code FEATURE: OpenGL 3 Renderer

[Archangel35757]

I don't know about max but for Softimage you can install a later version of crosswalk than that what shipped with the software to update the plugins.

I doubt Crosswalk 2014 supports XSI ModTool 7.5.

 

Artists who want to bake their maps using FBX are still free to do so. we can write a simple script or app to extract the normal/tangent/binormal data from modern FBX files... and supply that to the renderer.

Edited March 1, 2017 by Archangel35757

[Tempust85]

 

My test. FBX was exported with tangents and binormals checked, SBM was exported with tangent basis checked.

[minilogoguy18]

Now export all of those objects into the game as that's where it matters, the render engine of the 3d software being used is pointless to showcase. Put them up on turn tables next to one another.

[Tempust85]

Hm, good point. I'll export to MD3 as it's easier to export and load in-game. They both display normal maps the same.

[Archangel35757]

Just a quick comparison of tools (not an exhaustive list):

 

[Psyk0Sith]

Hm, good point. I'll export to MD3 as it's easier to export and load in-game. They both display normal maps the same.

 

I thought .glm was the reason why we needed xnormal in the first place, to see if dotXSI>glm yields good results. We want to know if carcass could do the job currently without any modifications.

[SomaZ]

I thought .glm was the reason why we needed xnormal in the first place, to see if dotXSI>glm yields good results. We want to know if carcass could do the job currently without any modifications.

No, actually we need it for every single model which is loaded by the renderer, as far as I understood. But glm has the highest priority imo. Testing tangents with md3 is totally legit.

[Archangel35757]

@@Psyk0Sith - send your resulting baked GLM box and maps to @@SomaZ and he can put up a video of it spinning in their test environment.

[SomaZ]

Let's play a little game. What am I working on?

 

EEEhm, I'm not crazy, no.

 

[Ramikad]

Let's play a little game. What am I working on?

 

EEEhm, I'm not crazy, no.

 

 

Being a total noob, env / cube maps? Reflection? Refraction? Two pounds of awesomeness?

[SomaZ]

Jep, on cubemaps. But they allready worked before. There is something special about this picture. Little hint: I havent changed the shaders.

[Tempust85]

Need moar hints!

[SomaZ]

Next Hint: When it does what its supposed to do, it will be part of the pbr evaluation.

[Tempust85]

.....AO? But then why would you be showing that off in a cubemap... Just tell me already! I must know, it's killing me!

[SomaZ]

Nonono, Last hint: Its part of the IBL evaluation.

[Xycaleth]

Implementing proper support for IBL?

[Tempust85]

....so this?

 

[Psyk0Sith]

@@SomaZ

 

Mind taking a look at this.GLM test?:

https://jkhub.org/topic/5120-feature-opengl-2-renderer/page-16?do=findComment&comment=129065

 

needs shader added tho.

[Archangel35757]

@@minilogoguy18 - I have spent my evening investigating Softimage's support for Tangents and Binormals... if you read all the sub-links in the one link you sent me, then you know it is an "Either" / "Or" deal-- NOT both.  Whichever one you apply first is what is written into the vertexes "COLOR" property... the other parameter is expected to be computed by the cross product of the vertex normal and whichever property (tangent or binormal) you applied first.  Does that make sense?  I verified this with multiple dotXSI 3.5 exports.

 

Also, Crosswalk 2014 does not support XSI Mod Tool 7.5 (...it only supports back to Softimage 2010).

 

@@DT85, @@Psyk0Sith, @@AshuraDX, @@Almightygir - So after chatting with DT I have a question... for normal map baking, is the low-Poly mesh suppose to have no smoothing groups(i.e., a faceted mesh)?  And then you could apply the resulting normal map onto a low-Poly version that does have custom-modified explicit vertex normals?  Please someone explain...

Edited March 2, 2017 by Archangel35757

[SomaZ]

Jep, proper support for IBL. What I did, was manually set the mip-maps for the cubemap. So the correct answer is: convoluting cubemaps.

[AshuraDX]

@@DT85, @@Psyk0Sith, @@AshuraDX, @@Almightygir - So after chatting with DT I have a question... for normal map baking, is the low-Poly mesh suppose to have no smoothing groups(i.e., a faceted mesh)?  And then you could apply the resulting normal map onto a low-Poly version that does have custom-modified explicit vertex normals?  Please someone explain...

You do not want to have no smoothing groups on your low poly mesh. That'd essentially make every edge on the mesh a hard edge, and to bake proper normal maps without artifacts you have to sepperate UV shells along hard edges on the mesh.

 

[minilogoguy18]

Your model generally just has relaxed normals with no hard edges for normal maps to work properly, in most instances you don't use hard edges.

 

It's strange what you say about tangents and binormals when I can export a dotXSI file with them applied and reimport that same model and they are there.

[AshuraDX]

Your model generally just has relaxed normals with no hard edges for normal maps to work properly, in most instances you don't use hard edges.

That's not correct either, not having any sort of hard edges or additional geoemtry in your mesh produces a wild amnout of gradients in your output normal maps, which tend to be more difficult to get rendered properly in different Engines.

[Almightygir]

Let's play a little game. What am I working on?

 

EEEhm, I'm not crazy, no.

 

 

You're either working on mip chains, or importance sampling. either of which are crucial to getting PBR IBL correct.

 

good job

 

 

Edit: oops, that was posted yesterday i guess... either way, i'm right >;]

 

Manually setting the mip-level is kinda what you have to do. it's kinda like... you importance sample across the hemisphere using a crazy mathematical equation to figure out where those samples happen, and the mip level you sample is usually just a divisor of the 8 mip levels by the gloss input.

[Archangel35757]

Your model generally just has relaxed normals with no hard edges for normal maps to work properly, in most instances you don't use hard edges.

 

It's strange what you say about tangents and binormals when I can export a dotXSI file with them applied and reimport that same model and they are there.

It's either tangent or binormals-- not both. If you will look at the dot XSI file in a text editor you will see that it is one set of vectors or the other not both... even though it will erroneously create both xsi_customPset templates. However, the color data represents which ever property you set first.