On 14/08/2012 11:51, clipka wrote:
> Way better.
>
> Still not perfect though, I think.

Just to be clear: I'm not going for perfection, I'm going for speed 
while maintaining believability.

 > Are you using highlights to fake the
> reflection?

yes

 > Note that the default highlight parameterization is not
> energy-conserving; try using the new "albedo" option (e.g. "specular
> albedo 1.0" for a surface that reflects 100% of all incoming light).

let's give that a try :)

 > You
> also need to take into account that your moon is not a full sphere, but
> only some 40% of it.

using specular highlights, the only way to take that into account, is by 
lowering the brightness of the light_source, which I of course do :)

..though I'm not sure, I'm doing it the physically correct way, but it 
looks OK

> Try with true reflection for reference.

I just did :o
Either my moon-"object" (it's not an object, it's pigment on the 
sky_sphere.. btw: I lied, the sky exists of more than just a sky_sphere: 
each cloud layer is a large sphere by itself) is too dim, or my 
light_source was way too bright, or the specular highlights were way wrong.

Probably a combination of all three :)

Anyway, here's the revised version, with a brighter moon-pigment, dimmer 
moon-light and redone specular hightlights.

Low resolution landscape and no area_lights this time though.


And yes, I know it's a very dark image, but what do you expect, it's the 
middle of the night over there! :)
Just for fun, I also tried to adjust the time, while keeping the 
location of the moon, so I'd get a more full moon.

Still have to work on an automatic moon placement.

Post a reply to this message

Attachments:
Download 'clearsky3.jpg' (80 KB) Download 'clearsky4.jpg' (101 KB)

Preview of image 'clearsky3.jpg'

Preview of image 'clearsky4.jpg'