EmptyARVideoLatent - ComfyUI Built-in Node Documentation - ComfyUI

The EmptyARVideoLatent node creates a blank latent representation for video generation. It builds a tensor of zeros using the requested width, height, frame count, and batch size, which can then be used to initialize a video generation process.

Inputs

Parameter	Description	Data Type	Required	Range
`width`	The width of the video frames in pixels (default: 832)	INT	Yes	16 to 8192 (step: 16)
`height`	The height of the video frames in pixels (default: 480)	INT	Yes	16 to 8192 (step: 16)
`length`	The number of frames in the video (default: 81)	INT	Yes	1 to 1024 (step: 4)
`batch_size`	The number of videos to generate in a single batch (default: 1)	INT	Yes	1 to 64

Note: The internal latent size is derived from these inputs. width and height are divided by 8, and the number of latent time steps is calculated as ((length - 1) // 4) + 1.

Outputs

Output Name	Description	Data Type
`LATENT`	A latent tensor filled with zeros, representing an empty video latent space with the specified dimensions, length, and batch size. The tensor shape is [batch_size, 16, lat_t, height/8, width/8], where lat_t = ((length - 1) // 4) + 1 is the number of latent time steps derived from the requested length.	LATENT

This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

Source fingerprint (SHA-256): 02ed3c96d94f2a3df9fb5315a5312e5280b9bee280369eb1218ba10bc122e609