pouët.net

Random distance field image thread

category: code [glöplog]
Now you know how to properly greet mercury.
added on the 2011-09-23 16:49:26 by xernobyl xernobyl
In order to annoy you guys not too much: All these old spikeball showing screenshots are nothing we are currently working on.
added on the 2011-09-23 17:04:28 by las las
las: that makes me sad :(
added on the 2011-09-23 17:19:01 by xernobyl xernobyl
BB Image
BB Image
added on the 2011-09-24 01:40:49 by unc unc
(O_O)
can anyone please tell me when distance fields where first invented?
added on the 2011-09-24 02:16:30 by rudi rudi
and presumably by whom?
added on the 2011-09-24 02:17:18 by rudi rudi
@unc Cries.

... Has new wallpaper.
added on the 2011-09-24 05:15:54 by Mewler Mewler
but i just put on these pants!
added on the 2011-09-24 05:54:38 by shuffle2 shuffle2
unc: Wow! I love the second one.
added on the 2011-09-24 06:48:35 by chock chock
@unc !! also, /r/ fullhd
added on the 2011-09-24 07:00:16 by provod provod
unc wins the thread.
added on the 2011-09-24 13:02:19 by gloom gloom
rudi: Distance fields are a pretty old thing, I guess really old - given that it's just a scalar field with some constraints.

There are several different methods to extract the surface from the (S)DF e.g. bisection, marching cubes/tetrahedrons, sphere tracing (which is what people here call "raymarching" - just another not properly defined term like "ray casting" (try to find a proper definition! :D)).

John C. Hart
Sphere Tracing: A Geometric Method for the Antialiased Ray Tracing of Implicit Surfaces (1994)

unc: how do you do the blurring/fake DoF?
added on the 2011-09-24 13:05:48 by las las
Quote:
extract the surface

extract/intersect...
added on the 2011-09-24 13:08:10 by las las
unc: can't wait to see it running in realtime.
added on the 2011-09-24 15:36:02 by pista pista
unc: the second shot looks like it was taken with a real camera. The "lens" producing chromatic aberrations.... :)
added on the 2011-09-24 20:11:26 by ara ara
las: ok. thanks. what ive seen from figures and things like that from SDF's is that when the sphere hit's and object, the ray starts on the radius (that is jumps), and then continues the ray along the direction, is that the trick? i just think that iq's NVScene slides where kinda difficult to grasp at fist sight. i thought the red lines where other rays, but i guess its just showing radius-distance and hasn't really anything practial use in the sence of the directions? :P
added on the 2011-09-24 23:10:37 by rudi rudi
Rudi: that's it. You calculate the distance to the closest surface from the current point, then you know it's safe to move the ray by that distance in any direction.

And of course the surface of a sphere is the same distance from the centre in any direction, so you could also say that you calculate the intersection with the nearest surface and a sphere with the centre at the current location, then move the ray to the surface of the sphere. Which is slightly more confusing, but sometimes useful :)
added on the 2011-09-25 00:56:42 by psonice psonice

login