In Part One we studied the importance of depth, and why you shouldn’t take it for granted. In this part we’ll see how your choice of depth affects everything else. We’ll start with the depth-budget.
How to prepare a depth-budget
We already know how to prepare a depth-chart and depth-script. I’m sure it looks intimidating to you. After all, what place does numbers and graphs have in art?
Well, the answer is: You asked for it. Deal with it.
Here’s the chart again for reference:
You can’t go haywire with your depth chart, no matter where your imagination takes you. You are constrained by many limits. Let’s look at them in detail.
Limits of Parallax
Since the 3D left and right images are only displaced horizontally, you only consider the horizontal width of your display.
No matter how far or close together you place your cameras, the end result will be seen by a pair of eyes with a fixed inter-ocular distance of 2.5″. The eye shouldn’t be forced to diverge beyond this point, and this sets the limit for how far the left and right versions of you object are in the final onscreen image. This difference is parallax.
Every screen has a fixed parallax limit.
Let’s say the width of your monitor is 50″ (inches). The screen limit of parallax is (2.5/50)*100 = 5%. Here’s the formula (courtesy of Tim Dashwood):
Maximum Parallax for a screen (%) = (Inter-axial distance / Screen width)*100
What this means is, the dots on the graph can’t move more than a certain limit in either direction (positive or negative). Remember, the dots on the graph represent real objects, maybe characters in your movie, etc.
What happens when you break this rule?
If you go higher than the maximum parallax limit, you will force the audience to diverge their eyes – which is painful. This is a definite no-no. If it happens for long enough, the audience will fight it, will feel uncomfortable, and eventually stop watching. This image should help you understand:
You can never create a parallax of greater than 2.5″ on any screen, because that will make your eyes diverge.
The reason we calculate the parallax limit in percentages is because screen sizes differ, but the inter-axial distance of the eye stays at 2.5″, and your parallax is a direct result of it.
In the first diagram, let’s say we replace the phrase ‘Screen Parallax in Pixels’ to Screen Parallax in %. You can easily draw an upper and lower line at the limit of parallax, ones that shout: “Do not cross!” Your graph can only dance within this range.
Should the interocular distance of the eye be the standard throughout your movie? Should your cameras always be 2.5″ apart?
No, of course not.
But it could, in which case we call the experience Ortho-stereoscopy. This isn’t mandatory. Changing the inter-axial distance has its benefits. Analogy-wise, it is as important as changing the focal length.
The ability to manipulate the inter-axial distance is the most important artistic tool in stereoscopy.
Long story short, increasing the inter-axial distance brings depth into faraway objects like mountains, skies, etc. If you’re doing an aerial shot this might prove useful. Overdo it, and the large-scale objects will look like miniatures.
Decreasing the inter-axial distance allows you to shoot objects smaller and closer to the camera. Overdo it, and the small-scale objects will look like giants. It’s different from how you tilt up your camera to make something or someone look menacing. Here, you actually feel the unnatural affect.
Finally, if you can’t find a fine balance with the inter-axial distance and the objects at different depths in your image, you will also get cardboard cutout figures.
Let’s look at some ‘rules’:
1/30th rule –
This rule of thumb states that the inter-axial distance must be 1/30 of the distance to the closest object.
E.g., if your actor is 6 feet away, the inter-axial distance = 72″/30 = 2.4″. On the flip side, it also means every time you want to shoot at 2.5″ inter-axial, you can’t have anything closer than six feet.
1/60th rule –
This rule of thumb states that the inter-axial distance must be 1/60 of the distance to the closest object.
E.g., to obtain an inter-axial distance of 2.5″, the closest object must be at 12 feet.
There are more rules, like the rule of 1/100, 1/120, etc. Which is your ‘lucky’ number?
I believe these rules are a total waste of time simply because you can’t go by rules on a shot-by-shot basis. How would you feel if someone forced you to follow the Hollywood system of long shots, mid shots and close-ups in every scene, regardless of content? I feel such ‘rules’ are detrimental. Many people start following it without understanding why it came to be in the first place.
How do you handle this problem? Study your situation carefully, and arrive at your own rules.
Many movies must shoot at an inter-axial distance of less than 2.5″ because actors have to get closer than six feet. You are limited by the laws of physics and optics. It’s a constant juggling act where the number of objects you’re juggling changes on a frame-by-frame basis.
This is why movies with visual effects are ‘willing to be made’ in 3D, while those without visual effects tend to stay away from this technology. When you shoot live action, you have to deal with millions of unpredictable elements that are for the most part simply beyond your control. How can you ensure correct parallax in such a scenario? My answer (which is going to be of no solace to you) is, you can’t.
Let’s go past that.
Positive or Negative?
Let’s assume you want your actor at 3 feet from camera. In that case, going by the 1/60 rule (nothing special about 60 it’s just a number) the inter-axial distance should be 0.6″ (1.5 cm).
Our parallax limit, assuming a 30 feet screen, is (0.6/360)*100 = 0.17%. We know that parallax is both positive (behind the screen) and negative (in front of the screen). If the maximum positive parallax is 0.17%, ideally the maximum negative parallax should also be 0.17%.
The total parallax, which we call the Maximum Depth-Budget, is 0.17+0.17=0.34%.
Important Note: In practice, some like to go beyond the maximum negative parallax, as our eye is able to converge a lot more than diverge. It is not mandatory to stick to the maximum negative parallax.
Whew! I know all this is not easy to take in. You might still be wondering: How does one determine the actual depth-budget of any scene?
My answer, to the best of my knowledge, is: You do it through trial and error. If you set limits through purely geometric calculations, you’ll be confounded by real-world limitations. Here are a few spanners in the works:
- Every member of the audience watches from a different distance, angle, height and perspective.
- A movie will be shown on many screen sizes, and even in the best of circumstances, one can only approximate.
- If your project is for TV or the web, you have a lot more choices.
- Humans don’t relate to 3D in the same way as a camera and lens does. Every property of a lens changes things.
- Real stereopsis is based on 3D objects in real space. 3D movies are always projected on to a flat surface.
- The resolution of the video and the lens; and the loss of resolution caused by sub-sampling, compression, light loss, artifacts, color science, display technology, camera filters, sensor noise, etc., are all beyond our control.
See? It’s not as simple as drawing a triangle on paper and finding a length or angle. You need a real good intuition about things. For this reason I’m with Lenny Lipton
“In order for stereoscopic cinematography to be a creative medium it’s got to be intuitive. If it’s not in the gut, it’s not going to succeed as an art form. If you have to rely solely on tables and calculations, composing stereoscopic images is not going to work.”
I strongly urge you to read Foundations of the Stereoscopic Cinema, by Lenny Lipton
Here, I’m with Tim Dashwood. Don’t converge if you haven’t done it before. The simplest way is to get it right on camera without converging. It also opens up many options in post.
If you do have to converge, do it under the insistence of an experienced stereographer. Even then, there must be solid irrefutable reasons for converging. When in doubt, don’t converge.
Resolution and Frame rate
Higher frame rates increase perceived resolution. Add that to high-resolution 3D imagery, and you get a heightened sense of realism. I’ve already explained about high frame rates here so I won’t go into detail.
Should you follow Peter Jackson (48 fps) or James Cameron (60 fps, maybe)? I believe so. I am a firm supporter of the 60p look. You may not, so stick to whatever suits your vision.
How many cameras per shot?
You need two cameras per shot, you know that already. But there’s also a school of thought that proposes three cameras:
Why not shoot with three cameras and get three inter-axial possibilities? Cool concept, though not easy to implement in practice, especially with beam-splitter rigs. It can be done with small cameras, though.
How many cameras per scene?
Finally, you have the consideration of having a multi-camera shoot. This is really tough to pull off, even in 2D. With 3D, you’ll need to always be juggling the parallax depth problem on every shot.
It’s not impossible, just tough for one person to do. It will play a big role in your decision to go 3D, so please don’t just ‘throw it out there’.
The three kinds of 3D rigs
The three kinds of 3D camera systems are:
- Beam-splitter Rig
- Side-by-side Rig
- Single camera rig
I’ve already gone into detail about these in Stereoscopy Rigs in the Comprehensive Guide to Rigging.
A single camera rig could include a mirror assembly system, but what I’m referring to is a 3D stereoscopic camera like the Panasonic AG-3DA1:
I’ll let you figure out the advantages and disadvantages of all these systems. Most of the time, your depth-budget and choice of inter-axial distance will decide which rig is best for each shot. On a feature length movie, you might find yourself using more than one type of rig. It’s common.
In Part Three we’ll look at how all these factors affect the budget and schedule. Of course, we’ll also look at what you need to look out for as far as post production is concerned.