Phantogram test
by JackDesBwa 2020-01-29 08:29
This is a phantogram test for a post of the 3dphoto.net forum: http://www.3dphoto.net/forum/index.php?topic=12357.0
-- 2019-09-06, LX15 manual sync
by JackDesBwa 2020-01-29 08:29
This is a phantogram test for a post of the 3dphoto.net forum: http://www.3dphoto.net/forum/index.php?topic=12357.0
-- 2019-09-06, LX15 manual sync
Comments
JackDesBwa 2020-04-29 12:31
Thanks for the advice Retroformat and for the associated explanations. I noticed that something was wrong, but I was not able to figure out what it could be on my own.
It sounds counter-intuitive at first, but it totally makes sense. I do not have much time those days, but I keep the idea in my head to test later and check how it improves the phantogram effect.
Retroformat 2020-04-25 09:41
Jack:
In a phantogram projection, there is a very simple test, to determine that you've got the perspective "stretch" done correctly. This image is perfect for testing, as it shows an object with an easily verified geometry: a bucket with presumably a perfectly circular, cylindrical shape. (i.e. I doubt that it is an oval bucket).
You've got the ground correctly aligned to be co-planar with your display surface, which might be a printed paper, or a screen, and which will be laid flat in a horizontal orientation. Now if you think about it, any geometric planar thing, that is parallel to the horizontal surface, must exhibit the same shape on that surface, as it would in real life. So for example, if you had a square or circle on the surface itself (i.e. no stereo disparity anywhere on that square or circle), then the phantogram rendering of that square or circle on the surface will be.... perfectly square and circular. But now, if you elevate your square or circle - keep it parallel to the floor, and only give it stereo separation - it will NOT change shape. It will only shift left or right, depending on if it is the left- or right-eye viewpoint. It will remain a perfect square or circle, regardless of stereo separation or disparity.
So that is the test for this image: are the top and bottom edges of this bucket perfectly circular? I assume so, in real life. But in the phantogram image? The top and bottom edges will be parallel in space, and you know that the real item has a perfectly circular top and bottom, and you know the bottom circle is co-planar with the phantogram display surface. So the bottom edge must be a perfect circle on the display surface, as well as the top edge!
Right now the edges are ovals. You have a little bit of work remaining on this one.
Boris