEC gives you both options. Some studio tech checkers like to do the colors by projecting DCP test patterns of white plus the three primaries, writing the numbers down, then entering them into the color calibration wizard while ignoring the uncorrected patterns that the wizard would have you project to take the measurements; and after that is done, project the white one again to verify that you have .314 and .351 (or within the .002 or .006 permitted variance, depending on if it's a screening room or theater). There is no reason why you couldn't calibrate your CLO/LightLoc/whatever NEC calls it the same way, if the difference in light level is significant enough for that to be worth doing.

[Marco Giustini]

As Leo said, use a peak white DCP instead, it's a known good pattern and it's the one you want to use as reference. Some projectors have special patterns designed to give you actual 14fL on screen but if you go via your server you can't be wrong. Same with everything else.

When converting from Photoshop, you need to understand that you are working with an 8 bit workflow and DCinema is 12bit. So there WILL be a conversion between your FFFFFF and the 12 bit one. I looked into making an accurate white patch many years ago, 12bit software were not available then (I don't know now) and even a patch made with a script would still translate ever so slightly incorrectly when converted by DCP-o-Matic, I remember chatting with Carl back then. The FS_White, FS_Red, FS_Green and FS_Blue DCP patterns are small and should be your go-to patterns when checking brightness.

On light, if you constantly run a bulb at its minimum, it might develop flickering at some point, even though the specs allow that. Sometimes you can get rid of that by running it at nominal current for a while. The Iris might give you a little extra contrast (it depends on where it is, I cannot remember on the NEC) so I wouldn't discard it straight away.

Finally, sometimes backing on the Z axis might affect uniformity. NEC were particularly affected before they changed a diffuser filter where the light sensor is located. Just a little heads up

[Marcel Birgelen]

There is a difference in definition between "FFFFFF" and the whitest white in DCI-P3, because the white point of both formats doesn't match. The bit-depth doesn't really matter here: More bits give you more steps between the extremes, but it doesn't necessarily move the extremes... Higher depth color formats often use more bit depth, because the range they need to cover is larger and those increased steps are needed to avoid banding, but you have those oddball HDR formats like HLG that use 8 bits but logarithmic mapping...

Back to the issue at hand:

If you convert FFFFFF from sRGB to DCI-P3 xyz, you get x≈0.328, y≈0.344, which is the D65 white point.
DCI-P3's white point is x=0.314, y=0.351, pretty close but NOT identical.

Color science isn't for the faint of heart.

[Steve Kraus]

Using an ND filter is something that I've thought about but I'd need one with no effect on color because I'd have to put it up for 1.85 and down for Scope. Or can color balance be set separately for each format?

ND in the path for 1.85 would let me reduce the amount of backing off on the Z adjust I had to do - back to peak or at least closer depending how dark the ND is - which then helps me with Scope. Common height screen so I'm at about 71% on 1.85 but with Z backed off I'm into to 90s for Scope.

[Marco Giustini]

ND filters should be Neutral as the name suggests. But if they're not, you can set two different MCGD, one for Flat and one for Scope and have them appropriately selected in the formats.

[Marcel Birgelen]

A new, out of box ND filter should be practically 100% neutral, like Marco indicated.

One gotcha though: ND filters do degrade over time when exposed to UV light and this can have an effect on their absolute performance, but also on their neutrality. We're using ND filters in our xenon machine in our screening room and we do replace them on an interval, only I'm not able to find the interval right now.

[Marco Giustini]

I'd also imagine the quality of an ND filter can wildly vary with their price tag. An ND filter from Ebay might not be so neutral and/or not last long when put in front of a cinema projector! I'm wondering whether a larger ND filter, installed in a way that the frame is as large as possible might be a good idea - to spread the heat and the UVs over a larger surface?

[Marcel Birgelen]

Yeah, the ones you'll find on AliExpress probably don't even last you a single show.

As long as the ND filter is a real ND filter and not just something pretending to be one, it doesn't really matter where in the light path you install it. You could even have an ND-coated porthole if you really wanted. The only thing you want to avoid is reflections from the ND filter back into the lens.

[Steve Kraus]

Anyone ever used one? Ideally it would be glass, if possible screwed to the lens. As I see it, it would only be used for Flat. Eating up light on flat enables me peak the Z which enables me to bring down the current on Scope from the 90+ percent it is now.

[Marcel Birgelen]

Anyone ever used one? Ideally it would be glass, if possible screwed to the lens. As I see it, it would only be used for Flat. Eating up light on flat enables me peak the Z which enables me to bring down the current on Scope from the 90+ percent it is now.

Yes, like indicated: We use an ND filter on our xenon machine, which is actually used for color grading. The high contrast lenses already eat a lot of light, but not enough, so we're using ND filters to correct for that. The machine gets calibrated multiple times a year, so that compensates for the degrading ND filter.

Edit: The filter is glass, suitable for high-intensity usage but not mounted to the lens as not to interfere with it, cause vibrations, etc, but has its own holder in front of the lens. It's also slightly angled, as to avoid reflections into the lens.