Looking for high-end Product Photography workflow advice: 3D Renders + Texture Ingestion + Custom Studio Lighting (Intern wanting to blow my boss away!)
Hey everyone,
I’m currently doing my internship and working on setting up an AI-assisted product photography pipeline for high-end furniture. I really want to deliver the best of the best and blow my boss away with what ComfyUI can do, so I’m reaching out to see if anyone has a battle-tested workflow or advice for this specific setup.
Here is the exact pipeline I’m aiming to build:
- Exact Geometry Retention: Starting with 4 Blender renders/passes of the furniture piece so the 3D form, proportions, and edges are 100% accurate (ControlNet Depth + Canny/Lineart).
- Multi-Texture Ingestion via Image Reference: Injecting real photo references for specific materials (e.g., a specific high-res wood grain for the body, leather texture for details/handles). I assume masked IP-Adapter or regional conditioning is the way to go here?
- Custom Studio Setting & Lighting: Placing it in a minimalist neutral grey studio setting, but with our own signature dramatic light and shadow style (considering IC-Light, specific ControlNets, or Style LoRAs).
- Ultra-High-Res Output: Multi-pass rendering / ultimate upscaler to get crisp micro-textures without plastic AI artifacts.
My questions for the experts here:
- Does anyone have a .json workflow template or a similar node architecture they’d be willing to share as a starting point?
- What’s the current gold standard stack for this? (Flux vs. SDXL for texture fidelity + IC-Light vs. regional IP-Adapter?)
- Any crucial custom nodes / techniques I shouldn’t overlook for keeping industrial product accuracy intact?
Any workflow links, node suggestions, or tips would be massively appreciated. Thank you in advance for helping me out!