<div dir="ltr">Thresholding like this is bound to have some pretty nasty failures. Any time you introduce any sort of adaptive criterion (whether spatial, temporal, intensity, whatever) you will start to find failure cases. This one assumes too much, and so it would have a lot of failure cases where it wouldn't refine when it should.<div>
<br></div><div>Welcome to the world of adaptive sampling. It's not an easy problem to solve. :)</div><div><br></div><div>-Mike</div><div><br></div></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Thu, Jun 19, 2014 at 9:22 AM, Antony Riakiotakis <span dir="ltr"><<a href="mailto:kalast@gmail.com" target="_blank">kalast@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">This can fail for too many reasons. Essentially in a global illumination renderer the color you get is not based only on the geometry of the first bounce (which is what the depth value corresponds to) but any other light colliders that may themselves be moving.<br>
<br></div>
<br>_______________________________________________<br>
Bf-cycles mailing list<br>
<a href="mailto:Bf-cycles@blender.org">Bf-cycles@blender.org</a><br>
<a href="http://lists.blender.org/mailman/listinfo/bf-cycles" target="_blank">http://lists.blender.org/mailman/listinfo/bf-cycles</a><br>
<br></blockquote></div><br></div>