Is it possible to shoot in 3-D using modern anamorphic lenses?

[Reuel Gomez]

Recently, a lot of cinema lens manufacturers such as Zeiss and Cooke have been unveiling anamorphic specifically for the purpose of being used on digital cinema cameras such as the Arri Alexa and Red Epic which both have sensors that can accommodate a 2x squeeze anamorphic lens. So I wonder, could lenses such as these, and even other recent anamorphic lenses such as the Panavision G-Seres and Hawk V-Lite 1.3x squeeze lenses be matched and used in 3-D rigs for stereo capture? I mean, they clearly don't have the burden of older anamorphic lenses such as of course the Panavision C-Series and Kowa Cine Prominar lenses of older coatings and just generally being unfit for that kind of work due to mismatch between lenses in color (at least in the PV C-Series) and a whole heck of a lot of aberrations, so is there something else I'm missing here as to why they aren't being used in 3-D productions?

[David Mullen ASC]

You could use rear anamorphic zooms like the new Angenieux on 3D rigs but what's the point? The newer front element anamorphic primes still have some distortions that make focus-pulling on 3D rigs difficult, except maybe the 1.3X Hawks, but then you're back to the same question -- what's the point? No anamorphic distortions and flares makes it look like you shot on spherical lenses, and then you are back to the question as to why you aren't using spherical lenses, especially for a 3D rig.

 

The only advantage then becomes the fact that with 4-perf 35mm you are using a bigger negative for a 2.40 image, but with 3D you'd probably be shooting digital. And since most 3D films are finished in 2K and projected spherical cropped to 2.40 in DCP's, I don't see much advantage in being able to use the full 4x3 sensor height on an Alexa recording raw, not if you aren't going for anamorphic lens distortions for a look. And if you really wanted more resolution, then you should just shoot on the sharpest spherical primes on something like the Sony F65 or a Red Dragon. So again, what's the point of using anamorphic lenses on a 3D rig?

[Reuel Gomez]

You could use rear anamorphic zooms like the new Angenieux on 3D rigs but what's the point? The newer front element anamorphic primes still have some distortions that make focus-pulling on 3D rigs difficult, except maybe the 1.3X Hawks, but then you're back to the same question -- what's the point? No anamorphic distortions and flares makes it look like you shot on spherical lenses, and then you are back to the question as to why you aren't using spherical lenses, especially for a 3D rig.

 

The only advantage then becomes the fact that with 4-perf 35mm you are using a bigger negative for a 2.40 image, but with 3D you'd probably be shooting digital. And since most 3D films are finished in 2K and projected spherical cropped to 2.40 in DCP's, I don't see much advantage in being able to use the full 4x3 sensor height on an Alexa recording raw, not if you aren't going for anamorphic lens distortions for a look. And if you really wanted more resolution, then you should just shoot on the sharpest spherical primes on something like the Sony F65 or a Red Dragon. So again, what's the point of using anamorphic lenses on a 3D rig?

For the look. That's what I meant really.

[David Mullen ASC]

It's the anamorphic look -- stretched bokeh, causing focus breathing during racks, and horizontal flares -- that makes it hard to pull focus and match two lenses in 3D rigs. You'd have to shoot at a deep stop to reduce this look, and minimize any flares... but then you're back to why are you shooting in anamorphic? At f/8, most anamorphic lenses start to look like spherical lenses. With 3D 2-camera rigs, you generally don't want lenses that have "personality", what you want is neutrality.

 

Plus most people want to use small zooms in 3D rigs to save time on lens changes and calibrating.

 

3D could be done with anamorphic lenses, I just can't imagine many camera assistants or stereographers thinking it was a good idea, it would just make it harder to get good 3D. I mean, you're already shooting in 3D for an unusual optical experience, tossing in anamorphics into the mix is sort of a distraction.

[Sean Cunningham]

The look of the photography in those 3-Depix process films from the early '80s (Friday the 13th Part III) always made me think of anamorphic, because it generally had a shallow depth of field and very dreamy quality to it. My understanding is this process used a single lens for stereo acquisition. I wonder if it's possible that a mid or rear mounted anamorphic might have been used in some of these films or if it's the optics of the stereo adapter that gives the photography some of the feel of anamorphic in the defocused areas of the frame.

 

I'll have to go back and look now and see if the bokeh is stretched or not. It doesn't "feel" like shallow DOF spherical photography at all.

[Chris Husz]

I asked the same Thing to Mr. Mullen some time ago in the REDuser Forums.
But i forgot to mention the 1.3x hawk's back then. I think they do have a good Portion of the anamorphic look.
Not as strong as the normal hawks or Panavision or even older Glass like Ultrascopes and Lomo's.

But it's stere. The bokeh that is. And if you try hard enough even the flares (i was actually surprised how well the Hawks flared, having their anamorphic element in the middle).
I doubt anamorphic 3D will be used that much in the near future, due to the more complicated use and the resulting costs mr. mullen mentioned.

Im waiting to see how well the new ARRI/Zeiss anamorphics perform though.

[Sean Cunningham]

I've worked post on a few stereo projects, both native and conversion (T23D, Meet the Robinsons). What I found is the human brain really wants to make sense of what you're looking at. It does an amazing job of filtering and averaging out certain discrepancies from each eye, particularly if they're specular in nature. It has to do this for natural human vision after all.

 

Strong anamorphic streaks would create an interesting depth dilemma. Placing them at the depth of their source could look very cool or it could look very weird and ultimately will create problems elsewhere in the frame, eventually, where they will overlap objects at conflicting depths.

[David Mullen ASC]

Since lens flares happen internally in the lens, they would have to be the closest thing depth-wise in a 3D image. I didn't see the whole movie, but the 3D trailer for "Star Trek Into Darkness" (post-conversion to 3D) drove me nuts because every anamorphic lens flare sweeps right in front of your eyeball in 3D, like someone waving their hand in front of your face now and then. At least in a 3D conversion you could adjust that depth but it still doesn't make much sense if it isn't in the foreground -- the flare doesn't start physically at the light source, it starts at the first lens element that is flaring.

[Sean Cunningham]

They have the option of doing that when they fake it but this means that they will often be forced to un-couple the flare from its source in the eye they're having to fake, which could cause odd shimmering or other strangeness. If you were to actually photograph a flaring light source in deep depth with an anamorphic system it would most likely create what appeared to be a "V" shaped structure pointing away from the view (if there were obvious areas of frame representing closer depth at the same screen height as the flare). Because the structural center of a flare will be over its illumination source in a strict two-dimensional relationship, for both eye. That means, in stereo terms, it's at the depth of whatever is causing it.

Edited October 9, 2013 by Sean Cunningham