UVs splitted by material, because of shaders? (Realtime)
polycounter lvl 6
Here and there sometimes i have read that people split/separate their uvs by material.
Like:
a Character has a metal-rod, a metal-shield, and wooden armour (what a bad character -.- )
So the uvs would be split like this, Character, Metal, Wood.
Is that really a thing? Is that so because different shaders?
If yes, isnt that obsolet with pbs-shader?
Or is that actually totaly wrong :-D
Like:
a Character has a metal-rod, a metal-shield, and wooden armour (what a bad character -.- )
So the uvs would be split like this, Character, Metal, Wood.
Is that really a thing? Is that so because different shaders?
If yes, isnt that obsolet with pbs-shader?
Or is that actually totaly wrong :-D
Replies
But sometimes its needed. If they change shading during the gameplay.
Or some objects are breakable. If you dont sure ask the Artist why its done that way on that Projects.
But its still a thing if you work for cinematics. Offline rendering still follow those rules.
Cause if you need to override shader attributes per shot you are in big troubles if everything is baked into textures.
You need the seperation per object. Build it like in real world. Every material is an other object.
If there is dirt or blood on objects its getting more complicated but in this case you have multiple textures to switch per shot.
Glad to hear that this is outdated (in most cases) for realtime-production.
So i can go on an merge/split my uvs by a more meaningfull group as i wish.
Greetings from the other side from austria
making seams on hard edge...because it is a hard edge?
For an object like a character it's often ok to have 3-4 drawcalls though. Because you're not going to have many characters on the screen at once. If you were making a large army tactical game (like shogun total war) with hundreds of characters you'd want to have a single drawcall.
Adding an unnecessary extra material is a much larger concern as far as performance goes - it's not just the number of materials that's the issue, switching between materials all over the place can royally fuck with the efficiency of a render pipeline.
It's probably worth mentioning that there isn't always a 1:1 relationship between materials you author and draw calls - graphics programmers/engineers are clever and have worked out ways to mitigate against these issues.
Any given engine (or indeed game) will have specific ways of handling this stuff and they're generally pretty unique.
From the artist's perspective the best thing to do is use the minimum number of materials/textures required to get the visual result you are looking for. If you have a specific shading effect to apply to a specific area of your mesh its probably a sensible idea to separate that part off and give it unique textures - otherwise just worry about texel density and making your UVs not shit.