[Bf-committers] Cycles-Bake: passes type revamp
dfelinto at gmail.com
Mon Dec 28 13:00:54 CET 2015
I've been giving it more thought, and I would rather hold on the
subsurface proposed change, and focus on getting the main revamp into
Even if we assume the performance would stay the same for diffuse
baking when no subsurface. How to merge the diffuse and subsurface
What is the point of baking Diffuse + Subsurface passes together as
weight only, if we won't be able to get the color to multiply them to?
Considering that you can bake the colored diffuse and subsurface
passes with the combine pass, I think it's fine.
I don't think fixing antialiasing in baking is a 'small thing'. That's
why it has never been tackled. But if I have a chance I will discuss
this with Brecht and see how would he propose addressing this. Maybe
it's simpler than I thought.
2015-12-23 13:14 GMT-02:00 Yury Baranov <cucumberer at gmail.com>:
> Hi. First of all, thank you for all things you did for Blender.
> I don't know if my suggestion is related to your task, but currently Cycles
> baking is missing antialiasing in everything: OSL shaders, normals baking
> and Combined baking, of course. I have switched my texture development
> pipeline from AD software to Blender completely, and this issue is
> harrasing me a lot ATM. It would be great if improving this small thing
> will be in your todo list. Thanks again!
> 2015-12-23 7:11 GMT+03:00 Dalai Felinto <dfelinto at gmail.com>:
>> I've been working in some changes on the way Cycles Baking handle the
>> combined and the light passes. You can find the current design and
>> discussions in the patch thread .
>> Besides what is there presented, Brecht proposed to unify diffuse and
>> subsurface baking. This way, it would be easier to bake the
>> view-independent passes at once.
>> Does anyone have something to say on this? I wonder if the patch as it
>> is would help for PBR materials, and if it makes sense for the users
>> to keep diffuse and subsurface separated.
>>  - https://developer.blender.org/D1674
>> Bf-committers mailing list
>> Bf-committers at blender.org
> Bf-committers mailing list
> Bf-committers at blender.org
More information about the Bf-committers