To toe in or not to toe in, that's the question

[Leonard Levy]

Thanks Alister,
Very helpful information and it sounds like a very logical way to work.

It's been a treat to have the benefit of all your experience both on this board and on the EX-1 board where I've been following your posts for a few years. I have a few more related questions if you don't mind, but I'll take a little while to formulate them.

Lenny

[Sareesh Sudhakaran]

This is a great thread. Let's say if one is shooting a simple drama with no visual effects or explosions or gimmicks, will I get acceptable theatrical results by just using an inter-axial rig and no toe in at all?

Mostly shot on wide angle in rooms without moving walls but not handheld. What should my main focus be in this case? Would I be better off with wide angle or long lenses in this case?

[Bruce Schultz]

Originally Posted by Petri Teittinen

I have the side-by-side 3DFF Indie rig, the cheapest of the lot at $499. Couldn't afford anything fancier.

Even easier but less accurate - you can loosen the camera mounting screws on one of the plates you have and toe one camera in towards the other until the foreground images (if that is where you are focused) seem to merge together and get reasonably sharp. That is basic convergence right there.

[Alister Chapman]

You should get perfectly acceptable shots either parallel or toe'd in. Provided the camera are truly parallel then that's the safer way to shoot if you are at all unsure about your abilities or rig setup. Toe in gives you more control over on screen depth and usually give a slightly rounder image which looks more natural, but needs more care and good on set monitoring.

The longer the focal length of the lens the more obvious any rig alignment errors will be. For most S3D shoots I will try to keep to a focal length of less than 75mm (on a 35mm sensor). Typically lenses that come close to the human field of view make shooting good S3D easier. Very wide lenses will be problematic with mirror rigs as you will often end up getting the extremes of the mirror or second cameras in the shot and with side by side rigs it can hard to get the cameras close enough together. A very wide lens will give you very flat looking 3D, so you end up moving the cameras apart to add depth, which then leads to cardboarding. 35 to 50mm is an easy range to work with.

I'm helping to run a 3D event in Chennai at the end of the month.

[Leonard Levy]

Originally Posted by Alister Chapman

Y A very wide lens will give you very flat looking 3D, so you end up moving the cameras apart to add depth, which then leads to cardboarding.

Yikes, I thought it was the opposite , that a long lens led to flat looking 3D and cardboarding.

[Petri Teittinen]

Originally Posted by Bruce Schultz

Even easier but less accurate - you can loosen the camera mounting screws on one of the plates you have and toe one camera in towards the other until the foreground images (if that is where you are focused) seem to merge together and get reasonably sharp. That is basic convergence right there.

Yup, I know... but I have nothing but the cameras and the rig. No external monitors, zilch, nada, zip.

[Alister Chapman]

Cardboarding or the cut out effect is one much discussed and debated. It occurs whenever the relationship in a shot between the total image depth(volume) and relative depth of individual parts of the scene are miss matched. It often occurs whenever there is a conflict within the scene between the scale of objects and their relative stereoscopic distances. It can occur at any focal length. It is very common with long focal lengths, as compared to what we normally see with our eyes the image is foreshortened or compressed. This has the effect of squashing the image into a series of flat looking planes. To compensate for this it is traditional to increase the interaxial as this increases parallax which then lengthens the squashed planes. With medium length lenses this can work, but it will depend on how far the subject is from the camera. A subject close to the camera will gain more roundness as the cameras sees more of the modelling of the individual object but subjects a long way from the camera will not become any rounder. Moving the cameras apart will compensate for the lens foreshortening as this increases the depth or volume of the scene, but if the planes within the image have no roundness because they are too far from the camera then all that happens (IMHO) is that the increase in depth or volume simply separates the flat planes still further creating a very un-natural effect. In addition exaggerated interaxial's lead to miniaturisation where the brain is confused by the scale of objects compared to their apparent distances. With very long focal lengths cardboarding or the cut out effect is very hard to eliminate. This planar effect is dependant on many factors, focal length, interaxial, scene depth and how near or far subjects within the scene are to the camera.

Going back to very wide lenses the opposite can occur as these will tend to give a scene with a small on screen volume relative to the depth of the real world scene. If you have a nice range of subjects in your scene, some close to the camera, some further away then generating a good stereoscopic image will be straight forward. However if everything in your scene is on a similar plane or close to the rear of the scene then the image may look very flat or lack volume. So to compensate the temptation is to increase the depth (volume) by increasing the interaxial. This "stretches" the scene depth or volume, exaggeration the separation between objects. If the camera is not close enough to those objects to show increasing roundness then cardboarding or planar effects will once again become apparent.

Of course this will happen at any focal length. So you must balance focal length and interaxial with both scene depth and subject distance. A wide angle lens will work best when you are close to your subject, but is not a substitute for a narrow interaxial. I like to try to work at FL's between 35 and 75mm (35mm equiv) and would much rather move the camera closer to the subject or further away than go outside that range. Obviously there are many time when you do have to go longer or wider and this is when creating a good stereoscopic image becomes much more challenging and much more subjective as in either case we are no longer presenting an image anything like the images seen by us humans.

[Leonard Levy]

Thanks again Alister for the best explanation of "cardboarding" I've seen anywhere else. This is helpful to me. a great thread. Thank again to everyone.
By the way, I just played with the method you suggested earlier ( for setting convergence with a slight toe in based on the furthest distance ) for a few minutes yesterday while mainly working on rig issues and it sure seemed to me like an elegant way to work on the set and see what your getting right away on the monitor without the other disadvantages of full convergence.. With parallel we have to go in to a Davio Box and readjust the horizontal offset constantly to get the convergence right for the monitor. Not that bad but slows us down with a small crew.

[Mathew Orman]

Originally Posted by Leonard Levy

Yikes, I thought it was the opposite , that a long lens led to flat looking 3D and cardboarding.

You are correct.
The wide angle setup will have normal depth in the center but progressively going to the left or right edge the volume gets contracted or squashed to 0 at 90 deg due to effective interocular distance decrease.

Mathew Orman

[Mathew Orman]

Originally Posted by Alister Chapman

Cardboarding or the cut out effect is one much discussed and debated. It occurs whenever the relationship in a shot between the total image depth(volume) and relative depth of individual parts of the scene are miss matched. It often occurs whenever there is a conflict within the scene between the scale of objects and their relative stereoscopic distances. It can occur at any focal length. It is very common with long focal lengths, as compared to what we normally see with our eyes the image is foreshortened or compressed. This has the effect of squashing the image into a series of flat looking planes. To compensate for this it is traditional to increase the interaxial as this increases parallax which then lengthens the squashed planes. With medium length lenses this can work, but it will depend on how far the subject is from the camera. A subject close to the camera will gain more roundness as the cameras sees more of the modelling of the individual object but subjects a long way from the camera will not become any rounder. Moving the cameras apart will compensate for the lens foreshortening as this increases the depth or volume of the scene, but if the planes within the image have no roundness because they are too far from the camera then all that happens (IMHO) is that the increase in depth or volume simply separates the flat planes still further creating a very un-natural effect. In addition exaggerated interaxial's lead to miniaturisation where the brain is confused by the scale of objects compared to their apparent distances. With very long focal lengths cardboarding or the cut out effect is very hard to eliminate. This planar effect is dependant on many factors, focal length, interaxial, scene depth and how near or far subjects within the scene are to the camera.

Going back to very wide lenses the opposite can occur as these will tend to give a scene with a small on screen volume relative to the depth of the real world scene. If you have a nice range of subjects in your scene, some close to the camera, some further away then generating a good stereoscopic image will be straight forward. However if everything in your scene is on a similar plane or close to the rear of the scene then the image may look very flat or lack volume. So to compensate the temptation is to increase the depth (volume) by increasing the interaxial. This "stretches" the scene depth or volume, exaggeration the separation between objects. If the camera is not close enough to those objects to show increasing roundness then cardboarding or planar effects will once again become apparent.

Of course this will happen at any focal length. So you must balance focal length and interaxial with both scene depth and subject distance. A wide angle lens will work best when you are close to your subject, but is not a substitute for a narrow interaxial. I like to try to work at FL's between 35 and 75mm (35mm equiv) and would much rather move the camera closer to the subject or further away than go outside that range. Obviously there are many time when you do have to go longer or wider and this is when creating a good stereoscopic image becomes much more challenging and much more subjective as in either case we are no longer presenting an image anything like the images seen by us humans.

Wrong assumption.
Wide angle shoots require wide angle viewing or else gross perspective distortions defeat the purpose.

Mathew Orman

[Mathew Orman]

Originally Posted by Leonard Levy

Thanks again Alister for the best explanation of "cardboarding" I've seen anywhere else. This is helpful to me. a great thread. Thank again to everyone.
By the way, I just played with the method you suggested earlier ( for setting convergence with a slight toe in based on the furthest distance ) for a few minutes yesterday while mainly working on rig issues and it sure seemed to me like an elegant way to work on the set and see what your getting right away on the monitor without the other disadvantages of full convergence.. With parallel we have to go in to a Davio Box and readjust the horizontal offset constantly to get the convergence right for the monitor. Not that bad but slows us down with a small crew.

"cardboarding" is due to viewing narow view angle shoots with normal view angle. Example: 10 deg telephoto and 40 deg on monitor.


Mathew Orman

[Alister Chapman]

Cardboard predominantly occurs when there is an imbalance within the on screen image between the total image depth or volume and the apparent distance between the objects in the scene.

Viewing an S3D screen using different FOV's will not change this relationship as what we are viewing is a pair of flat images, not a true three dimensional scene. All it will do is make the image bigger or smaller which will increase/decrease the apparent disaprity, but the relation ship between the dissparity and on screen distances will remain exactly the same as what you are looking at is essentially a 2D image not a 3 dimensional one.

I don't follow you at all Mathew. If you shot a distant object with long lenses against an even more distant background to give some depth to the scene you would most likely use a wide interaxial. This shot would almost certainly have some cardboarding. Lets consider a shot of a soccer game from the back of the stadium. The camera is 150ft from the pitch using a long lens and lets say 18" interaxial. In the shot you can see the width of the pitch and a couple of payers. The two players A and B appear to be of similar size (due to foreshortening) even thought they are on opposite sides of the pitch. The players themselves show little depth as the cameras are simply too far away compared to the interaxial to resolve any significant depth in the players themselves. Our parallax depth cues however are telling us that the players are 40ft apart as the disparity difference between the front of the pitch and rear of the pitch are quite large due to the hyperstereo. Our brain will interpret this as flat cardboard cutout players on a pitch with some depth to it.

Mathew, are you suggesting that if I went to the cinema to watch this using a pair of binoculars this would some how help? The magnification introduced by the longer focal length of the bino's would increase the apparent disparity thus increasing the volume(depth) of the pitch, but the players will still appear flat as the cameras were not far enough apart to resolve any meaningful 3D in them, the relationship of the size of the players will stay the same and the parallax will also increase so if anything it will make the situation worse not better.

If we were projecting a true three dimensional image then I would agree with you, but we are not. We are showing flat 2D images presented in such a way as to create the illusion of depth and changing the scale of these flat 2D images or magnification of the them will not alter the ratio between parallax, scale or disparity within those 2D images, that was determined when the images were created.

[Mathew Orman]

"Mathew, are you suggesting that if I went to the cinema to watch this using a pair of binoculars this would some how help? "

No it is the opposite, you would match the camera's view angle by using -5 diopter
correction glasses. Which would give you a screen size appropriate to match the 10 deg telephoto image.
Human eyes have no ability to zoom.
Telephoto stereo shots are only good to emulate using a pair of binoculars
and in such case carboarding is natural as experienced by many people.
Appropriate use at cinemas should include circular mask to inform viewers that camera is emulating binoculars.

Also if you want to see cardboarding of flattening by mismatched view angle then just get you face about 3 inch from your desktop monitor.
Opposite happens (excessive stretching) if you get away from normal position something like 10 feet .

Mathew Orman

[Alister Chapman]

So your proposal Mathew is that we only ever shoot and display 3D using ortho-stereo or provide the viewer with corrective glasses to match the FOV of the camera?

Corrective lenses will still not eliminate cardboarding as I described and as being discussed as the cardboarding is recorded in a 2D image, so no corrective viewing devices will eliminate it because the ratios contained in the image are incorrect. Magnification or reduction of the image will not change those 2D ratios, only the scale. This will reduce or expand the distance between the flat planes, but they will still be flat planes because unless you had an extremely wide interaxial (with associated disparity issues) in my soccer game example the camera separation would not have been wide enough to resolve any depth in the players themselves. We are talking about the real world here with limited camera and lens resolutions (and thus depth resolution) and not a hypothetical world with unlimited resolution.

[Mathew Orman]

"Corrective lenses will still not eliminate cardboarding"

Wrong again, corrective lenses make distant object blend with background
totally flat as perceived by human eyes. So there is no cardboarding effect observed.
Corrective lenses with negative diopter are used in the cinema system that has the sweet spot seat at the last row and in such system every one has the same FOV that matches displayed content.
Also, orthographic stereo images are always made with orthographic cameras.
Matching stereoscopic geometry between viewer and the content has nothing to do with orthographic stereoscopy.
I suspect that you have never seen an orthographic stereoview.
But if you did then please provide an example.
Finally you can test your stereoscopic geometry knowledge
by providing an answer for a simple question:

What is the infinity parallax in inches for a screen that is 30 feet away?

Mathew Orman