Latent Keyframe Node Documentation

Overview

The Latent Keyframe node is part of the ComfyUI-Advanced-ControlNet extension. It is used within the ComfyUI framework to manipulate the strength of ControlNets applied to specific latents, enabling you to precisely control the influence of ControlNets across different elements within a batch.

Functionality

The Latent Keyframe node allows users to define the strength of a ControlNet for specific latents in a batch by using a keyframe system. This means you can apply varying levels of ControlNet influence across different elements within the same processing batch, providing detailed control over the output of the model at each step.

Inputs

• Previous Latent Keyframe (prev_latent_kf) (Optional): This input can be used to chain multiple Latent Keyframes together. When multiple keyframes are interconnected, the one with the corresponding batch_index will take precedence and its strength value will be used.

• Batch Index (batch_index): The index of the latent in the batch to which the ControlNet strength should be applied. This acts as a unique identifier for the latent within the batch.

• Strength (strength): The magnitude of the ControlNet’s influence to be applied to the selected latent. A higher value means a stronger influence from the ControlNet.

Outputs

• Latent Keyframe (LATENT_KF): This output provides a Latent Keyframe that can be connected to other nodes in a workflow to define the strength schedule of a batch’s latents.

Usage in ComfyUI Workflows

In ComfyUI workflows, the Latent Keyframe node is utilized when specific alterations in ControlNet influence are needed for different latents within a single batch. This detailed level of control can be particularly useful in situations where outputs need to be finely tuned across multiple stages or scenes.

• Chaining: Multiple Latent Keyframes can be linked together using the prev_latent_kf input to create a schedule of influences across various batched latents.

• Batch Specific Adjustments: By applying different strengths to specific batch_index, users can achieve diverse effects within a single processing batch, ensuring each latent is manipulated to contribute to the desired outcome.

Special Features and Considerations

• Selective Influence: The node facilitates selective influence on individual latents, which allows for the creation of detailed and varied outputs from the same initial input batch.

• Chain Management: Care should be taken when chaining multiple Latent Keyframes, as overlapping batch_index will result in prioritization of the most recent keyframe in the chain.

• Flexibility in Design: This node allows for flexibility in workflow design, enabling users to experiment with different settings and find the optimal configuration for their use case.

By leveraging the Latent Keyframe node, users can fully explore and exploit the capabilities of ControlNets within their ComfyUI environment, tailoring outcomes with precision and creativity.