Purpose of a texture atlas

[Norgus]

I can imagine it would be efficient to texture a group of objects with one image texture vs multiple image textures, but I’m struggling to see the benefit of doing in in this particular use-case.

Surely if we are scaling down the UV to give particular groups of faces solid colours, it would be better to just assign groups of faces to single colour materials?

Is there a benefit, or is the only reason we do it this way in the course to practice the workflow?

[NP5]

No idea about game engines but it implies they do not have plain colours to add, but do it all with images. Hence these tiny colour atlases give several plain colours for infinitesimal image size, memory cost.