"Andy Cocker" <big### [at] mariner9fsnetcouk> wrote...
> "Christoph Hormann" <Chr### [at] schunteretctu-bsde> wrote in
> message news:39F1DD3A.133D3770@schunter.etc.tu-bs.de...
>
> > when you only want camera *movements*, you could simulate them by moving
> all
> > objects in the scene (at least as i understand motion blur).  I you want
> to
> > change the camera angle, this does no more work of course.
>
> Yeah, that's basically what I had to do in my animation. It's a very
> restrictive way of working... I wonder if there's any possibility that
this
> could be changed?

Ideally, camera motion blur would be done differently than object motion
blur.  Object motion blur (as implemented in MegaPov) increases speed while
also increasing memory usage.  It does this by making multiple copies of the
object (one for each time sample).  That way, multiple rays only need to be
traced where they will intersect a motion-blurred object.

With camera motion blur, on the other hand, you would need to trace multiple
rays for every pixel in the image.  Therefore, the way MegaPov does it would
only waste memory and not give any speed advantage.  Therefore, camera
motion blur should be implemented by tracing the scene multiple times and
averaging the images.

-Nathan

Post a reply to this message