wanx-troopers

Wan T2V Advanced

Note: Wan Wiring page contains initial information on constructing Wan VACE workflows, both native and wrapper. This page contains additional/advanced examples.

Ditto is a stylized fine tune of VACE

drozbay has shared the following advanced VACE workflows relying on his ComfyUI-WanVaceAdvanced node pack, with ClownShark and VACE:

Dancer - DWPose Estimator and Depth Anything V2 as 2 separate VACE control videos
Falling Into The Water - Canny Edge limited by depth masking: “keep the outlines, … don’t … details inside the objects”

Note: WanVacePhantomDualV2 from WanVaceAdvanced is compatible with Phantom but not linked to it otherwise. vace_reference and phantom_images on this node are fundamentally different: “they go to different model layers and they are treated differently”.

vace is still pretty iffy when trying to use the reference input for style it’s not a style-transfer model, even if it occasionally does work the reference is really for either background, setting, or subject(s)

it will not be great for character unless your character reference is pretty much exactly where the character will be in the first frame. In that case I recommend you use Phantom

and don’t use the skeleton+depth mix; send in the depth and the skeleton into separate Vace control inputs; either chain WanVaceToVideo nodes, or use WanVacePhantomDualV2 node

Q: load FUN VACE module for High Model first, and then VACE 2.1 for the LOW model? A: sometimes

can try WAN 2.1 + Wan 2.1 to Wan 2.2 LOW LoRA + VACE 2.1 on LOW
original Vace was pretty weak with pose control

Vace + Phantom on Low Noise

Combining Latent, VACE Masks, Phantom, and Context Windows

If you use only Vace for inpainting without a latent noise mask, it samples the entire video, so even the “unchanged” part of the image gets degraded

drozbay has thoughtfully developed a WF for using VACE, latent masks, VACE, Context Windows references together. His considerations were

“You can’t split a generation using a latent noise mask into two samplers natively. For that I modified the RES4LYF code because it already had the infrastructure to be able to add that feature. But you can do it with just one sampler just fine”
“if you want to use Vace reference, because then you need to account for that frame’s noise mask. For that I modified my WanVaceAdvanced node to handle it and I’ll push that soon”
“the comfyui built-in interpolation … when you are taking the “average” of the masks for 4 frames … the resulting masks really fail with fast motion … I made a small node that I will add to WanVaceAdvanced as well to fix this issue”

WIP wf from him: droz-dancing-man

My ref image was just the old man with a white background for both Vace and Phantom. The outpainting was handled by Vace entirely and the prompt

VACE With Qwen Image Edit

TrentHunter82 has shared a workflow that combines Qwen Image Edit with Wan Vace Keyframe Builder from his TrentHunter82/TrentNodes node pack: Keyframe Party-v1

wanx-troopers

Wan T2V Advanced

Vace + Phantom on Low Noise

Combining Latent, VACE Masks, Phantom, and Context Windows

VACE With Qwen Image Edit

See Also