Theatre Video Projection & Milky Blacks

I recently got pulled into an indie films post for some advice on the final grading. It was shot on the PD-150 and contains some very low-key available night scenes. The first thing I noticed was that the blacks were severely milky. I brought them down and adjusted the gamma a bit to make the image snappier and he told me he didnâ€™t want to crush the blacks. I showed him on the waveform monitor that absolutely no shadow detail was being lost. He told me that he had screened the film in town on a digital projector with the DP and they were shocked to see the loss of detail in the blacks and as a result dropped the contrast. Now, I know for a fact that the theatre he screened it at takes less than perfect care of their equipment and probably has never calibrated their video projector properly.

Since I have very little knowledge of the dynamic range of some of the video projectors being used in art house theatres right now, I was wondering if some people out there who have screened DV material could give me some advice. Is it common practice to raise the blacks to compensate for theatrical screenings?

His film is very well shot and Id hate to see him sacrifice the contrast of his image because it didnâ€™t hold up in one theatre. Im trying to convince him to run some tests in other local theatres as well. If there is anyone in the Portland, Oregon area who has experience with this sort of thing and is interested in helping, Id appreciate it immensely.

Nick Peterson
Portland, OR
www.sampofilms.com

>I recently got pulled into an indie films post for some advice on the final >grading. It was shot on the PD-150 and contains some very low-key >available night scenes.

REPLY :
It could be that some of the places you are viewing your material are calibrated to a bottom black level of 7.5 IRE units, while others are calibrated to a bottom black level of Zero IRE units.

A lot of digital environments the color bars that everything is aligned to are set to O IRE black, while analog is still 7.5 IRE black. Yet some digital houses use 7.5 IRE! I have seen confusion over this in the most sophisticated studios and trucks, with all the proper test gear, scopes, buzzers and bells.

If this O IRE/7.5 IRE standards confusion is happening in a high end broadcast environment, I'd bet money its happening in a DV movie projector environment.

Lew Comenetz
Video Engineer, Seattle

>The first thing I noticed was that the blacks were severely milky... >skipping some details]... I was wondering if some people out there who >have screened DV material could give me some advice. Is it common >practice to raise the blacks to compensate for theatrical screenings?

One place to look for the problem may be in the playback machine itself. Digital editing is done at zero ire and output back to a DV tape at zero. The tape deck itself can have a menu function "add set up". When this is selected the deck plays back the zero blacks with 7.5 ire added in.

If the edited material already has been given 7.5 ire setup or there are some 7.5 analogue scenes in it, the play back deck will again add another 7.5 ire's to the black giving you just the look you describe...15 ire blacks.

Al Emer
Lighting Cameraman
Holmdel, NJ

My own guess is that people are just running the projectors too damn bright, say at the setting they'd use in a convention auditorium or ballroom with lights up slightly, and not the setting one would use for a fully darkened theatre. The projections certainly seem very bright compared to film projection, but I don't have a footlambert meter.

Steven Bradford
Film HD program Chair
Collins College
Phoenix AZ

> He told me that he had screened the film in town on a digital projector >with the DP and they were shocked to see the loss of detail in the >blacks and as a result dropped the contrast.

One of the biggest flaws with digital projection compared to film is that you no longer have control of your black levels! In addition to the 0%/7.5% setup issue, projectors that are used for presentation displays (i.e. PowerPoint) often have blacks crushed even beyond the setup difference, to add "snap" to the image.

I've been burned by this before, and now I make sure (whenever I can) that the projector is properly calibrated to *my* black levels prior to exhibition. At the very least I put bars with PLUGE on the head of the tape, and add a greyscale with end-markers too, if I have a chance.

Sometimes it helps.

Adam Wilt

/ video geek / menlo park ca usa

Adam Wilt wrote :

>One of the biggest flaws with digital projection compared to film is that >you no longer have control of your black levels!

I wanted to thank everyone for their replies to this issue and especially Lew for some assistance off list. I think everyone helped narrow down the issue, which was two fold:

1) Since his levels are indeed correct (now that we've brought them down once again and had them checked), we can assume the projection he saw was setup incorrectly (probably at 7.5). I'm stressing that he has properly referenced color bars at the head of his festival tapes and to follow this in the future.

2) Some night exterior scenes were apparently shot with gain activated on the camera; these scenes are especially milky and needed some additional adjustment.

Thanks everyone!

Nick Peterson
Portland, OR
www.sampofilms.com

Adam Wilt wrote:

> One of the biggest flaws with digital projection compared to film is that >you no longer have control of your black levels!

Would it be rocket science for projector mfgrs to incorporate an auto-detect system to determine the black level of an input feed and calibrate itself accordingly? Seems like all it would take is one embedded bit of metadata code.

Failing that, it could constantly search for the blackest black, and calibrate to that.

Yes? No? Why not?

Dan "recent victim of milky blacks at a festival screening. Grrrr.. " Drasin
Producer/DP
Marin County, CA

>Would it be rocket science for projector mfgrs to incorporate an auto->detect system to determine the black level of an input feed and calibrate >itself accordingly?

----Either that or "look up tables", presets to SMPTE, ITU, etc....something akin to the scene file setups in better cameras. I have a sneaking suspicion that the very high end projectors probably DO have something akin to this but I like your idea of metadata because it could conceivably be an "add-in" or "add-on" to many different models. I sure hope manufacturers look in from time to time around here for good suggestions.

Jeffery Haas
freelance editor, camera operator
Dallas

>Would it be rocket science for projector mfgrs to incorporate an auto->detect system to determine the black level of an input feed and >calibrate itself accordingly?

It would be a Frankenstein's monster, perhaps.

Let's assume there's a tape supplied with embedded calibration metadata (i.e., bars). Fine, but what about the next tape without such calibration?

Auto-leveling on blacks is the whole auto-exposure problem, turned upside down. Do you want the blacks to pump? With what time constant? Do you clamp to picture? To bars? The bottom PLUGE patch? Blanking? Breezeway (assuming composite). Some offset from sync tip? With or without setup?

Without a standardized test pattern, this can't be done. The best a projectionist can do, presented with a bars-less tape or disk file or bitstream or whatever, is use judgment. I don't see the need for an aware projectionist going away any time soon.

And even with bars, there's variation: does the bar signal have the full PLUGE? Most RGB-rendering NLEs can't create the sub-black PLUGE bar.

Is your circuit a closed-loop system, measuring actual light output? Or does it simply set a black level, and assume no offsets or biases in the light-emitting or light-gating system that need to be compensated for?

Harry Mathias told of a shoot spent breathing theatre fog to get nice atmospherics, lifting the shadows to 20 IRE. As soon as it got to the edit bay and the engineer saw it on a 'scope (this being back in the linear days), he promptly pulled the blacks back down to 7.5 IRE.

Bingo! Crisp blacks! DP's intent destroyed!

And that's with a PERSON. How do you tell a circuit not to fiddle stuff?

Adam Wilt

/ video geek / menlo park ca usa

Dan "recent victim of milky blacks at a festival
screening. Grrrr.. " Drasin wrote:

>Would it be rocket science for projector mfgrs to incorporate an auto->detect system to determine the black level of an input feed and >calibrate itself accordingly?

You're not far off Dan. The Panasonic PT-L500U has an "AI" mode that will auto adjust white/black levels based on content (rather than feed). Fortunately, I've calibrated my projector fairly well and actually don't need to use the "AI" mode (uses up the lamp faster), if you combine a good LCD/DLP projector with the right screen you can get very black blacks.

With today's technology there's really no excuse to have milky blacks anymore.

Illya Friedman
Home Theater Fan
Van Nuys, CA

Adam Wilt writes:

>Let's assume there's a tape supplied with embedded calibration >metadata (i.e., bars). Fine, but what about the next tape without such >calibration?

Two answers :

1) I was suggesting that such a system would auto-calibrate itself for each tape.

2) The kind of metadata I had in mind would simply indicate whether a given tape's setup was at zero or 7.5 IRE. That's not very sophisticated, but it would eliminate one major headache when tapes of mixed setup are played. This metadata would not be in the form of bars or a chart, but would be in the form of numbers embedded in the video stream.

>The best a projectionist can do, presented with a bars-less tape or disk >file or bitstream or whatever, is use judgment. I don't see the need for >an aware projectionist going away any time soon.

No doubt you're right.

>the engineer saw it on a 'scope .. he promptly pulled the blacks back >down to 7.5 IRE.

Presumably with metadata that tells the circuit (directly) or the engineer (via a display) what the desired result or calibration should be.

Or are my thoughts so utterly simplistic that I should be thoroughly embarrassed for having expressed them but don't know any better?

Dan Drasin
Producer/DP
Marin County, CA

>1) I was suggesting that such a system would auto-calibrate itself for >each tape.

Here, an idea for Sony: since the use of built-in memory-chips on tape has never really kicked off, it could be used to store such metadata.

>the engineer saw it on a 'scope .. he promptly pulled the blacks back >down to 7.5 IRE.

I always write all the set-up details on the tape and the tape-sleeve, so that any monkey at the duplication-house or broadcast-line, can understand what we meant it to be. Fortunately in PAL-land we don't have to deal with the 0/7.5IRE mess (except when converting and sending a tape to an NTSC-country, but I have never had to do that, but certainly will include that IRE-setting then as well).

Cheers

Martin Heffels
--
filmmaker/dp/editor/filmschool techie
Sydney, Australia