How it's possible to enhance the depth effect of 3d pictures without increasing the cameras distance?

Question

I always thought that for a 3d picture, which is in fact 2 pictures that are displayed one for each eye, the more far the cameras are, the more "depth" you will see. And it's a fact i think.

So, for example, if you take the picture of a house and the cameras distance is let's say, 10m, you'll have the impressions that the house is a miniature, (and you could also take the pic using a very large aperture to enhance the effect).

However, today a little friend showed me the 3Ds which is a portable videogame that can also take 3d pictures, and the first thing i notice is that the distance between the cameras is minimal, i thought it would only be capable of taking close photos of small objects and at the photo they would look larger, so you'll have the same impression of an eagle for example.

For my surprise however, it's not what happens, the thing can take pictures of even very far (and big) objects and their scale look right, or even smaller by adjusting some options. You have the impression of an even larger distance between captures than the human eyes.

How can it be possible? I think it's "physically impossible", I mean, you could even map the picture on a 3d mesh and enhance the effect, but i think we'll be able to notice artifacts in this case, because, for example, if you have a box, and the left camera can't see the side on the right, and at the distance of a human eye, the right camera will be able to see, but if you take the right camera closer to the left one, it will not see the right side of the box, so it's a visual information that can't be captured when you have a small distance between the cameras, and i can't see how it's possible to simulate something like this.

I couldn't make extensive tests, but I want to know how that little device tricked my eyes so impressively. Because obviously it's a trick, i think...

Any clues? Thank you.

Answer 1

For the reasons that you already mentioned in the question, it would not be possible to modify stereoscopic depth without adding artifacts.

So instead I would argue if the depth is actually increased. It probably isn't.

We have gotten used to watching monoscopic (regular) photographs with both our eyes. Theoretically the perceived scale of the scenes on those photographs should be huge. But we have gotten used to it.

So, the depth effect of an stereoscopic image that is captured in the range of monoscopic (0 camera distance) up to a regular human eye-distance will be perceived natural. A small amount of depth is already obvious to the brain.

If you would increase the camera distance beyond human eye distance, then your brain would get an unusual stimulus hence it might conclude that the scene is a miniature.

Also note that the viewing angle (field of view) of a photograph or mobile screen is also significantly smaller than the captured angle. Not that it would compensate for reduced depth, but just another example of our tolerance in accepting the illusion of a photograph.

If you are shooting images for a more immersive viewing experience (eg. 3D cinema), then the tolerances are probably much smaller.