Confusion over UV mapping with texel denisty in mind
null
I'm currently very confused about two different approaches I keep seeing regarding UV mapping and texel density. I suspect it may have to do with differences in workflows.
Most videos I watch regarding UV unwraps usually result in the person completely unwrapping an object and then scaling the UV shells to make the texel density uniform but not to a certain scale. They are not setting the density to a floor reference (say 400cm = 2048x2048). I am however under the impression that texel density should be consistent all the way through from ground plane to objects and architecture. I understand that there will be some fudging but figure it probably shouldn't deviate from more than 5% or so.
Does all of this depend on the workflow? If you UV map objects to a precise texel density then it makes packing UV into 0-1 much harder and can lead to the need of stacking UVs. This however becomes an issue when using Unreal and creating lightmaps correct? What about if I am also using Substance Painter to bake high poly into a low poly?
Is this also affected by what you're modelling? If you are creating a modular kit would you approach this differently to if you were modelling a small game asset.
Another questions I have is if the texel density required is 400cm = 2048x2048 but I am UV mapping a small object would I use the same 5.12 tx/cm scale with a 512x512 texture (100cm = 512x512) so as to not have massive unused space in the texture?
I apologize for throwing so many questions out at once but I've been running in circles trying to figure this out!
PS I should clarify that the workflow I am using is Maya , Substance Painter/Designer , UE4
Most videos I watch regarding UV unwraps usually result in the person completely unwrapping an object and then scaling the UV shells to make the texel density uniform but not to a certain scale. They are not setting the density to a floor reference (say 400cm = 2048x2048). I am however under the impression that texel density should be consistent all the way through from ground plane to objects and architecture. I understand that there will be some fudging but figure it probably shouldn't deviate from more than 5% or so.
Does all of this depend on the workflow? If you UV map objects to a precise texel density then it makes packing UV into 0-1 much harder and can lead to the need of stacking UVs. This however becomes an issue when using Unreal and creating lightmaps correct? What about if I am also using Substance Painter to bake high poly into a low poly?
Is this also affected by what you're modelling? If you are creating a modular kit would you approach this differently to if you were modelling a small game asset.
Another questions I have is if the texel density required is 400cm = 2048x2048 but I am UV mapping a small object would I use the same 5.12 tx/cm scale with a 512x512 texture (100cm = 512x512) so as to not have massive unused space in the texture?
I apologize for throwing so many questions out at once but I've been running in circles trying to figure this out!
PS I should clarify that the workflow I am using is Maya , Substance Painter/Designer , UE4
Replies
Stacking UVs is fine, because in unreal you'll generally have a second UV set for the lightmaps. When you're baking maps with stacked UVs, you offset the stacked parts outside of the primary 0-1 UV space, and they'll be ignored by the baker, but still receive final textures.
And yes, if your chosen texel density for a particular object only uses a quarter of a 2k map, you definitely want to use 512 instead. If you're using Maya, you've probably found the texel density widget in the UV tools.
I was thinking that if I invest more time in creating texture atlases for modular assets or assets with repeating features I will be able to better maximize my texel density as well.
I just came across a video that did a decent job at explaining the guideline vs rule part of my question as well for anyone else looking for reference. The meat of it starts at 12:55 - https://www.youtube.com/watch?v=OJoIb8KsXWk
If everyone sticks to the same density a team can be pretty certain everything will look similar in terms of fidelity and that if they're sharing texture resources (eg. a tiling brick texture) then they can be sure there's not wild inconsistencies in scale between assets that use it.