Xycaleth Posted May 9, 2016 Author Posted May 9, 2016 The meaning of the albedo/specular maps in a specular workflow isn't the same as the albedo/specular maps in the traditional workflow though. You can't use one set in the other, so I don't think that's really a good reason for using specular workflow in this case. With regard to metallic workflow, the edge artifacts along material transitions is kind of a deal breaker for me though Archangel35757 likes this
Archangel35757 Posted May 9, 2016 Posted May 9, 2016 The meaning of the albedo/specular maps in a specular workflow isn't the same as the albedo/specular maps in the traditional workflow though. You can't use one set in the other, so I don't think that's really a good reason for using specular workflow in this case. With regard to metallic workflow, the edge artifacts along material transitions is kind of a deal breaker for me though You're right they aren't the same exact maps-- but in that Marmoset tutorial link above. The author goes into his process for converting traditional diffuse/specular maps into a PBR Specular workflow. Perhaps that conversion could be automated in your code for legacy textured characters.
SomaZ Posted May 10, 2016 Posted May 10, 2016 The meaning of the albedo/specular maps in a specular workflow isn't the same as the albedo/specular maps in the traditional workflow though. You can't use one set in the other, so I don't think that's really a good reason for using specular workflow in this case. With regard to metallic workflow, the edge artifacts along material transitions is kind of a deal breaker for me though I'm not sure if that's only an artifact based on the implementation. I never noticed that in Unreal 4 and I only use the metallic workflow. The only thing is that I have to set the specular in the Materieal to 0.5 and everything looks like it should in UE4. (They kind of combine both workflows)
AshuraDX Posted May 10, 2016 Posted May 10, 2016 @@DT85 summoned me I myself prefer the metal/roughness workflow BUT there is one thing the metalroughness workflow fails at - iridescence, which we probably won't need too often anyway I'm not sure if that's only an artifact based on the implementation. I never noticed that in Unreal 4 and I only use the metallic workflow. The only thing is that I have to set the specular in the Materieal to 0.5 and everything looks like it should in UE4. (They kind of combine both workflows)I second this, unless you have a lot of mid values in your metallic map you won't see artifacts around material transitions That specular value you set there is most likely used as a base specular value for dialectric materials. You're right they aren't the same exact maps-- but in that Marmoset tutorial link above. The author goes into his process for converting traditional diffuse/specular maps into a PBR Specular workflow. Perhaps that conversion could be automated in your code for legacy textured characters. a conversion from PBR to legacy textures is easy to perform since you can get everything you need from your bakes and created PBR maps converting PBR metal/rough to pbr spec/gloss is also no problem at all - reverting that is a bit more difficult. But converting legacy textured models to PBR is a pain in the butt, you have to cancel out all lighting information from the image which can be difficult to do without destroying detail or the base color value an automated process will surely cancel out lighter or darker values but may also lose things like dirt or scratches in the process. there also isn't anything that tells your automatic converter which parts of the texture it's looking at are metallic and which are not it simply can't be automated with just the pure legacy maps to work from. If you went ahead and hand painted a metallic map forthe legacy objects it might be possible to create something semi useable. I'll put something together that illustrates the differences between the 3 discussed systems in a minute Edit: my substance material for converting PBR to JKA materials is not yet perfect, it needs more testing and tweaking until I can say I'm happy with how it turned out If anyone wants to try it with his meshes pm me, it works as a smart material in substance painter or as a node for substance designer which also makes it useable in the free substance player Boothand, Raz0r, minilogoguy18 and 9 others like this
Xycaleth Posted May 11, 2016 Author Posted May 11, 2016 Great, thanks for the input everyone I think I'm going to go for metallic workflow since @@AshuraDX says the artifacts on material transitions can be avoided. I think rend2 already has code for metallic workflow - I just need to enable it and make sure it's working properly. SomaZ likes this
AshuraDX Posted May 11, 2016 Posted May 11, 2016 @@Xycaleth compare the rendered images I supplied and you won't find artifacts
minilogoguy18 Posted May 12, 2016 Posted May 12, 2016 While we're on the subject, will Rend2 support the legacy shaders?
AshuraDX Posted May 12, 2016 Posted May 12, 2016 Something I was curious about: can we use object/world space normal maps for our models? I've only seen tangent space maps used with rend2 yet.
AshuraDX Posted May 12, 2016 Posted May 12, 2016 @@ensiformIt saves me from rebaking which isn't hard or anything - just annoying ! take a look at this:https://jkhub.org/albums/bzyoR the difference between the 2 types is the the tangent space map contains relative normal offsets that are added to the face normals of your mesh while the object space normal map completely overwrites the normal information. I can reproject the detail from the worldspace normal map into a new tangent space map for the model with no problem, but I figured asking wouldn't hurt anyone
Archangel35757 Posted May 12, 2016 Posted May 12, 2016 Something I was curious about: can we use object/world space normal maps for our models? I've only seen tangent space maps used with rend2 yet. an option like that would likely need a dropdown option list so the modeler/artist could specify which type of normal map it is, right? Or would different normal map specification be handled in the shader only?
Xycaleth Posted May 12, 2016 Author Posted May 12, 2016 I'd never considered the same normal map would be used for multiple models I'll look into it. It don't think it would be too much work to do. I imagine it would just be a keyword in the shader to say this is an object space map instead of tangent space. Archangel35757 likes this
AshuraDX Posted May 12, 2016 Posted May 12, 2016 I'd never considered the same normal map would be used for multiple models I'll look into it. It don't think it would be too much work to do. I imagine it would just be a keyword in the shader to say this is an object space map instead of tangent space.LODs? they have different a topology
Sovietskiy Partizan Posted May 19, 2016 Posted May 19, 2016 @@ensiformIt saves me from rebaking which isn't hard or anything - just annoying ! take a look at this:https://jkhub.org/albums/bzyoR the difference between the 2 types is the the tangent space map contains relative normal offsets that are added to the face normals of your mesh while the object space normal map completely overwrites the normal information. I can reproject the detail from the worldspace normal map into a new tangent space map for the model with no problem, but I figured asking wouldn't hurt anyone So I've made simple bump map for Almighty_Gir's Revan https://jkhub.org/images/R87aLuz.png and it looks prety flat in-game - https://jkhub.org/images/6mLApPZ.png . So this effect because renderer project it this way or its because of ineffective shader code ?
Xycaleth Posted May 19, 2016 Author Posted May 19, 2016 Did you create a replacement shader which makes use of the normal map?
Sovietskiy Partizan Posted May 19, 2016 Posted May 19, 2016 Did you create a replacement shader which makes use of the normal map?Ofc, I copied Ashura's one. models/players/revan/revan{ cull twosided { map models/players/revan/revan rgbGen lightingdiffuseEntity alphaFunc GE128 } { stage normalmap map models/players/revan/revan_bump normalScale 1 1 } { stage specularmap map models/players/revan/revan_spec specularReflectance 0.5 }}
ensiform Posted May 19, 2016 Posted May 19, 2016 Is it in a .mtr of the same name as the other .shader? And is it in a pk3?
Sovietskiy Partizan Posted May 19, 2016 Posted May 19, 2016 Is it in a .mtr of the same name as the other .shader? And is it in a pk3?Its In .pk3 and shader is revan.shader only. What about .mtr file ? I see it have same code.
ensiform Posted May 19, 2016 Posted May 19, 2016 A .mtr file is the same as .shader except it takes precedence over .shader if it exists to have the new features and .shader be legacy without new features. Tempust85 likes this
eezstreet Posted May 19, 2016 Posted May 19, 2016 So I've made simple bump map for Almighty_Gir's Revan https://jkhub.org/images/R87aLuz.png and it looks prety flat in-game - https://jkhub.org/images/6mLApPZ.png . So this effect because renderer project it this way or its because of ineffective shader code ?Are you using Rend2? Did you set cl_renderer to rend2?
Tempust85 Posted May 19, 2016 Posted May 19, 2016 You use .shader for the vanilla renderer and .mtr for the rend2 renderer. Bascially, it's to make sure your model can be used with both renderers without errors. @@AshuraDX Why is your tangent normal map so borked? I've never had results like that unless I'm flipping channels or have screwed the low poly models normals. I do agree though, object space is better and also more colorful when opening up the texture.
AshuraDX Posted May 19, 2016 Posted May 19, 2016 @@DT85 well by chopping the tricount down by about 30% I did exactly that Tempust85 likes this
Xycaleth Posted May 20, 2016 Author Posted May 20, 2016 I can't say for sure why it's looking flat, but my guess would be because there's not many "real" lights in JKA so you don't get the lights shining across the surface. Try turning your saber on and see if that makes any difference. Also what does your normal map look like? You said you're using a bump map when I hope you mean a normal map
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now