Non-Standard Aspect Ratios

[Martin McCaffery]

When directors and dps just on aspect ratios...they are not thinking of cinemas...

I'm sure there are enough bad screens in Hollywood that any director who wants to do something in other than flat or scope should be required to have a screening at a randomly selected theatre before he signs off on a cool new aspect ratio. Maybe he/she should also be forced to watch their previous feature under similar conditions, just to know what's going on.

[Steve Guttag]

It SEEMS like metadata could indicate the number of active pixels in each direction, and theater automation could adjust the lens, masking, and light level based on that metadata. ANY aspect ratio would work (though some pictures may be masked down a lot).

Harold, the problem is not the cueing. Setting aside metadata, the creation of a cue for an aspect ratio is a no-brainer for any competent installer. However, when this cue fires, what happens? If the lens does not have a corresponding setting, done by the installer, what does that cue do? If the masking (if available) is not a zillion-stop variety (most are in the 2-stop, 10-stop categories with some that can do other variants), how do you fit more than 10 aspect ratios into a 10-stop machine? 10 sounds like a lot but most of my sites that have them also tend to have film and other media so the stops get eaten up pretty fast. And then there are my sites that have 4-way masking and the screen presents the various ratios such that they all "feel" big and with similar square footage. Again, lots of stops in the masking machines.

But okay, you're a 2nd rate theatre without masking. But you do want to zoom the image such that one edge is hitting one of the screen's limits. If you zoom the lens, by definition, you've changed the lighting requirements because you are covering more or less screen area. Who calibrates that for the metadata issued new-fangled format?

And speaking of maskingless theatres...so you zoom the image, or you get an S-220 movie and use your normal Scope preset, what about the keystone on those pillar boxes? You need a separate screen file to ensure that they don't look like they are leaning. It's fine if one can prepare and also get the sense that they the effort will be used. But for crackpot ratios like 2.1, who is going to set up for a 1-off? And I say crackpot because 2.0 and 2.2 are already well established for over half a century...the goldilocks that wanted 2.1 really didn't like either established ratio? That sort of thing is a nuisance.

[Harold Hallikainen]

I was thinking masking could be continuous, not any specific presets. This could be done with stepper motors or synchronous motors with a timer and limit switches (I did a satellite dish steering system that way. Same with lens zoom and light level. Come up with a few points on the curve (perhaps standard aspect ratios) and interpolate for inbetween. I had not thought about keystone correction. I imagine that could also be interpolated between several fixed levels. It seems like all these controls are continuous and could just be calculated out from active pixels horizontally and vertically instead of having to come up with a bunch of presets that work great until the next odd AR comes out. Clearly this whole projection thing is outside my area of expertise.

[Steve Guttag]

Harold, while it is all doable, even if you got your stepper motor masking, it wouldn't be so perfect due to cord stretch and would constantly need recalibrating. Yes current ones do as well but they only worry about the few stops...not worrying about if a stepper loses its position. The 10-stop motors by MDI seem to be done via stepper. You'd need to convince all of the projector and lens manufacturers to get in on the plan too...with all of the varieties of lenses. Heck NEC can't repeat a lens position, as it is, let alone adding another degree of precision.

You know, if we're going to do metadata...let's start with the easy ones...have the movie select the right picture and sound formats rather than have a person map out that "F" means flat and "S" means scope. TMS systems tend to have this capability (but are not hip to the variants within a container, like F-200).

[Mike Blakesley]

However, be careful what you wish for. If 2:1 really takes off...you watch...exhibitors will put in 2:1 screens (they aren't masking, by and large, anyway) and do what some did in the film days...EVERYTHING will be shown 2:1.

That'd be OK with me. Our screen has been 2:1 since 1953. In the film days we used to crop scope movies off at the sides, which I hated. When we went to digital, we show it in the proper aspect ratio, but we move the picture to the top of the screen and leave a gap at the bottom...which I hate. However, it does make it easier for shorter people to see the picture -- it gives us an extra foot or so at the bottom.

An item on my list of projects is to create some kind of masking panels to fill in that space during the 85% of movies we show that are in scope. For flat films we'll just remove the panels. I've already got the panels cut, I just need to get hold of some masking cloth and figure out how to put them up and make them easy to remove.

[Steve Guttag]

Mike, you can do a half-scissor masking (and motorize it, if desired). You, basically rig it like a scissor masking (common center where top and bottom panels move together) except omit the upper panel and use counterweights to balance the system out.

This, normally, presumes that you have enough space below the screen for the rigid masking to go. If not, use soft goods and let it pile up with, perhaps a mini-skirt to hide the folding masking drape.

[Rick Cohen]

In some theaters it's not just as simple as creating lens position file, screen file, and macro. Some masking systems do not have enough presets. At least one major chain installs masking motor/control units that are actually garage door openers, and have only two positions (open and closed, or, in a movie theater, flat and scope).

The whole point of standardizing a couple of ratios is to enable theaters to be able to present them properly. OK, digital cinema makes it easy to create movies from anything between cellphone-style vertical videos, to as wide as 3:1 if you really want that. But they are not mandated in the DCI standard, and as a result, some theaters will not be able to mask them properly. The same applies to ratios that were a widespread industry standard once, but are not any more; 1.37 being the obvious example.

One of the sites I am servicing is a drive-in with five screens, so masking isn't an issue. Three of those screens were constructed at 2:1 (approx 60' x 30') so the F-200 looks perfect on those screens, with almost 100% of the screen area being used without any cropping. Not saying that I prefer F-200, just that it was worth the effort to create the lens files on those screens under the current circumstances.

If it was a one-off oddball format or a less prominent movie, I probably wouldn't notice or have made the effort to create a special lens file. Barbie is currently the number one grossing movie and breaking records at the box office for the third straight week. Oppenheimer is holding well at #4 in its third week. Both movies look better on the screen when projected in their intended aspect ratio. There have been several wide release F-200 titles recently, and Christopher Nolan seems to be fully committed to releasing all of his movies in 2.2:1. Instead of pretending the non-standard formats don't exist and using standard scope or flat lens files, which compromises the presentation quality, I went ahead and created the correct lens files for those two features. I'm sure these files will come in handy down the road, as I expect that F-200, F-220, and S-220 aren't going to go away, and might even become more frequently used with filmmakers than they are now.

[Steve Guttag]

Remember, the S-220 format is specifically designed to work with standard Scope presets (no lens or light setting differences). Likewise, F-220, on 1.85 screens, is designed to use the normal Flat lens and light settings. The only addon is to create a dedicated screen file, if you have enough keystone to cause the pillars to highlight the fact. On F-220 on 1.85 screens, odds are, most cinemas do not have enough lateral keystone to worry about. Where things go amuck is when only the F-220 version is offered and it is run on a 2.39 screen. To a lesser extent, the 2.39 screen with F-200 can be an issue:

Screen Shot 2023-08-05 at 8.03.34 PM.png

Since Oppenheimer offered teh S-220 format, the F-220 issue was eliminated.

[Jon Dent]

Our preshow and ads are in flat, then we switch aspect ratio if necessary for the feature. If the feature is in 2.0:1, we'll run the previews in flat, and then insert the 2:0:1 macro just before the feature. It's not that difficult and I do not understand why anyone would think this is such a big deal. If your managers don't understand how different aspect ratios work, then teach them.

Because we have a series 2 NEC that takes 45 seconds to switch presets. That's nearly an eternity for the audience to stare at a blank screen after seeing ads running for the previous half hour. Plus as it's been mentioned before, focus isn't the most solid with the presets, so changing the lens setting multiples times a day seems like a recipe for blur.

For our ads I ended up using a flat container and shrinking the image down a bit so it would still work with minimal cropping when the lens is set to scope. It's not ideal but it beats having to have two (or more) separate advertisement DCPs.

[Andrew Leist]

It's not just the macro for the Zoom that you'd have to build. You have to adjust the brightness as well. Flat and Scope always have different lamp power settings, and setting up a macro for 2.0:1 or 2.2:1 would be no different. The new macros would require new lens and lamp settings. I've considered it, but have so far not done it because of what it would add to our maintenance procedures.

[Jason Sharp]

Another non-standard wide release this week.
GranTurismo_FTR-2_F-190_EN-XX-CCAP_OV_51-HI-VI-DBOX_4K_SPE_20230724_DLX_SMPTE_OV

[Steve Guttag]

Yeah, there needs to be pushback on this sort of thing. 1.9 is nothing more than full-container. However, F-190 was a full container shrunk to fit within the Flat container. Alternately, some goof rounded up 1.85 to 1.9 the same way people will round Scope from 2.39 to 2.4.

[Marcel Birgelen]

I guess we're now entering into the realm of cinematographers and directors that have never ever shot anything on celluloid film before and all came from the iPhone-school of film-making. For them, choosing a proper, standard aspect ratio isn't a thing... aspect ratio is a choice and can be anything you want at any time of the day...

Before our last remodeling, our screening room had a "native" aspect ratio of about 2.0, with the masking fully open, we've now changed that to 2.39, after the last screen replacement. Even though 2.0 gave the biggest possible picture, I never really liked the looks of it. It's neither fish nor meat... It's like a director not daring to make a choice. I always preferred the looks of either 2.39, proper 2.2 or just 1.85 over any other oddbal aspect ratio. As for 1.3x, it always feels a bit cramped for me and I'm happy that cinema evolved to wider aspect ratios over time. Maybe it's because I grew up with those aspect ratios, just like I grew up with 24 FPS, but, in the end, a movie looks like a movie, because it's being shot like a movie... If you start to mess with the basic canvas of what a movie is, then you start to mess with the very concept of that thing in your head. Maybe the "vertical" or "letterbox" generation, as it's being called in another topic doesn't mind this, but I still do.

[Scott Norwood]

I'm still waiting for someone to make a movie where the entire image is warped into a circle. I guarantee that it will happen.

[Marcel Birgelen]

I'm still waiting for someone to make a movie where the entire image is warped into a circle. I guarantee that it will happen.

That used to be called Omnimax

Granted, it looked more like a squashed egg than a circle.