Forum: Poser - OFFICIAL


Subject: Poser 8 SR1 now available

thinkcooper opened this issue on Sep 22, 2009 · 178 posts


bagginsbill posted Tue, 22 September 2009 at 7:35 PM

OK. You may know some of this, but I'll say it all anyway.

Ray-traced shadows can be set to blur the edges. The amount of blur is chosen to correspond somewhat with how big (physically) a real-world light source would be, as opposed to the tiny point sources we use in Poser. This is applicable to infinite, spot, and point lights (lights with a specific origin or direction), not IBL.

In real life, the amount of blurring is not just because of the size of the light source. It also depends on the distance from the object casting a shadow to where the shadow falls. As the shadow falls farther away, the shadow should get more smeared out. Poser ray-traced blurred shadows do this.

But until now, we had no control over the quality of this simulation. For mildly blurred shadows, this was OK. But for quality and realism, we want some control.

The problem is that to blur a ray-traced shadow, the renderer has to do more than just send one ray from the lit spot to the light source. It has to send multiple rays, and average them together, to figure out if something is fully lit, partially lit/shadowed, or completely shadowed by another object. So how many rays should it send out to do each sample? 

Before you had adjustable Shadow Samples, the value was fixed at 19 samples. This is pretty fast and works OK for some situations, but not others.

Click this render to observe the details. Each prop is identical, but a different distance from the backdrop. The rightmost prop is far enough away that nearly its entire shadow is blurry. Because the blurred shadow is so obvious on that one, we notice the quality of it more, which is not very good. It's grainy, and unrealistic.


Renderosity forum reply notifications are wonky. If I read a follow-up in a thread, but I don't myself reply, then notifications no longer happen AT ALL on that thread. So if I seem to be ignoring a question, that's why. (Updated September 23, 2019)