this is promised experiment with two twisting along perpendicular axes.
there are two white light sources, uv_mapping, camera and pure deformed box

box {
  -1.8 1.8
  rotate Start*<0,45,0>
  rotate Start*<0,0,45>
  rotate Start*<0,20,0>
  deform { twist 8+Twist1 }
  rotate start*<0,0,90>
  deform { twist 16+Twist2 }
  uv_mapping
  texture{pigment{checker color rgb 0 color rgb 1}finish{specular 1}scale 1/8}
  no_shadow
 }

where Start, Twist1 and Twist2 depend of clock
memory used was like previous no more than 200 KB
parsing time was near 0 sec
but tracing time incresed as I expected - becouse there is deform of deform -
this makes 8 minutes pere frame on PII233 - I have idea to speed it up but I
have troubles with implement it - unfortunatly it shows bugs.

I can't promise that this is last twisting sended here ;-)

ABX

Post a reply to this message

Attachments:
Download 'DEFORM.MPG' (396 KB)

I think it would look interesting, if you let the deforms constant, and just
rotate the box ...

Of course, good work so far... I wish you good luck for the speedup-thingies
...
But I think it's just time-consuming another factor between radiosity,
Media, focal-(and other)-blur, HighRes-AntiAliasing, Isosurfaces and what
the hell we all know ...
what's the difference, if one pixel renders 3 mins, or 200 secs ???
(yes 20secs, who cares ;-)

--
Jan Walzer

Post a reply to this message