I started reading the vbprog.pdf, http://www.vr32.de/content/downloads/documents/vbprog.pdf

I wonder about image 2.1, page 3…

Aren’t close objects supposed to be wider apart (in affine or composite mode) than objects far away?

Why then are the blocks (I suppose by the orientation and background that the picture is showing three blocks of the same size at different places in the picture) farthest away most spread apart?

The picture is wrong, isn’t it?

Yeah, I have reacted on that one too… the picture must be wrong.

Actually, if you look at the first two images as a stereo pair, you’ll see that it’s correct (not “good”, just correct ;-)).

Here it is made into an anaglyph for those with 3-D glasses (red over the left eye). I left the labels on so you can compare the two by covering each eye in turn.

When I’m thinking about 3-D images (such as when making art for VB demos) I try to remember that things farther away move right in the right image, left in the left image, or both.

Attachments:

1. 

   stereo_explanation.png
   

That’s correct. It’s confusing if you don’t approach it the right way, so I made a quick illustration:

Attachments:

1. 

   stereo.jpg
   

I guess it all depends on where your focal point is…

If you focus your eyes on an object in the distance, and alternate between closing your right and left eye, then objects closer to you will move further apart and objects further away will “stand still” more and more the closer they get to your focal point.

But if you focus your eyes on an object close to you, the opposite occurs, and the object in the distance will move the most apart… so it’s the angle of the eyes, or cameras that decide.

In Yeti3D, I have put the cameras a bit apart, but their line of vision is completly parallel to each other, like staring infinitly into the distance. It looks pretty good right? And it causes the walls in the distance to appear closer to each other than the ones in your face, aka the opposite of that picture in the manual.

Is this the correct way to do it though, or how do I know how far into the screen the player should focus (the angle between the cameras)?

I just thought of DanB:s avatar and it looked wrong.

@DanB:

DEFINITELY, always keep the “cameras” parallel when rendering stereo CGI. Simulating the swiveling of a person’s eyeballs is called “toe-in” and almost always causes more problems than it solves (we don’t want to make all those evil reviewers from the mid-90’s correct about the VB causing headaches, do we? :-P).

(You probably already know this, but) if you want to control where the “screen” plane is (no screen in the VB, of course, but you know what I mean, I hope) you need to leave some columns off of the left and/or right of the two generated images. This allows for things to “pop out” of the center of the frame, but can cause problems if things at the edge of the frame lie in front of the virtual screen plane. According to the “gun hand” sprite in the screen-shots (I still haven’t seen the real thing) you seem to be doing that now in Yeti. Are the renderings themselves also being shifted?

It’s up to you, of course, but I would avoid “piercing” the VB’s screen. It’s probably okay for the gun-hand, though, since it can’t move out of the frame.

@KR155E:

That is a very useful diagram (despite the English mistake ;-))! Quick, find a spot for it in the Wiki! 😀

It’s “Farer away”, right? 😉 Before posting it in the Wiki, I think I should make a better one, maybe even several for objects at different distants instead of putting all in one image, which can get confusing quickly, with all the lines and colors and blocks. 😉

KR155E wrote:
It’s “Farer away”, right? 😉 Before posting it in the Wiki, I think I should make a better one, maybe even several for objects at different distants instead of putting all in one image, which can get confusing quickly, with all the lines and colors and blocks. 😉

It is farther or further away. But in your case it is farther as it is about distance.

Or just check here to learn more about it:

http://www.lessontutor.com/eesfarther.html

Thanks! I knew “farer” sounded weird. 😀

See also: Figures 25 through 27 of U.S. Patent No. 5,682,171 and their related text.