[Bf-cycles] Motion Blur Implementation
Constantin Rahn
crahn at vrchannel.de
Mon Jun 10 20:52:48 CEST 2013
Gavin,
and take a look at the Mitsuba solution for moving lights, please.
http://www.mitsuba-renderer.org/devblog/?p=786
Happy Blending,
Conz
Am 10.06.2013 20:39, schrieb Matthew Heimlich:
> Gavin,
>
> You may also want to poke around in the Appleseed Render source code,
> as if I recall correctly they have a pretty modern approach to
> deformation blur, and a very similar architecture to Cycles as a whole.
>
>
> On Mon, Jun 10, 2013 at 2:33 PM, Gavin Howard
> <gavin.d.howard at gmail.com <mailto:gavin.d.howard at gmail.com>> wrote:
>
> Hello all,
>
> Brecht, thank you. I will get going on that patch. Stuart, if
> you don't think that's a good idea, let me know. Otherwise, I'll
> use that to focus my code exploration.
>
> Nate, because of the nature of deformation motion blur, it
> will include curved blur. It has to. The reason is because most
> deformation is done by rotation. My current plan is to simply add
> multi-step export to the Cycles addon. (All you smart devs should
> yell at me if that's a bad idea.) I am also going to browse
> LuxRender's source code to see how they did it. Regardless, at the
> end of this summer, not only will Cycles have deformation motion
> blur, it will get multi-step export (steps will be specified by
> the user) and curved motion blur by side effect.
>
> God Bless,
> Gavin Howard
>
> On Jun 10, 2013 9:39 AM, "Nate Wiebe" <natewiebe13 at gmail.com
> <mailto:natewiebe13 at gmail.com>> wrote:
>
> Not entirely related, but will the scope of this GSOC cover
> curved motion blur?
>
> -NateW
>
>
> On Mon, Jun 10, 2013 at 11:29 AM, Brecht Van Lommel
> <brechtvanlommel at pandora.be
> <mailto:brechtvanlommel at pandora.be>> wrote:
>
> Yes Blender supports this, it does scene.frame_set(frame,
> 0.0) now,
> the second value is a subframe value between 0.0 and 1.0.
>
> On Mon, Jun 10, 2013 at 4:58 PM, Gavin Howard
> <gavin.d.howard at gmail.com
> <mailto:gavin.d.howard at gmail.com>> wrote:
> > Hello,
> >
> > Does Blender allow exporting at arbitrary
> subframes? If not, that might
> > be a problem. If so, then Stuart has me exploring the
> code, and I think that
> > I could write the patch while doing so. Does that sound
> okay?
> >
> > God Bless,
> > Gavin Howard
> >
> > On Jun 9, 2013 6:34 AM, "Brecht Van Lommel"
> <brechtvanlommel at pandora.be
> <mailto:brechtvanlommel at pandora.be>>
> > wrote:
> >>
> >> We should support exporting at the exact shutter time.
> At the time I
> >> implemented motion blur evaluation at subframes was
> actually broken
> >> for shape keys and physics systems so I didn't use it,
> but that has
> >> been fixed since.
> >>
> >> However, exporting motion from the frame positions may
> be useful to
> >> keep as it may give more predictable results in some
> cases. When
> >> you're editing animations that's the only thing you
> actually see in
> >> animation playback, so the motion between those frames
> may contain
> >> some unexpected things that you don't want to render.
> >>
> >> I don't know how common that is, did not look at .blend
> files yet to
> >> see how different the results are, if it is an issue we
> should
> >> probably have an option (off by default) to only export
> motion from
> >> integer frame numbers and no subframes.
> >>
> >> On Sat, Jun 8, 2013 at 7:36 PM, Gavin Howard
> <gavin.d.howard at gmail.com <mailto:gavin.d.howard at gmail.com>>
> >> wrote:
> >> > Hello all,
> >> >
> >> > I have a question about implementation. I have
> been told by
> >> > Brecht that currently, the Cycles addon exports three
> steps for motion
> >> > blur to the renderer: one at the previous frame, one
> at the current
> >> > frame, and one at the next frame.
> >> >
> >> > Now, my memory may not be serving me well here,
> so I may have
> >> > gotten the details wrong. But if not, shouldn't the
> addon actually use
> >> > the shutter time to determine the first and third
> steps? For example,
> >> > shouldn't the addon export a step at the start of the
> shutter time,
> >> > one at the current frame, and one at the end of the
> shutter time?
> >> > Also, I think this type of reasoning should apply
> even when multi-step
> >> > (more than 3) motion blur is added. (Multi-step blur
> is needed for my
> >> > Deformation Motion Blur GSoC project, so I will add
> it over the
> >> > summer.)
> >> >
> >> > In short, I think that the shutter time should
> be the goto for
> >> > exporting motion blur steps. If it is, then just
> ignore this. But if
> >> > it's not, should we change it? Anyway, that's my two
> cents.
> >> >
> >> > God Bless,
> >> > Gavin Howard
> >> > _______________________________________________
> >> > Bf-cycles mailing list
> >> > Bf-cycles at blender.org <mailto:Bf-cycles at blender.org>
> >> > http://lists.blender.org/mailman/listinfo/bf-cycles
> >> _______________________________________________
> >> Bf-cycles mailing list
> >> Bf-cycles at blender.org <mailto:Bf-cycles at blender.org>
> >> http://lists.blender.org/mailman/listinfo/bf-cycles
> >
> >
> > _______________________________________________
> > Bf-cycles mailing list
> > Bf-cycles at blender.org <mailto:Bf-cycles at blender.org>
> > http://lists.blender.org/mailman/listinfo/bf-cycles
> >
> _______________________________________________
> Bf-cycles mailing list
> Bf-cycles at blender.org <mailto:Bf-cycles at blender.org>
> http://lists.blender.org/mailman/listinfo/bf-cycles
>
>
>
> _______________________________________________
> Bf-cycles mailing list
> Bf-cycles at blender.org <mailto:Bf-cycles at blender.org>
> http://lists.blender.org/mailman/listinfo/bf-cycles
>
>
> _______________________________________________
> Bf-cycles mailing list
> Bf-cycles at blender.org <mailto: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
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.blender.org/pipermail/bf-cycles/attachments/20130610/4fd46736/attachment.htm
More information about the Bf-cycles
mailing list