ComfyUI question: can premade Blender depth maps skip the ControlNet preprocessor?

DiddlyDinq · reddit · 2026-09-28

A Reddit beginner question: ComfyUI ControlNet tutorials all use Load Image → Preprocessor → Apply ControlNet, but the poster wants to use premade constraint images (depth maps, masks etc.) from Blender.

He wonders whether the Apply ControlNet node actually knows what type of image it receives — if there's no hidden metadata, he assumes he can ditch the preprocessor entirely and go straight from Load Image to Apply ControlNet, and asks if that's correct.

Original post →

More from Multimodal

Multimodal channel →