ComfyUI Node: Frequency Combination Node

Frequency Combination Node:

The FrequencyCombination node is designed to merge high-frequency and low-frequency components of images, creating a combined image that retains the sharp details from the high-frequency input while preserving the smooth gradients and overall structure from the low-frequency input. This node is particularly useful in image processing tasks where you want to enhance or manipulate specific frequency components of an image, such as in texture synthesis, image restoration, or artistic effects. By leveraging linear light blending, the node ensures a seamless integration of the two frequency components, resulting in a visually appealing and coherent output.

Frequency Combination Node Input Parameters:

high_freq

The high_freq parameter represents the high-frequency component of the image, which typically contains the fine details and sharp edges. This input is crucial for preserving the intricate details in the final combined image. The high-frequency image should be provided as a tensor with dimensions corresponding to (batch, channels, height, width). There are no specific minimum, maximum, or default values for this parameter, but it is essential that the input is a valid image tensor.

low_freq

The low_freq parameter represents the low-frequency component of the image, which generally includes the smooth gradients and overall structure. This input is essential for maintaining the broader visual context and color information in the final combined image. Similar to the high-frequency input, the low-frequency image should be provided as a tensor with dimensions corresponding to (batch, channels, height, width). There are no specific minimum, maximum, or default values for this parameter, but it is important that the input is a valid image tensor.

Frequency Combination Node Output Parameters:

IMAGE

The output parameter IMAGE is the combined image resulting from the merging of the high-frequency and low-frequency components. This output is a tensor with dimensions corresponding to (batch, channels, height, width), and it retains the sharp details from the high-frequency input while preserving the smooth gradients and overall structure from the low-frequency input. The combined image is suitable for further processing or direct use in various image-based applications.

Frequency Combination Node Usage Tips:

• Ensure that both the high-frequency and low-frequency input images have the same dimensions to avoid any shape mismatch errors during the combination process.
• Use this node in conjunction with frequency separation nodes to first decompose an image into its high and low-frequency components before recombining them with the FrequencyCombination node.
• Experiment with different high and low-frequency inputs to achieve various artistic effects or to enhance specific details in your images.

Frequency Combination Node Common Errors and Solutions:

Low frequency image at index {i} does not have 3 channels

• Explanation: This error occurs when the low-frequency image does not have the expected three channels (typically RGB).
• Solution: Ensure that the low-frequency input image has three channels. If your image is in a different format, convert it to RGB before using it as an input.

Shape mismatch between high_freq and low_freq inputs

• Explanation: This error happens when the dimensions of the high-frequency and low-frequency input images do not match.
• Solution: Verify that both input images have the same dimensions (batch size, height, width, and number of channels) before feeding them into the node. Resize or crop the images if necessary to ensure they match.