This ComfyUI workflow is used to test and pick which preprocessors/controlnets will work best for your images. Simply add a image (or single frame) and analyze the results. It uses the auxiliary controlnets, marigold depth, SoftEdgePlus, and threw in a automask node.

*When running the workflow for the first time, It will auto download the models it needs. Currently there is a bad link for the "Zoe_Depth_Anything" model. If you get a giant error. Download the model from HuggingFace here and put it in:

...\ComfyUI\custom_nodes\comfyui_controlnet_aux\ckpts\LiheYoung\Depth-Anything\checkpoints_metric_depth\depth_anything_metric_depth_indoor.pt