"And" <49341109@ntnu.edu.tw> wrote:
> "And" <49341109@ntnu.edu.tw> wrote:
> > Another concept diagram, the principle of this little program. Each pixel on the
> > output image generates a "primary path" to calculate the scattering sun light
> > with many segments. then return the sum of them. the variable
> > 'sky_light_primary_path_res' in the .pov is control the resolution of it.
> >
> > If the 'calc_secondary_scatter' turn on, it will sampling secondary paths on
> > every segment of the primary path. then add the effect "sunlight --> skylight
> > --> skylight". It will much slower when you turn it on. But for a dense/thick
> > atmosphere it will make some difference.
>
> This is the demo. The left is only the primary scattering, it takes 34s to
> render. the right is calculate the primary + secondary scattering, it takes
> 18m33s to render.

...., I must say, if you want to use it for the environment(lighting), it is not
very well yet. First of all, what's the real extinction coefficient of our
atmosphere. I don't know that very well. The current value is just my guess.
Secondly, the sampling method of the secondary scatter which I implement is not
very good. But you can turn off the ground reflection by setting 'ground_albedo'
zero. It will eliminate the most obvious artifact. Finally, I should calculate
the output color of the sun again(in the output text file).

This image is the comparison of the ground_albedo 0.2 and 0.0. You can see the
artifact disappear.

Post a reply to this message

Attachments:
Download 'demo3.png' (58 KB)

Preview of image 'demo3.png'