Slime wrote:
> > The caustics are completely different here.
>
> The shape of the glass object is a bit different, too. That's why. Maybe a
> different IOR also.

No, the IOR is the same. As far as the glass goes, just a bit different
frequency shouldn't cause that difference IMO. That being said, don't you
find Lightflow's caustics a bit approximative?

> In your image, if you look in the sphere's reflection, you can see lots of
> radiosity artifacts behind the area that the camera sees.

 Funnily there shouldn't be any radiosity at all ;)
I used precalculated radiosity data from a *dummy* scene without
reflections. Next time I will render the save_file scene with reflections
turned on to avoid the problem.
Besides, the Lightflow version consists of a real box. POV's does not. The
reason is that in Lightflow _no_image_ objects (the wall which you can see
reflected but that theoretically should be visible just in front of the
camera) are taken into account during radiosity calculations. In POV they're
not, so I couldn't use that trick (and placing the camera inside the box
would cause ugly distortions to the spheres due to the wide angle needed).

--
Jonathan.

Post a reply to this message