"Kari Kivisalo" <ray### [at] engineercom> wrote in message
news:3C1B2F05.3E150CF6@engineer.com...
>
> The tube doesn't seem to be 36x1 in dimensions though, and my
> previous advice on the fade_distance doesn't apply here unless you
> recalculated for the new geometry.

The "tube" you see is actually the plastic cover of the lamp. Inside it is a
35 units long cylinder with the same diameter so the fade_distance is
roughly 3.

> Is the whole room modelled? Just blank walls are enough for the
> out-of-frame parts and maybe black patches for the window(s).

Yep. White walls, no windows because I thought they wouldn't make a
difference.

> If you add glass to the scene the render time will explode unless it's
> done is two steps.

I've been using this with glass objects since it was discussed here. It's
been a really great help.

But how about reflective surfaces in general? A quick test shows that you
could trace any scene with this method. It's really fast because without
reflections you can use a lower max_trace_level and get the same result.
See the picture. The upper half was done in one trace with max_trace_level 7
(2m37s) and the lower was done in two traces, first save radiosity without
reflections with max_trace_level 4 (48s)and then load it with reflections
and max_trace_level 7 (19s). Difference in time is 1m30s and in quality
unnoticable.

Ari-Matti

Post a reply to this message

Attachments:
Download 'Balls3.jpg' (36 KB)

Preview of image 'Balls3.jpg'