Kenneth wrote:

> Could this be implemented as easily as I imagine?

http://graphics.pixar.com has a paper on combining focal blur, 
anti-aliasing, motion blur, specular highlighting, area lights, and 
blurred transparency into one feature.  To improve the quality of any of 
these, one merely increases the anti-aliasing level.

The way POV-Ray handles this now is rather cumbersome.  If I set 4x4 
anti-aliasing, and uses a 4x4 area light, the area light is fully 
sampled for each sample of the spatial anti-aliasing, and if I throw in 
10x oversampling for motion blur, then the scene takes 2560 times as 
long as a scene with no area lights, anti-aliasing, or motion blur.

The Pixar method rolls all of this together, so that the anti-aliasing 
level sets the quality level for all of these features that call for 
extra samples, and the rendering time is based strictly on the highest 
level required, and not the product of all the different quality levels 
multiplied together.  Essentially, extra quality for all features is 
paid for with the increase in quality to any feature.

In the example above, setting the anti-aliasing to 4x4 automatically 
sets the area lights to 4x4 subsampling, and also turns on 16x 
oversampling for the motion blur (if the objects receive the frame_start 
and frame_end settings that are necessary for motion blur), and the 
sampling level only goes up by a factor of 16.  If one of these features 
isn't sampled well enough, increasing the anti-aliasing level raises the 
sampling for that feature (and all others that are turned on for the 
render).  The Pixar docs say that 8x8 anti-aliasing is usually enough to 
take care of all the effects covered by this technique; the 2560x 
oversampling level from the current POV-Ray method would translate to an 
anti-aliasing level of 50x50 per pixel, which is overkill for almost any 
scene.

Regards,
John

Post a reply to this message