Hm. No, I didn't use media_attenuation, but I just tried
it with using it, and to be honest, I see no significant change,
other than tracing slower. I wasn't sure what to expect though,
after all, I'm using scattering media, which absorbs light, so
how should there be much difference? I don't understand the
keyword really, at least not what it implies on using with media
(I do get it if its required for fog), other than doing something which
the media already does itself: swallow some light.

About your shell-idea: I think that'll result with some strange
shadows: where parts of the shell overlap (shadow-wise speaking),
the shadow will be really dark, but when passing right through the
middle of such a bubble, it will be brighter than the shadowrim.
When I have the time, I'll fiddle with it, but I thank for realism's sake
I should stick with the approach I'm using right now. I've gotten used
to it, at least in this image.

But I'll experiment with it using a sphere, scattering media etc, just to
see the effect.

Regards,
Tim

--
Tim Nikias
Homepage: http://www.digitaltwilight.de/no_lights/index.html
Email: Tim### [at] gmxde

> >
> > > Have you used media_attenuation in the light source?
> > Sure, how else would media interact with lights?
>
> media_attenuation, not media_interaction which is already used by default.
> Attenuation so the light falls off with distance through the media. I wasn't
> sure if you had already used both those keywords in the light_source.
> However, now I'm thinking it might affect the entire media object instead of
> only where the clouds appear. I say that because I just checked on that by
> using a cloud animation I was trying out after someone had asked about a
> fly-thru of clouds.
>
> What I was suggesting about the density before would probably go like this:
>
>         [0.50 rgb 0.0]
>         [0.50 rgb 0.25]
>         [0.55 rgb 3.0] // whatever is best for this part, a shell effect
>         [0.60 rgb 0.5]
>         [1.00 rgb 1.0]
>
> But not having tried it I couldn't say it is right.
>
>

Post a reply to this message