Help remind me how to set up smaller maps similar to micro detail maps

DustyShinigami

Hi

Not sure if my head is just full of sleep, or it's because I've still yet to make them for myself, but I'm struggling to process how you'd go about setting up smaller maps similar to the Micro Detail Normal maps for skin pores etc. This was discussed a bit in another thread of mine, but I don't think the exact workflow was covered.

For instance, I've downloaded a model of Dr Robotnik (Eggman - though he'll always be Robotnik to me) from Sonic Unleashed, which I plan on including in a render or two once my project for making the Death Egg Robot is finished. My sketchbook with that can be found here: https://polycount.com/discussion/232465/sketchbook-tim-smith#latest

Like other models I've examined, it includes additional maps like Specular and Reflectivity that are smaller than the base maps.  The base maps are 512x512, but these smaller maps are 128x128, or 256x256.

What's the exact workflow for setting these up again? What's the common process? Do you just add them like you normally would, either painting them manually...? Extracting them from other texture bakes...? And then export them out at a lower resolution...? Or if it's just for a specific area, like on the torso, do you crop it so it's just that area, and then export and apply that?

Thanks

Alex_J

Probably its tiled and then a mask is used to select where it appears.

Only reason I can think that it might be a separate map is in case you wanted to selectively enable high resolution detail for close up shots, but then not require as much resolution in the base textures otherwise.

I wouldn't fuss over all of that unless I was actually optimizing a game. If you just want to make a character, can just include the pore details as part of your regular texture maps.

Fabi_G

In what application do you want to render? As far I know in several applications the default materials already have a detail textures input, at least I remember seeing inputs in standard Unity material, Mamorset material and there is a detail texture material function in Unreals material editor too.

DustyShinigami

@Alex_J
Yeah, but I'd like to learn new things as and where I can, especially if it's something I'm curious about. And if they're optimisation techniques, I'd like to learn them; they're useful to understand and know when to implement. One thing I'll be experimenting with in my current project are LODs.

It looks to be a reflective map for the torso.



Body Texture (512x512):

And then the 128x128 reflective map:


Also, looking at the Robotnik model I have, and I've noticed this with a number of character models I've checked, that they have the UV maps set out like this, in different UV grids:


Isn't this a UDIM setup? I've only used UDIMs once. Although I thought this wasn't common use in games? And I've noticed some model's UVs are splayed out over everything. Why is this done? Wouldn't it be better to have them in a UV grid with other objects?


Actually, I'm guessing this is purely for giving something much more resolution as they look to be for stitches. But again, surely something that small wouldn't be seen clearly anyway...?

DustyShinigami

Fabi_G said:

In what application do you want to render? As far I know in several applications the default materials already have a detail textures input, at least I remember seeing inputs in standard Unity material, Mamorset material and there is a detail texture material function in Unreals material editor too.

It'll be in Unreal.

poopipe

The offsetting of UVs into grid squares is just a convenient way of organising UVs assigned to different materials.  
The big UVs are - as you surmise - a means of increasing texel density 

There is nothing special about UDIMs - it's just a naming convention for a standardised system of managing multiple material assignments in a single mesh.  

That reflectivity map is a tricky one. It definitely does not look like a detail texture. 
Is there a second UV set that fits it?
Is it possible that it is intended to be used as a reflection map (i.e spherically projected) ? 

It's not unusual for eg. ambient occlusion to be a lower resolution than a normal map - you just don't need the resolution a lot of the time

DustyShinigami

poopipe said:

The offsetting of UVs into grid squares is just a convenient way of organising UVs assigned to different materials.  
The big UVs are - as you surmise - a means of increasing texel density 

Ahh, okay. Cool.

poopipe said:

That reflectivity map is a tricky one. It definitely does not look like a detail texture. 
Is there a second UV set that fits it?
Is it possible that it is intended to be used as a reflection map (i.e spherically projected) ? 

It's not unusual for eg. ambient occlusion to be a lower resolution than a normal map - you just don't need the resolution a lot of the time

I'm not sure about the spherical projection. Is there a way of finding out in Maya? As to a second UV set - nope. It looks like there's just the one. I've temporarily moved other UVs out the way and placed the torso/jacket UVs in the first UV slot and then enabled that texture. It's only covering the breast region and arms as can be seen on the model and screenshots. The map is added in the Reflected Color slot for the material in Hypershade.

poopipe

looking at any renders of the asset on whatever site it came from might offer a clue. 

doesn't seem like it's a texture in the conventinal sense.   see what happens if you give it spherical/reflection coordinates in Maya?  

I can't remember how you do that in Maya and explaining how it works is redundant because everything in maya has a stupid name but you're looking for ps2 era style reflection mapping

pior

"But again, surely something that small wouldn't be seen clearly anyway...?"

That's the whole point of using such clever local tiling - by having these little details nice and sharp, it creates the illusion of the whole character being high detail when the camera flashes by (as opposed to using a massive but still relatively blurry texture just to carry these small details).

DustyShinigami

poopipe said:

looking at any renders of the asset on whatever site it came from might offer a clue. 

doesn't seem like it's a texture in the conventinal sense.   see what happens if you give it spherical/reflection coordinates in Maya?  

I can't remember how you do that in Maya and explaining how it works is redundant because everything in maya has a stupid name but you're looking for ps2 era style reflection mapping

There aren't any renders from where I've got it from. People have used rippers and importers for countless games and they upload them to the likes of DeviantArt. Also, the game was released on PS2, but PS3 and Xbox 360 as well, so not sure if the map was used on all versions.

Ultimately though, regardless how you would set it up in Max, Maya, Blender, Unreal etc., what I'm after finding out is how you would go about making something like that. If I wanted to make/add a reflectivity map similar to that, what would I need to do? Would it need to be made from another baked map? Painted manually? And how does a lower resolution work? Surely if the diffuse is 512x512, a simple spec/reflect map would need to be the same...? Otherwise it would be dwarfed and barely show...? Just trying to understand the logic and process behind it all.

DustyShinigami

For another model, and from the same game, there's some similar smaller maps too. One's called lens_01_spc, which is the Specularity map. Easy enough. Applying that, I see that it only shows in a specific region of the light on Robotnik's Eggmobile. So that's answered one of my questions with why they're smaller - they're only made for a specific region/area.

However, there's another map called lens_01_spp. I've no idea what that one is. 'spp'...? I'm guessing that's some sort of glossy/roughness map...?

pior

The 128 texture represents the environment being reflected. How it is mapped to the asset is a whole topic in and of itself.
https://en.wikipedia.org/wiki/Reflection_mapping
On the consoles side of things you can find examples of this all the way to Gran Turismo 1/Ridge Racer Type4.

DustyShinigami

Ah-haa, I see. Hmm. Is that set up in Blender or Maya? So, I take it you just apply that small Reflective map to the body, like I have done, and then you need to set up some specific environment lighting for it to work correctly?

Thanks

pior

No, you don't "apply the reflection map to the body" - you start by look up how reflection mapping (of any kind) is done in the Maya realtime viewport, setup some test scenes accordingly to understand how it works, and then apply what you learned to display this model as desired.
(of course ultimately the reflection texture will be part of the material/shadergraph applied to the model. But my point is, you won't find out how to create the reflective mapping look by starting there, but rather, by researching the effect itself).

In a sense that's why a console like the Switch can last for as long as it did, with gorgeous first party titles : if a gamedev team knows the effect they want to achieve, they can selectively chose a low-end technique that allows them to pull it off within the limitations of the hardware, as opposed to using a dogmatic "current day" way of doing the effect.

Depending on the intent, the reflection map could remain the same on a character for whole duraction of a game, and/or vary per material, and/or vary per environment.

DustyShinigami

pior said:

No, you don't "apply the reflection map to the body" - you start by look up how reflection mapping (of any kind) is done in the Maya realtime viewport, setup some test scenes accordingly to understand how it works, and then apply what you learned to display this model as desired.

I'll definitely look into it, but are you sure it's not applied to the body...? Looking at the naming convention of the textures, it suggests to me that it does...?

'chr_eggman_body_01' and then followed by 'dif', 'ref', and 'spc'.

This one looks more like the reflective environment map, right...?:

pior

It's a 128*128 texture, and its design doesn't seem to correspond to anything related to the mapping of the body. Past such a low resolution threshold, masking would be better off done per-polygon rather than per-pixel. It also looks great when applied as an environment map, without showing too much detail in the reflection (as opposed to what one may want on car body for instance).

There is a 99% chance that this is not a masking texture, but rather a reflection texture. I think you are being too dogmatic about assuming that the naming convention dictates everything - It's the other way around really : the gamedevs designed the effect first, and *then* decided on a naming convention for the various components.


It's also possible that it may be overriden fully by other env maps in the actual game, depending on context. There's no way to know ...
It really is a fantastic model BTW ! A perfect illustration of how low spec game art can look better than something fully sculpted/baked down to a highres texture.

poopipe

I think..

there are two reflection maps applied to the model based on surface.  the smaller one is used on non-shiny, non-metal  surfaces because a blurry reflection makes the surface less mirror like, the larger will be used on chrome or whatever cos that'll look more like a mirror

 I would expect they were using spherical environment coordinates to sample the textures in the shader rather than pulling a reflection map from the environment but that's purely because it was a long time ago and pulling them from the environment is not cheap. Masking per polygon would work well with this theory. 


in short  - i agree with Pior on everything except that I think the reflection textures would be directly applied in the material using hacky UVs

DaveGarza

DustyShinigami said:

Hi

Not sure if my head is just full of sleep, or it's because I've still yet to make them for myself, but I'm struggling to process how you'd go about setting up smaller maps similar to the Micro Detail Normal maps for skin pores etc. This was discussed a bit in another thread of mine, but I don't think the exact workflow was covered.

For instance, I've downloaded a model of Dr Robotnik (Eggman - though he'll always be Robotnik to me) from Sonic Unleashed, which I plan on including in a render or two once my project for making the Death Egg Robot is finished. My sketchbook with that can be found here: https://polycount.com/discussion/232465/sketchbook-tim-smith#latest

Like other models I've examined, it includes additional maps like Specular and Reflectivity that are smaller than the base maps.  The base maps are 512x512, but these smaller maps are 128x128, or 256x256.

What's the exact workflow for setting these up Ask AI APK Mod - an application to chat and communicate between you and Chat GPT through artificial intelligence? What's the common process? Do you just add them like you normally would, either painting them manually...? Extracting them from other texture bakes...? And then export them out at a lower resolution...? Or if it's just for a specific area, like on the torso, do you crop it so it's just that area, and then export and apply that?

Thanks

The general workflow for creating smaller maps like this would involve either painting them manually or extracting them from other texture bakes. Once you have the smaller map, you would export it at a lower resolution than the base maps and then apply it to the appropriate area of the model.

If you're using a painting software, you can create a new document with the desired resolution for your smaller map (e.g., 128x128 or 256x256). Then, you can use various painting techniques to create the desired effect, such as using brushes or stamps to add details like skin pores.

Alternatively, if you're using a baking software, you can extract the smaller map from a larger texture bake. For example, you could bake a high-resolution normal map for the entire model and then use a tool like Substance Painter to extract a smaller normal map specifically for the area with skin pores.

Once you have your smaller map, you can apply it to the appropriate area of the model. This can be done by assigning it to a specific material slot or by using a masking technique to apply it to only certain areas.

DustyShinigami

DaveGarza said:

DustyShinigami said:

Hi

Not sure if my head is just full of sleep, or it's because I've still yet to make them for myself, but I'm struggling to process how you'd go about setting up smaller maps similar to the Micro Detail Normal maps for skin pores etc. This was discussed a bit in another thread of mine, but I don't think the exact workflow was covered.

For instance, I've downloaded a model of Dr Robotnik (Eggman - though he'll always be Robotnik to me) from Sonic Unleashed, which I plan on including in a render or two once my project for making the Death Egg Robot is finished. My sketchbook with that can be found here: https://polycount.com/discussion/232465/sketchbook-tim-smith#latest

Like other models I've examined, it includes additional maps like Specular and Reflectivity that are smaller than the base maps.  The base maps are 512x512, but these smaller maps are 128x128, or 256x256.

What's the exact workflow for setting these up Ask AI APK Mod? What's the common process? Do you just add them like you normally would, either painting them manually...? Extracting them from other texture bakes...? And then export them out at a lower resolution...? Or if it's just for a specific area, like on the torso, do you crop it so it's just that area, and then export and apply that?

Thanks

The general workflow for creating smaller maps like this would involve either painting them manually or extracting them from other texture bakes. Once you have the smaller map, you would export it at a lower resolution than the base maps and then apply it to the appropriate area of the model.

If you're using a painting software, you can create a new document with the desired resolution for your smaller map (e.g., 128x128 or 256x256). Then, you can use various painting techniques to create the desired effect, such as using brushes or stamps to add details like skin pores.

Alternatively, if you're using a baking software, you can extract the smaller map from a larger texture bake. For example, you could bake a high-resolution normal map for the entire model and then use a tool like Substance Painter to extract a smaller normal map specifically for the area with skin pores.

Once you have your smaller map, you can apply it to the appropriate area of the model. This can be done by assigning it to a specific material slot or by using a masking technique to apply it to only certain areas.

Thank you. Thank you for the no fuss, no nonsense response to my question. Good to know and be reassured that it's as simple as that.

Noors

If this is sphere mapping, in your shader you use your mesh normals in view space as input UV for the "reflection" texture.
It's the same thing as zbrush matcaps. It's cheap and easy to do.
https://www.youtube.com/watch?v=osgMB0mf23w
Now it can be hard to tell what map does what without seeing the shader or all UV/vertex color channels