[Bf-cycles] Future of the non-progressive renderer and Cycles strand rendering

[Brecht Van Lommel]

This paper indeed looks like it would better starting point, mainly
because it includes support for all terms R/TT/TRT instead of just R.
The Hery paper indeed seems to have a slight error but I wouldn't call
it unstable, the difference is quite difficult to see (and Pixar used
it successfully for a year ...). There is also code in the
supplemental material so it should be easy to implement too.

Brecht.

On Mon, Jan 7, 2013 at 6:56 PM, Matthew Heimlich
<matt.heimlich at gmail.com> wrote:
> Brecht and Stu,
>
> I've done some poking around with render-guru buddies and the
> consensus seems to be that the Pixar paper has some inherent numerical
> problems that make it unstable. A couple of people pointed me to the
> paper at http://www.cs.dartmouth.edu/~ouj/site/Research/Entries/2012/6/21_ISHair__Importance_Sampling_for_Hair_Scattering.html
> instead. Thoughts?
>
> On Sun, Jan 6, 2013 at 9:39 AM, Brecht Van Lommel
> <brechtvanlommel at pandora.be> wrote:
>> I don't have specific plans for when such features would be done. I'll
>> try to help solving the limitations and assist Stuart, but for now I
>> plan to follow my original roadmap and work on the planned 2.66
>> features and then subsurface scattering for 2.67 probably.
>>
>> Mainly we need (in no particular order):
>> * Improved attribute support
>> * Preview/render resolution
>> * Curve/ribbon primitive
>> * Minimum pixel width
>> * Transparency optimizations
>> * Memory usage optimization
>> * Hair BSDF
>> * GPU support
>>
>> Maybe Stuart has more specific plans for what he wants to work on first.
>>
>> Brecht.
>>
>> On Sun, Jan 6, 2013 at 3:37 AM, Matthew Heimlich
>> <matt.heimlich at gmail.com> wrote:
>>> Sorry for the delayed response, but I'm interested, what is currently
>>> on the table as far as hair development goes? For the near future I
>>> mean. I'm very pleased with the surprise reveal of a strand patch, and
>>> would like to know what the roadmap for such an implementation looks
>>> like between where the patch started and an 'end' point where features
>>> like light cache and automatic alpha blending are included.
>> _______________________________________________
>> Bf-cycles mailing list
>> Bf-cycles at blender.org
>> http://lists.blender.org/mailman/listinfo/bf-cycles
> _______________________________________________
> Bf-cycles mailing list
> Bf-cycles at blender.org
> http://lists.blender.org/mailman/listinfo/bf-cycles