[Bf-committers] Re: [Bf-blender-cvs] Re: CVS commit:
blender/source/blender/blenkernel/intern ipo.c node.c
DNA_camera_types.h DNA_node_types.h blender/s
trip0o at gmail.com
Fri Dec 22 21:53:22 CET 2006
Not to nit pick, but have you thought of using hardware to compute the
dof or fake dof?
In no way do I mean to use pixel shaders. I mean instead that the
computation be sent to the hardware gpu to compute. I have lots of
docs on it and examples of its use.
That way it is exactly the same from any computer, just much faster.
It works for Audio mixxing apps
On 12/22/06, GSR <gsr.b3d at infernal-iceberg.com> wrote:
> trip0o at gmail.com (2006-12-21 at 2254.01 -0500):
> > GSR that dof blur looks realy good. The smoothest I have seen yet with
> > blender.
> Yeah, it is nice.
> I wonder if the "brush" method could be done in Blender like in
> http://sudakyo.hp.infoseek.co.jp/gimp/fblur/focusblur_e.html as that
> would allow real bokeh, the one that depends in blade shape and glass
> distortions too http://www.kenrockwell.com/tech/bokeh.htm
> Using fg and bg separation, it would even allow for different look as
> in real camera http://www.luminous-landscape.com/essays/bokeh.shtml
> > On 12/21/06, GSR <gsr.b3d at infernal-iceberg.com> wrote:
> > >Hi,
> > >ton at blender.org (2006-12-21 at 1911.07 +0100):
> > >> http://www.blender3d.org/cms/Composite__Defocus.836.0.html
> > >
> > >The DoF distance should be in the node as you do not need a camera to
> > >have zbuffer data, and anyway jumping between 3d and nodes is rather
> > >slow.
> I read the code a bit more and found that lens is also used, same
> reasoning, image can come from other place, cameras can change and
> swapping to separate contexts is not fast.
> Bf-committers mailing list
> Bf-committers at projects.blender.org
More information about the Bf-committers