ACN_AdvancedControlNetApply Node Documentation

Overview

The ACN_AdvancedControlNetApply node is a powerful feature of the ComfyUI-Advanced-ControlNet repository. This node extends the capabilities of ComfyUI's vanilla Apply Advanced ControlNet node by integrating advanced features for managing and applying ControlNet influences. These enhancements make it particularly valuable for tasks requiring precise control over ControlNet application, such as integrating keyframe scheduling, applying attention masks, and using custom weights.

What This Node Does

The ACN_AdvancedControlNetApply node applies loaded ControlNets to input conditions over a video or image processing workflow. It automatically upgrades any supported ControlNet loaded into it to an advanced version. This feature is essential for leveraging advanced scheduling and attention mask functionalities, enabling fine-grained control over where and how ControlNets intervene in the transformation of inputs.

Inputs Accepted

The ACN_AdvancedControlNetApply node accepts the following inputs:

• positive (required): Input for positive conditioning, determining aspects that should be emphasized.
• negative (required): Input for negative conditioning, defining elements that should be minimized or excluded.
• control_net (required): The loaded ControlNet, which will be automatically converted to its advanced version by this node.
• image (required): The guiding image for ControlNets. If only one image is provided, it is applied to all latents. More images allow for per-latent application, and insufficient image provision will result in recycling images to match latent needs.
• mask_optional (optional): Attention masks to indicate which parts of an image the ControlNet applies to. The strength application varies based on the binary or nonbinary nature of the mask.
• timestep_kf (optional): Timestep keyframes dictating ControlNet effect transitions across sampling steps.
• latent_kf_override (optional): Overrides for latent keyframes, applying uniformly across all time steps unaffected by keyframes.
• weights_override (optional): Overrides for weights across all sampling steps, substituting any defined in timestep keyframes.
• strength: The overall strength of the ControlNet's effect, ranging from no effect (0.0) to full effect (1.0).
• start_percent: The percentage of the sampling step at which ControlNet application begins, delaying keyframe effect activation until reaching this threshold.
• stop_percent: The percentage of the sampling step at which ControlNet application ceases, ceasing keyframe effects post this threshold.

Outputs Produced

• positive: Outputs the positively conditioned data with ControlNets applied, reflecting the desired enhancements.
• negative: Outputs the negatively conditioned data with ControlNets applied, showing the suppression of unwanted features.

How It Might Be Used in ComfyUI Workflows

In ComfyUI workflows, the ACN_AdvancedControlNetApply node offers nuanced control over the application of ControlNets:

1. Scheduling Control: By incorporating timestep keyframes, the node allows for dynamic adjustment of ControlNet application throughout different sampling steps.
2. Multi-Latent Processing: Users handle multiple latents with distinct images or masks, ensuring each latent receives specific attention as intended.
3. Layered Effects: Combining this node's capabilities, such as masks and keyframes, with other ComfyUI features, users can create more sophisticated and layered image transformations.
4. Advanced Integration: Particularly effective when used in combination with the Load Advanced ControlNet Model node, where automatic upgrading and complex scheduling create powerful workflows.

Special Features or Considerations

• Advanced Conversion: The node automatically boosts compatible ControlNets to advanced versions, simplifying the user's task of manually configuring settings.
• Granular Control: Through keyframe scheduling, weights, and masks, users achieve high granularity in determining how and where effects are applied.
• Interoperability: Works seamlessly with previous configurations and additional advanced nodes in the ComfyUI ecosystem, allowing flexibility in designing expansive workflows.
• Attention Masks: The use of non-binary masks enables proportional strength settings for ControlNet effects, allowing precision in targeting certain image areas.

The ACN_AdvancedControlNetApply node is an instrumental tool in comprehensive image and video processing tasks within the ComfyUI framework, offering unmatched control and flexibility.