I have a performance-related question. Given an engine and a video adapter can support both, what will be faster (for detailed, AAA-quality characters and props): using limited number of big, unique textures (4 or 8 K resolution), or smaller textures which are combined in modular fashion (detail texture overlays, trim sheets, decals). The first option will need more video memory of course. But the second, while lighter on memory, may need more complex shader instructions, extra UV channels, and extra texture samples. And I do not know which method’s drawbacks are worse for performance