Meant to be Seen

Hi,

This comes from a Wired online article:

http://www.wired.com/wiredscience/2012/02/3d-moon-map/" onclick="window.open(this.href);return false;

Get out your red and blue anaglyph glasses and check this out.

Its a revolutionary new stereophotographic method which allows a stereo image to retain its scale 'independently' of where the viewer is located.

Shown here is an anaglyph map constructed from lunar survey data. However the method can be applied to full colour 3D applications too.



The patent can be found in the article above or here:

http://www.google.com/patents?id=sRkLAAAAEBAJ" onclick="window.open(this.href);return false;


Thanks

I dont see any "greater POP OUT".

It's not about "greater pop-out", it's about being able to move relatively to the image (forward, backward, left, right, up, down) and seeing less distortion than with standard anaglyphs. I thought it was bogus at first but when I tried I was really convinced by the result.

If you look at this anaglyph of a similar scene but using the standard method, you'll see that the peaks seem to move according to your head movement, that's not the case on the first image.



I haven't read the patent yet so I don't know if this technique can be used for other types of scene but that looks very interesting.

Umm.. isn't that just a red/cyan anaglyph image? I call BS.

Hi,

I dont see any "greater POP OUT".

Firstly, it made you look, just like any newspaper headline .....

Secondly, it is true according to the author of the patent. Its just the height scaling has to be made so initially.




Thanks.

First the versions of the maps you are looking at are relatively low resolution so the effect is not so dramatic. The real trick that makes this special is that the effect allows you to look from almost any angle without losing the 3D image - even when displayed on a monitor or an i-Pad. you can visit http://www.popviewmaps.com" onclick="window.open(this.href);return false; and see some better examples.

If you visit http://www.kickstarter.com/projects/165 ... d-moon-map" onclick="window.open(this.href);return false; you can get a high res image emailed to you for supporting the work the inventor has done.

I partially read the abstract of the patent.
I don't undestand everything but I seem to understand the basics.

It seems this technique uses something we'd call an orthographic camera (most 3D modeling tools can use such a camera) but slightly modified to be able to make maps of round objects like planets without interference from the curvature of the earth (you could make it with a normal ortho projection but the end result wouldn't be usable as a map, it would be a simple picture).
But this type of projection has a cost : the picture for the left and right eyes would be identical (since translation has zero effect on an ortho image).
In order to produce 3D, you then need to fake a perspective by applying separation between the eyes according to the height of the map and apply a height scale that matches the size of the picture.
The root of the problem is that traditional cameras create a perspective that viewers must match. When the viewer stands in the ideal place in front of the screen the 3D picture is perfect. When the viewer walks away from this ideal place, the picture gets more and more distorted. The further away you stand from the ideal place, the more distortion you get.
With this special projection, there is no perspective and no ideal place to stand in front of the screen (if there was one it would be at an infinite distance of the screen).
Therefore, any position in front of the screen has virtually the same distortion, and so the picture looks the same no matter where you stand in the room (as long as you look straight at the picture).

In essence : it is similar to what we'd understand as a 2D + depth picture from an orthographic camera (but raytraced in order to fill gaps of traditional 2D + depth)
Should be great for representing 2D environments like maps or for a menu interface or for 2D games but it wouldn't work for gameplay since this is type of camera (ortho) is unusable to navigate a 3D environment.
Actually I think if you were to add depth to the various planes in a 2D game, you'd probably doing something very similar to this technique without realising it.

Thanks for the explanation BlackShark, makes more sense now.