Distance disolving fractal ?

  • 3 Replies
  • 256 Views

0 Members and 1 Guest are viewing this topic.

Offline hamtarodeluxe

  • *
  • Fractal Freshman
  • *
  • Posts: 3
« on: July 20, 2018, 11:14:43 AM »
Hi everyone  ;)
I hope I'm in the right place to ask my question.
Theses days, I'm playing with fractal shapes in raymarching. It belongs to a bigger project and a major feature is to be able to freely fly in and around the shape, without any near clip glitch, and staying the more artistic as possible.
My main problem is how to handle the penetration of shapes.
Alpha fading is expensive and not really raymarching compatible.
Using a raw clipping sphere doesn't look good. A "smooth" substraction operation is really cool on an apollonian fractal, giving a feeling of progressive deconstruction. But this do not always work especially not on Kaleidoscopic IFS fractals which I'm playing with right now. (It reduces the spheres composing the fractal and looks noisy).
A generic manner is to hollow the shapes by taking the absolute value of the distance function. Then, using the clipping sphere technique is also nice, for example on mandelbulb.

I'm trying to find a better method, especially for the Kaleidoscopic IFS fractals (any de based on folding space) , that gives a nice feeling of dissolving shapes to avoid penetration. Any idea about it ? Maybe something less generic, passing a distance to camera parameter as an argument to the distance estimator ? For example, on a truchet towers DE, I managed to reduce the radius as we get closer and it gives a nice effect without any performance issue.

Any thought about it is welcome :)

Offline fractower

  • *
  • Fractal Fanatic
  • ***
  • Posts: 28
« Reply #1 on: July 27, 2018, 01:28:32 AM »
Quote
Alpha fading is expensive and not really raymarching compatible.

Have you considered a speckle fade. The idea would be to have a ray statistically ignore a surface hit based on the distance from the eye. At high enough frame rate it would likely approach Alpha fade... with some chance of induced seizures.

Adding a color fade as well would help to avoid bright flashing pixels.   

Offline hamtarodeluxe

  • *
  • Fractal Freshman
  • *
  • Posts: 3
« Reply #2 on: July 31, 2018, 11:23:11 AM »
Hello :-)
Thank you, I did'nt think of speckle fade in raymarching. I implemented this and the result is cool, but the noisy aspect is to obvious, especially in VR.
But I might use this somewhere else :)

Offline fractower

  • *
  • Fractal Fanatic
  • ***
  • Posts: 28
« Reply #3 on: July 31, 2018, 07:41:43 PM »
Evolution seems to have favored motion detection over scene averaging. I guess all our potential ancestors who would have liked this approach missed seeing the charging tiger at some point.


xx
M-set distance estimation

Started by gerrit on Fractal Mathematics And New Theories

11 Replies
751 Views
Last post July 08, 2019, 07:08:19 PM
by claude
xx
Distance estimator

Started by gerrit on UltraFractal

0 Replies
139 Views
Last post May 14, 2018, 11:18:59 PM
by gerrit
xx
Distance (Logarithm) test

Started by claude on Fractal Image Gallery

4 Replies
305 Views
Last post September 14, 2017, 10:08:09 PM
by claude
clip
Rotations that retain the Minkowski distance.

Started by eiffie on Fractal Mathematics And New Theories

22 Replies
610 Views
Last post May 25, 2019, 02:08:19 AM
by 3DickUlus
xx
burning ship distance estimation

Started by claude on Fractal Image Gallery

14 Replies
699 Views
Last post September 11, 2019, 07:44:43 PM
by claude