You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, shadow is produced by ray marching from the light (or rather: sufficiently far outside of the fractal) to the surface point to check if the path is blocked. This approach has several shortcomings (right now):
There are many artifacts, probably due to a rather unprecise DE
The current implementation has quite a few magic constants for fine tuning several values. Reducing the number of parameters would be a good idea.
It's slow: even with at most 10 iterations, the framerate on my slow notebook GPU drops below 60. Querying the DE a bunch of times in the shader is not the best idea, I guess...
I see three possibilities:
improve the current algorithm
reverse the direction of the ray: start at the surface point and march towards the light
scratch ray marching altogether and use a simple shadow map approach (which might be extremely complicated when "zooming in")
The text was updated successfully, but these errors were encountered:
Currently, shadow is produced by ray marching from the light (or rather: sufficiently far outside of the fractal) to the surface point to check if the path is blocked. This approach has several shortcomings (right now):
I see three possibilities:
The text was updated successfully, but these errors were encountered: