MORE THAN A GIMMICK – The Art of 3D Filmmaking

I’m in a unique position as a film blogger and critic when it comes to discussing the art of 3D. I worked in stereo conversion (the 3Difycation of 2D movies) for three years as a production coordinator. During that time I talked regularly and in depth (get it?), with some of the best 3D artists in the business. It was my job to know when and why a shot had problems, and help organize the team who could come up with solutions. As always in film production, there was a tight schedule to keep.

So while I’ve never worked as a stereo artist, and I only have an average “stereo eye,” the contents of this 101 may be the most informed topic I can write on concerning the basics of filmmaking and review. These aren’t my opinions or even my company’s “style.” (No trade secrets here.) They are the fundamentals of 3D that seemingly nobody in the general public is aware of.

Not that they should be. The magic of film is that it’s supposed to move you without your even noticing. Unfortunately, studios and theaters introduced the 3D Surcharge, and suddenly all that changed. We don’t pay extra for well-mixed surround sound, more impressive special effects, or big name actors. Yet suddenly audiences were asked to put a dollar amount on the value of an immersion factor deemed extraneous to the “core” experience. 3D was Optional. Not in the sense that you were allowed to also see a movie in 2D if you chose but in the sense of being upsold on a car. (You want cupholders? Optional! That’s another grand.)

3D Walk by Don Hertzfeldt
The Animation Show by Don Hertzfeldt

A large part of what made people view and even implement 3D as a “gimmick” was the nickel-and-diming practices that came along with it. Now reviewers (like myself), who happen to give a damn about the art and value of 3D have to make special notes about whether it’s worth your hard earned dollar. The problem? While most reviewers are well educated when it comes to story structure, mise-en-scène, and performance, 99% of them know fuck-all about what’s actually going wrong if they’ve had a bad 3D experience. Sure, they might know it was boring or nauseating, but most of them could only give a wild guess as to the all important WHY.

MISCONCEPTIONS ABOUT 3D

I’ll give Cinema Blend props for making an attempt. Their “To 3D or Not to 3D” features are maybe the only consistent examples I’ve seen of reviewers regularly and succinctly trying to point people in the right direction. Unfortunately, the metrics they use for their 1 to 5 grading scales show how woefully unprepared they are to be taking on the role of expert on the matter.

“BRIGHTNESS”

We’ll start with “Brightness,” since it’s their most egregious category. I already discussed in my most recent Mailbag that a huge number of problems in a 3D experience can occur because of improper theater calibration. Providing a 3D grade for brightness is as misguided as calling out a cinematographer for filming a movie “too dark” when your TV picture settings are off. Yes, wearing 3D glasses slightly darkens the image on the screen. But believe it or not, we have the technology to make screens shine a bit brighter to compensate for this!

So why do some theaters forgo displaying their movies at an adequate level of brightness for a 3D experience? A) It costs more money to power the projectors, B) It burns their projector bulbs out faster, and C) They’re too cheap to get their equipment professionally calibrated. If your local theater’s screen isn’t bright enough, it’s not the 3D’s fault. You’re just in a shitty theater. (With great sadness, I must report that most theaters in the U.S. are pretty shitty.)

Fun Fact: Technical perfectionist that he is, Michael Bay demanded that theater screens be calibrated to an exact level of brightness to display Transformers: Dark of the Moon. That way, people wouldn’t blame a lack of clarity on the 3D.

Bumblebee in Transformers Dark of The Moon
If he hadn’t, it may have been difficult to tell what was going on here.

“THE GLASSES OFF TEST”

Cinema Blend’s next absurd category? The “Glasses Off Test.” In case you’re not familiar with the basic concept of how it works, 3D screens are displaying two images at once – one for your left eye and one for your right. The glasses make sure you’re seeing the correct information with each eye. When you take off the glasses, you see both images, which are slightly offset from one another. The further objects get from the screen plane, the more split apart the two images are, and the blurrier the edges of objects get.

In my three years working in conversion, do you know what I never saw anybody do to test the quality of 3D? Yeah. You guessed it. The very concept of taking your glasses off to quantify 3D quality is absurd.

DEPTHINESS ≠ QUALITY

Depthiness and dramatically divided spaces in no way represent 3D quality. That would be like using the number on a sound meter to grade how well an orchestra was playing. It would be like scoring works of art based on how bright its color palette is. Sure, there are times when volume or vibrancy will bolster works of art, but using these factors as a blanket measure of quality is egregiously missing the point. Some of the best instances of 3D I’ve seen have been subtle – like the disparity between Carl’s face and the flattened photographs of his past in Up. How about in The Avengers, when Loki’s spittle hits the wall of his glass cage as he calls Black Widow a “mewling quim”? Some of the most effective and beautiful 3D shots I’ve seen haven’t been big but had the texture and subtlety needed to make them beautiful. Take off your glasses during these shots, and you might not even be able to register the double image.

“AUDIENCE HEALTH”

The other thing about excessive Depthiness – it tends to make people vomit. It’s amusing that Cinemablend also has an “Audience Health” score discussing how comfortable the 3D is to watch, because it contradicts their “bigger is better” attitude. Larger than life 3D may feel more impressive momentarily, and it sure is blurry when you take those glasses off, but a good percentage of brains will reject it (and subsequently their lunches) if it’s used too much or in the wrong moments. This is actually one of the huge lessons that Hollywood learned too late. Avatar‘s 3D was big and amazing, but for the weak of stomach, its 3D turned a lot of people off to the format. (More on this in a moment.)

Speaking of that “Audience Health” quantifier, Cinema Blend regularly cites “motion sickness” as the reason for a bad score in this category. I’ve also heard this described as the “Dashboard Effect” among 3D professionals, (as in, you get sick behind the dashboard of a moving car). But describing this as “motion sickness” belies the actual cause of the problem and therefore misrepresents its solution.

So what metrics actually tell you you’re getting a superior 3D experience? I’m glad you asked.

3 TRUE TESTS OF 3D QUALITY

Test #1 – THE DEPTH BLEND

What Cinema Blend is actually discussing when they talk about comfort of viewing is the film’s Depth Blend (also referred to as a Depth Continuity or Smoothing). Depth Blending is a little difficult to describe without a 3D demonstration, so let’s sidestep a moment and discuss a similar concept in 2D.

Here are a few seconds of a particularly hectic scene from Mad Max: Fury Road.

Max and his Interceptor
Max is about to be crushed by his own Interceptor.

While the rules of art and cinematography tell us that an image is going to be prettier with the focal point offset to the 1/3rd or 2/3rd point, director George Miller wisely focuses all these shots dead center. Why? Because in just a brief moment of screen time, we see the positions and reactions of Max, Furiousa, and Slit. There’s no time to slow down. This is a car chase! Miller wants you to understand Max’s danger in an instant, without confusion, and certainly without motion sickness. As long as you’re focused dead center (while on the edge of your seat), you’re not going to miss a thing. (Although, you may be an unfortunate exception if you were born before the advent of MTV.)

So what does this have to do with 3D? Well, Depth Blending works the same way, just on a different axis. Let me show you another example.

Shane Come Back
Shane… come back!!!” (Shane never did pay Joey the 500 buckeroos he owed him.)

Imagine for a moment, that you were watching the final scene from Shane in 3D. (If you can manage, try not to think about how wildly inappropriate Shane is for a 3D conversion.) There’s very little movement in this scene – just cuts between Shane riding off into the wilderness and a Close Up of little Joey Starrett iconically crying after him. But think about the disparity in depth between these two shots.

Here’s where minds behind a 3D presentation have to get creative, because if the cut between these two depths was “realistic”, it would make this touching moment incredibly disorienting. Near. Far. Near. Far. The scene cuts back and forth between these camera setups four times. It’s enough to make someone sick… which is often what is actually happening when people begin to feel “motion sickness”.

CHEATING IT – More Real Than Reality

Ideally, the 3D of little Joey’s close up would be played at Screen Plane (so he looks like he’s right in front of us). For the purposes of this scene, however, he might be played a bit deeper into the screen. That way, when our focus shifts to the Wide, we’re not jarred by the perception shift. In the Wide, the foreground might be pulled closer to the audience than usual, with Joey placed at roughly the same Focal Plane that he was in the previous shot. This way, when our attention shifts from Joey to Shane in the distance it feels like a natural transition.

Yes… this is cheating. Doing this gives a more or less “false” 3D impression, in that the audience’s location in 3D space is going to differ slightly from the camera’s. But as long as everything in the image is properly proportioned, the effect will be seamless. For those who haven’t worked in film before, I hate to break it to you, but cheating of this sort happens all the time – lens distortion, spatial orientation, object location, lighting, performance pacing and line delivery. NOTHING you are seeing is real. That teary-eyed performance Joey gave? He was probably delivering it to the director while Alan Ladd (Shane) was screwing the Makeup artist in his trailer.

Jake Sully meets Pandora dandelions
Space dandelions on Pandora.

Jumping back to Avatar‘s 3D (like I promised), the film had BIG 3D but no real depth blend, which made some viewers sick. In James Cameron’s defense, not enough was known about stereography at the time for him to realize this was necessary. It’s not that his film’s 3D placement was “incorrect,” it was just frequently disorienting. So if you gave up on 3D after Avatar, maybe try it again at a good theater. Stereographers know a lot more now, and 3D comfort has improved by leaps and bounds.

Test #2 – THE 3D ENVIRONMENT

There are three ways to create a 3D movie environment. One is completely digital with VFX. This isn’t easy or cheap, but (for better or for worse) anything you can film now, you can create in a computer. There’s a reason why “we’ll fix it in post” is an ongoing industry attitude.

The next, and old-fashioned way, is to film a live environment with two cameras simultaneously. One camera for the left eye and one camera for the right. This is called “Shooting Native”.

Finally, we have Conversion. Aside from being the most convoluted, it’s also the Black Sheep of the bunch. In this case, standard 2D footage is taken, a painstakingly created digital 3d environment is applied to the image, and a distortion is applied to break it into two unique images. These images trick your brain into thinking that what you’re seeing is in three dimensions. A lot can go wrong in a conversion, and it sometimes does, with uncomfortable or unnatural results. Yet Conversion is quickly becoming Hollywood’s #1 method of producing 3D content, so we’ll start there.

THE CHALLENGES OF 3D CONVERSION

The conversion of a single 2D shot into theater quality 3D can take days of work. I’ve seen longer and more complicated shots take weeks (sometimes with artists working weekends and 12-hour shifts). First, the 3D environment needs to be flawlessly mapped to the 2D image. This involves building out and animating faces, bodies, and everything else you see in a movie. The closer an object gets to the screen, and the bigger the depth you’re using, the more detail it needs to be a convincing approximation.

Everything in the frame must be placed at the proper depth in relation to everything else. Otherwise, an effect occurs where objects appear to be miniaturized or enlarged from their actual sizes. In addition, a lack of finesse or detail is a sure sign of a cheap and rushed conversion. If you watched Ant-Man in 3D, you may have noticed that in Close Ups, the characters’ faces were often flat often sloped unnaturally. (Cinema Blend didn’t seem to notice, awarding the film a 4/5 in “Planning and Effort”.)

But wait! There’s more. Now the edges need to be cleaned up.

Distorting an image into two separate perspectives does A LOT of damage to an image, and repairing that damage is usually harder than placing the shot in depth in the first place.

Just think for a moment how difficult it must be to track the borders of individual flyaway hairs or foliage. (In rushed conversions actors are missing any flyaway hairs and branches of leaves often looks like blobs.) Often, fine details like these need to simply be erased and then digitally redrawn. That’s right. Human flyaway hairs are often a special effect when you’re watching them in 3D!

Martha Kent in Man of Steel
Martha Kent’s hair – The most underappreciated Special Effect in Man of Steel. STEREO ARTIST (to ME): “Dammit! Would somebody get this bitch a comb!?!”

But wait! We’re not done yet!! Let’s do a quick experiment.

  1. Close one eye
  2. Hold your finger up between your eye and a fixed object
  3. Holding your position, open your other eye
  4. Notice how with two eyes, you can now see the previously obstructed object
  5. Imagine that object is a moving human face
  6. Imagine having to digitally recreate moving human faces from scratch behind partially obstructed objects throughout a movie
  7. Suddenly realize how underappreciated the thousands of uncredited VFX artists are that tirelessly work to entertain you

That’s right. Not just the edges of 3D objects may be digital fakes, the stuff just behind those objects likely are as well. Depending on what’s going on in the shot, this process may be as simple as sampling the surrounding frames for background information, or building a complicated puzzle of details from surrounding shots, frankensteining them together and praying to God it looks okay. Depending on how lenient the client is, or how the shot is framed, sometimes the background can be simplified, or foreground objects can simply be blown up to cover up the gap. But the bigger the depth difference – the bigger the gap – the more the original composition must be altered to cover that gap. (If the edges around a foreground object are warped, this is because that missing information wasn’t properly filled. It’s yet another sign of a rushed conversion.)

So why the hell would anybody do a 3D conversion over filming Native!?!?

This is a fair question. Think of all the drawbacks:

  1. Rushed 3D Conversions look like garbage
  2. Good 3D conversions take a TON of time, money, and manpower
  3. Distortion from Conversions GUARANTEE that the image will never look identical to the original

THE CHALLENGES OF SHOOTING NATIVE

Despite the allure of Native 3D’s “purity” and perceived simplicity, there are a number of downsides. The biggest of these is that shooting Native is way more complicated than you’d think.

Peter Jackson Showing off The Hobbit 3D camera rig
Peter Jackson shows off his fancy HFR 3D camera rig.

COST

This may be surprising to folks who haven’t worked on a film set, but the cost of shooting in 3D goes well beyond the amount of the camera rig. During an average film shoot, the most likely element to hold up a film crew will be camera setup – lighting adjustments, lens changes, focus measurements, tracking practice, character placement, environment cheats. This is why even lower end films tend to look so much better than most television. It takes exponentially longer to compose a shot that looks cinematic than it does to just capture an actor saying a line.

Now imagine adding a whole additional step to that process – having to know in advance how your 3D is going to play in a scene – fiddling with the settings until you get the exact depth you wanted. Imagine how much more likely you are to have equipment trouble when you add a whole extra camera to the mix.

Imagine your entire crew just sitting there waiting for your camera team to figure it out, and having to pay to block off that street in New York City for an extra day because you thought Native 3D would make your movie look better.

FLEXIBILITY

In short – Native 3D gives you very little flexibility. This is creatively the biggest problem with the format. Normally, a lot of happy accidents can happen in the editing room but locked and awkwardly blended 3D footage could feasibly prevent you from easily making a cut that works GREAT in 2D.

So let’s say a director is worried about this. They play it safe by focusing every shot at Screen Plane. But now you have a boring 3D experience. Coincidentally, this is how I felt about The Hobbit trilogy. Sure, the 3D looked “accurate” a lot of the time, but many moments that should have felt grand were shrunk back into the screen.

Compare this to Conversion, where the 3D environment of a scene can be changed on the fly to make maximum impact. Complete scenes can be more easily “Scored” (like a soundtrack) to ease off the 3D effect, then slowly build it to a crescendo during an exciting moment for MAXIMUM IMPACT!

Water on LensARTIFACTS

Another thing that not many people know about shooting in native, is that it’s often impossible. Our company did some work on Enchanted Kingdom, a documentary filmed by the folks responsible for Planet Earth. (For my money, this film is the best 3D experience I’ve ever had.) A good chunk of the movie was filmed Native, but there were times that doing so was impossible.

Some of the time, this was because their 3D equipment broke, or was too unwieldy to get to their shooting locations. We did a lot of Macro Close Ups – I can only assume because capturing 3D Close Ups of lizards in the wild is difficult.

And then some images looked similar to the above image of the wave. You’ll notice that there are flecks of water on the lens of the camera. If you were trying to film this Natively, you’d have flecks hitting one lens that didn’t match the other. This is what we call a Natural Artifact. It may be “realistic” to have objects appearing in one of your eyes that doesn’t match the other, but it’s certainly not pleasant.

The wave sequence in Enchanted Kingdom is mesmerizing, but it would have been impossible to see it so clearly in “real life” or with Native 3D because you’d be half-blinded during the experience.

COHESION

So let’s say you’ve painstakingly filmed your entire movie in flawless Native 3D. But since you’re doing a fantasy film, there are all kinds of separate takes and VFX elements that you have to merge with your footage. This process is difficult enough with a malleable playing field, but it’s going to be even trickier when your principal cast and parts of your environment are locked in place.

My 3D eye wasn’t good enough to see it, but friends have told me that The Hobbit exhibited a number of problems in this area. Environments clipping with VFX and actors clipping with the environment. Suddenly the “flawless” Native experience isn’t so cohesive.

THE PERFECTION OF VFX & COMPUTER ANIMATION

Finding Dory watching School of Stingrays
Fortunately, these Stingrays already exist in a 3D environment.

This all leads us to Visual Effects and Animation, the golden children of the 3D world. Placing a second virtual camera in an already virtual 3D environment only requires additional rendering time from the computer. No additional manpower needed! And even better, these environments can be manipulated just like a conversion if necessary, for all the flexibility that brings.

Computer animation (and 3D VFX scenes in live action movies) should always be flawlessly detailed, with perfectly defined edges, and an easily organized Blend and dramatic flow. There’s really no excuse for them not to be. The only thing that could go wrong with 3D animation is…

Test #3 – 3D FIT

Let’s wrap this up on a positive note and discuss the scoring metric that Cinema Blend gets unequivocally right – the “3D Fit Score”. This determines whether or not your 3D was done purposefully, or tacked on as a gimmick.

There’s a tendency to think that BIG movies are naturally going to be better in depth, but the truth is that regardless of size, some movies are simply filmed better for 3D than others. Gravity, Life of Pi, The Walk, and Ghostbusters are prime examples of films that simply aren’t the same if you watch them in 2D. These were built from the ground up with a 3D experience in mind and were directed for both immersion and comfort. Well staged framing and multiple layers of depth. Heavy use of transparencies, clouds, or particle effects. Long steady takes with room to explore the frame. Use of the Z axis to not only grow the stage but enhance the motions within it.

The Amazing Spider-Man 2 animated
The Amazing Spider-Man 2 had it’s fair share of problems, but those swinging scenes were the most effective use of 3D I’ve ever seen.

Then there are movies like Man of Steel, which was given a subtle and beautiful conversion, but with unfortunately limited appeal. Director Zack Snyder uses excessive shaky-cam, opens scenes with doors closing on the camera, and fills the foreground of many frames with massive, out-of-focus elements. All these techniques can work fine in 2D, but it takes a tremendous amount of finesse from a stereographer to make it not look like garbage in 3D. Rather than the movie working to earn the effect, it works against it and it never truly feels “right”.

Movies like Man of Steel may be technically perfect 3D experiences, but they sure don’t do anything to illustrate how the format can be so much more than a gimmick.

In Conclusion…

This may be my bias showing, but 3D is incredible when done right. And despite all the ways that it can go wrong…

  1. Filmed improperly
  2. Converted improperly
  3. Displayed improperly

…I still think that the benefits of a stellar 3D experience are worth the trouble of finding it. And hey – I’ll keep reviewing 3D to steer you readers in the right direction, and with any luck, this 101 course will help other reviewers to speak knowledgeably about 3D as well.

Happy viewing!

It's like I Can Touch You by Don Hertzfeldt


101 is a series of editorials whose purpose is to discuss the fundamentals of how we relate to art and movies in particular. Is there some element of film that has you’d like The Powerwolf to explore more deeply? Let us know at mailbag@powerwolf.ink!