[Bf-cycles] GSoC - Hair shaders for cycles strand

[Girish Ramesh]

Yes, but full path tracing can take up to 22 hours to render a normal hair
coupe. So Marschner's original paper(2003) only accounted for single
scattering and multiple scattering was handled in later papers by him,Zinke
and others using photon mapping and global illumination techniques to try
and reduce the complexity. However, Dual scattering was the first good
approximation which allowed multiple scattering to be handled in real time.
Hope that clears it up. Anyway, I will work on the proposal, will trouble
you if I have any other questions :) Thanks again!

On 21 April 2013 05:02, Brecht Van Lommel <brechtvanlommel at pandora.be>wrote:

> Hi,
>
> On Sat, Apr 20, 2013 at 11:50 PM, Girish Ramesh <rgirish28 at gmail.com>
> wrote:
> > Thanks for the reply. I understand the Dual scattering model is only an
> > approximation but it would be a good addition to the marschner model as
> > marschner only handles single scattering of fibres and not multiple
> > scattering within the hair volume. Dual scattering is hard to integrate
> with
> > the system, but does provide better results for blonde and lighter
> coloured
> > hair.
>
> My understanding is that Dual Scattering gives you similar results to
> Marschner with full path tracing, just that it's a faster
> approximation? It's been a while since I read this paper though.
> Anyway, the fact that it's an approximation is no problem, it's a very
> good one and in used in production by e.g. Disney. So I'm confident
> this would be a great feature to have.
>
> > I am quite experienced with rendering as I have worked on hair models and
> > physically based rendering extensively and will be pursuing my phd in the
> > area in July.    I could take up some of the ideas presented on the ideas
> > page but wanted to tackle hair rendering in particular as it is my
> research
> > area. Do let me know what you think.
>
> Ah, if this is your research area then it's a great fit of course!
> Definitely no need to pick another topic then.
>
> Thanks,
> Brecht.
> _______________________________________________
> Bf-cycles mailing list
> Bf-cycles at blender.org
> http://lists.blender.org/mailman/listinfo/bf-cycles
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.blender.org/pipermail/bf-cycles/attachments/20130421/13c25388/attachment.htm