[Bf-cycles] Shadow Pass including IBL + mesh Lights (VFX)

Thomas Beck software at plasmasolutions.de
Sun May 11 11:51:40 CEST 2014


As Thomas Dinges pointed out it's just a matter of deleting one check in
kernel_emission.h to support mesh lights in the shadow pass (or is there
more to do, Brecht?). And frankly I don't see any issue with making this
the default... every other PB render engine I know take mesh lights into
account in the shadow pass.

Generally I agree 100% with Matt, a shadow pass should represent all
shadows in the scene - and when we use a HDRI to cast them they should be
included too. I really hope that we have that (and a shadow catcher)
available in the near future as this is crucial for serious compositing in
VFX. It's really a bit awkward that we speed up cycles by 30% and then have
to use a node setup that doubles our rendering time ;) Until then, I'll try
the node setup Brecht posted for simple, non computational expensive scenes.

Greetings, Thomas


2014-05-11 8:35 GMT+02:00 Matthew Heimlich <matt.heimlich at gmail.com>:

> All light sources should contribute to the shadow pass, environment lights
> included. Some environments with HDRI cast a very distinct shadow. A shadow
> catcher would indeed be the best option, though, as it could be given its
> own pass without any additional issues.
>
>
> On Sat, May 10, 2014 at 5:29 PM, Brecht Van Lommel <
> brechtvanlommel at pandora.be> wrote:
>
>> The mesh lights choice was a bit arbitrary and could be changed. For the
>> environment light, if we included that in the current shadow pass you would
>> basically get ambient occlusion, which is not so useful because you have
>> that already in a pass.
>>
>> The shadow pass could be changed to handle this case better, which would
>> be roughly the same as implementing a shadow catcher material.
>>
>> Brecht.
>> On May 10, 2014 8:57 PM, "Greg Zaal" <gregzzmail at gmail.com> wrote:
>>
>>> If I understand Thomas correctly, he's simply asking if we can have mesh
>>> lights contribute to the shadow pass, and perhaps environment lighting as
>>> well if possible.
>>>
>>> Was it a design choice not to have mesh lights included in the shadow
>>> pass? If so, why? Just a performance issue, or is it somehow useful not to
>>> have them? I don't have much experience in compositing renders with GI, so
>>> forgive my ignorance if the answer is obvious.
>>>
>>> Cheers,
>>> Greg
>>>
>>>
>>> On 10 May 2014 00:27, Brecht Van Lommel <brechtvanlommel at pandora.be>wrote:
>>>
>>>> In V-Ray or Mental Ray, is really all you use the shadow pass to get
>>>> this working with IBL? Or do you also use "shadow catcher" or "shadow
>>>> matte" materials? To me it seems what you are really asking for is the
>>>> latter, but maybe I misunderstand how this works.
>>>>
>>>> You can emulate shadow catchers using render layers, but the render
>>>> time is longer than it would be if it was done natively:
>>>> http://www.pasteall.org/blend/29068
>>>>
>>>> It's unlikely that I will work on a shadow catcher material in the
>>>> next months, if someone wants to work on that I can help them, but I
>>>> have other priorities at the moment.
>>>>
>>>> Brecht.
>>>>
>>>> On Fri, May 9, 2014 at 9:44 PM, Thomas Beck <
>>>> software at plasmasolutions.de> wrote:
>>>> > Hi Brecht,
>>>> >
>>>> > as we are creating a lot VFX for the advertising industry and we're
>>>> > converting our projects slowly but steadily to Cycles we're now at a
>>>> point
>>>> > that bothers me and my colleagues a lot: A proper shadow pass that
>>>> includes
>>>> > HDR environment lights and mesh lights.
>>>> >
>>>> > In our production work flow we're taking light probes on the set
>>>> whenever we
>>>> > need to create cg elements afterwards, this way we match the lighting
>>>> > situation perfectly with the real world conditions.
>>>> > When we compose those elements into the shot we're used to use the
>>>> shadow (&
>>>> > AO) pass (in vray / mental ray) and were a bit surprised as we saw
>>>> that
>>>> > we've got no possibility to mimic that behaviour with Cycles and our
>>>> > lighting requirements as mesh and environment lights are not
>>>> considered in
>>>> > cycles shadow pass..
>>>> >
>>>> > Today I asked DingTo on IRC and he told me that there is a way to
>>>> take mesh
>>>> > lights into account by deleting "ls.prim == PRIM_NONE && " in
>>>> > kernel_emission.h
>>>> > That is already a big leap forward but does not help us on our light
>>>> probe
>>>> > (IBL) work flow...so the question is: Do you know that this is not
>>>> only a
>>>> > small missing feature but is really crucial for VFX work? I talked
>>>> with
>>>> > Sebastian too and he's also missing it very much - without this, many
>>>> > projects are very cumbersome to do atm.
>>>> >
>>>> > I searched a bit and was led to the wiki page where you collect all
>>>> todos
>>>> > and especially this:
>>>> >
>>>> > Option to include environment lighting in shadow pass.
>>>> >
>>>> > http://wiki.blender.org/index.php/Dev:2.6/Source/Render/Cycles/ToDo
>>>> >
>>>> > I really don't want to bash here - just raise the awareness of this
>>>> issue.
>>>> > It's not only a missing feature, it's really a show stopper for us.
>>>> >
>>>> > Is that feature hard to implement? Is there any chance you could
>>>> tackle this
>>>> > issue in a relatively short amount of time? Should I write Ton and
>>>> beg for
>>>> > dev time?
>>>> >
>>>> > If you'd like to talk just drop me a message @plasmasolutions or g+
>>>> and I'll
>>>> > be online in IRC...
>>>> >
>>>> > Many greetings, Thomas
>>>> >
>>>> > P.S: I'm happy that I'm not the only one with this request but I still
>>>> > wonder why there are not many more? Is Blender rarely used in the VFX
>>>> field?
>>>> > https://developer.blender.org/T34272
>>>> >
>>>> > --
>>>> >
>>>> > ------
>>>> > Plasmasolutions
>>>> > Design | Development | Training
>>>> >
>>>> > Website: Http://www.plasmasolutions.de
>>>> > Blog: Http://blog.plasmasolutions.de
>>>> >
>>>> >
>>>> > _______________________________________________
>>>> > 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
>>>>
>>>
>>>
>>> _______________________________________________
>>> 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
>>
>>
>
> _______________________________________________
> Bf-cycles mailing list
> Bf-cycles at blender.org
> http://lists.blender.org/mailman/listinfo/bf-cycles
>
>


-- 

------
Plasmasolutions
Design | Development | Training

Website: Http://www.plasmasolutions.de <http://www.plasmasolutions.de/>
Blog: Http://blog.plasmasolutions.de <http://blog.plasmasolutions.de/>

Telefon: 0176 2017 9565
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.blender.org/pipermail/bf-cycles/attachments/20140511/bfd0c842/attachment.htm 


More information about the Bf-cycles mailing list