[Bf-gamedev] Cycles Baking Development

metalliandy metalliandy666 at googlemail.com
Wed Mar 26 15:15:50 CET 2014 

Hi Fredrik,

If you read my reply again you will notice that I never actually said 
anything about removing OS normal maps. WS and OS normal maps are not 
the same thing. ;)

The reason I suggested removing WS normal maps is that they are 
virtually identical to OS with the exception that they need to remain in 
the same position and can't be rotated etc. and as such they are the 
least flexible normal maps of all. OS normal maps would serve exactly 
the same purpose with less code to maintain in the future.

Handplane actually requires Object space and not World Space btw. so you 
don't really want to be using WS normal maps.
I'm not sure what you mean by less errors really. Baking to Object space 
does nothing to reduce errors in bakes, it just ignore the crude 
smoothing of the low poly mesh which when set up incorrectly for Tangent 
space can give you seams etc.

Cheers,

-Andy

On 26/03/2014 09:14, Fredrik hansson wrote:
> i was wondering other than using the cage as a blocker mesh(great 
> feature to have btw) why not use shapekeys instead if you are supposed 
> to have the same topology and control the ray direction with those.
> easier to use imo and less destructive since you could change your 
> underlying mesh and not have to redo the cage every time.
>
> secondly removing object/worldspace normalmaps isn't actually a good 
> idea sure everyone uses tangetspace normals but if you use a tool like 
> handplane for example then you actaully always want to bake worldspace 
> normalmaps since that generally leads to less errors in the bake and 
> can easily be transformed into a tangentspace normalmap for your 
> target game engine(provided that the tool supports that engine). it 
> also allows for topology changes without rebaking well as long as you 
> don't change the uvs that is.
>
>
> On Thursday, March 20, 2014 6:34 PM, metalliandy 
> <metalliandy666 at googlemail.com> wrote:
> Awesome work, Dalai.
>
> I have spoken to you about this on IM but I figured it would be good 
> to get this on the list too, so please forgive any overlap between the 
> list and our conversations. :)
>
>  1. Any ideas when we are going to get to see the cages? This
>     shouldn't ship with Blender until we can see how far the cages are
>     extruding. The cage feature should definetly be a modifier as it
>     allows the flexibility to modify the cage and other gui such as
>     picking the colours of the cage with perhaps some sliders for
>     offsetting everything easily.
>  2. AA is very important, but not as important as being able to view
>     the cage in real time. A good default for AA would be 4xAA.
>     Anything more isnt worth the extra rendering time and anything
>     less doesnt match up well enough in terms of quality.
>  3. The Custom cage title is not correct. Custom cages are when the
>     user has modified the topology of the cage by moving
>     verts/topology and meshes under this slot do not need to have
>     modified verts/topology in order to work. They could use a regular
>     mesh or have one imported from another application so just calling
>     it Cage mesh or Cage object would be perfect. Calling it cage mesh
>     goes against convention for all other 3d applications so it's just
>     going to cause confusion.
>  4. Tangent space should be default for normal maps. I would even
>     think about removing the option to bake world space as no one uses
>     them ever in production as they are the least flexible of all
>     normal maps.
>  5. X+Y+Z+ should be default and is what is being baked currently by
>     default even though the GUI says X-Y-Z- (for TS nmaps at least.
>     IIRC OS and WS bake an odd order that doesnt match any convention.
>     Something like X+Z-Y+ of the top of my head.)
>
>
> Cheers,
>
> -Andy
>
> On 20/03/2014 03:08, Dalai Felinto wrote:
>> Hi all,
>>
>> Some update:
>>
>> The *bake-cycles* branch is near feature complete for its first round.
>> The only missing bit is anti-alias which I'm tempted to post-pone
>> until after the first round of code-review.
>>
>> Recent/New features:
>> * User Interface (no more need to fiddle with Python scripts)
>> * File Formats (save as PNG, JPG, OpenEXR, Tiff, Targa and BMP -
>> RGB/RGBA - 8/16/32)
>> * Custom Cage
>> * Normal swizzle (change the axis that gets mapped to RGB)
>> * Normal space (save as tangent, world, object)
>>
>> As usual the code is here:
>> *https://github.com/dfelinto/blender-git/tree/bake-cycles
>>
>> The Blender-Artist thread [1] is still getting a lot of attention, but
>> I would rather if testers could use the github issue list to report
>> bugs:
>>
>> https://github.com/dfelinto/blender-git/issues?state=open
>>
>> I'll try to do some code cleaning up and address some of the known
>> issues in the remaining of the week. Hopefully the code can be
>> available for code-reviewers by next Blender Sunday Dev Meeting.
>>
>> [1] -http://blenderartists.org/forum/showthread.php?326534
>>
>> Cheers,
>> Dalai
>> --
>> blendernetwork.org/dalai-felinto
>> www.dalaifelinto.com  <http://www.dalaifelinto.com/>
>>
>>
>> 2014-02-17 23:41 GMT-03:00 Dalai Felinto<dfelinto at gmail.com>  <mailto:dfelinto at gmail.com>:
>>> Hi there.
>>> I just pushed support for "Selected to Active" in the bake-cycles
>>> branch. The new parameters for the operator are:
>>>
>>> * use_selected_to_active=False
>>> * cage_extrusion=0.0
>>>
>>> The idea is to bake from a high-poly mesh to a low-poly mesh.
>>>
>>> Notes:
>>> * There is no pre-visualization of the cage
>>> * The cage always look inwards
>>> * The cage cannot be 'edited' (all you can change is the cage extrusion)
>>> * The calculation is done in the local space of the objects (as if the
>>> objects had scale 1,1,1, and were in origin with no rotation)
>>> * I tested against XNormals (managed to install "here") and the result
>>> seem to be matching, though the color is coded slightly different for
>>> object normal maps:
>>>
>>> Cycles RGB = XNormal R(1-B)G
>>> XNormal RGB = Cycles RB(1-G)
>>>
>>> Cheers,
>>> Dalai
>>> --
>>> blendernetwork.org/dalai-felinto
>>> www.dalaifelinto.com  <http://www.dalaifelinto.com/>
>> _______________________________________________
>> Bf-gamedev mailing list
>> Bf-gamedev at blender.org  <mailto:Bf-gamedev at blender.org>
>> http://lists.blender.org/mailman/listinfo/bf-gamedev
>>
>
>
> _______________________________________________
> Bf-gamedev mailing list
> Bf-gamedev at blender.org <mailto:Bf-gamedev at blender.org>
> http://lists.blender.org/mailman/listinfo/bf-gamedev
>
>
>
>
> _______________________________________________
> Bf-gamedev mailing list
> Bf-gamedev at blender.org
> http://lists.blender.org/mailman/listinfo/bf-gamedev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.blender.org/pipermail/bf-gamedev/attachments/20140326/abc4e16c/attachment.htm

More information about the Bf-gamedev mailing list