ComfyUI Node: Frequency Separation HSV Node

Frequency Separation HSV Node:

The FrequencySeparationHSV node is designed to decompose an image into its high and low-frequency components using the HSV (Hue, Saturation, Value) color space. This technique is particularly useful in image processing and editing, allowing you to isolate fine details (high frequencies) from broader color and lighting variations (low frequencies). By converting the image to HSV, the node focuses on the Value channel to perform the separation, ensuring that the color information remains intact while processing the luminance. This method is beneficial for tasks such as texture enhancement, noise reduction, and detail preservation, providing a more nuanced control over the image's visual elements.

Frequency Separation HSV Node Input Parameters:

image

The image parameter is the input image that you want to process. It should be a 3-channel RGB image. The node will convert this image to the HSV color space to perform the frequency separation. Ensure that the image has the correct dimensions and color channels to avoid errors during processing.

blur_radius

The blur_radius parameter determines the radius of the Gaussian blur applied to the Value channel of the HSV image. This blur helps in separating the low-frequency components by smoothing out the fine details. A larger blur radius will result in a smoother low-frequency image, while a smaller radius will preserve more details in the high-frequency image. The value should be an odd integer to ensure proper Gaussian blur application. Typical values range from 3 to 21, with a default value of 5.

Frequency Separation HSV Node Output Parameters:

high_freq_result

The high_freq_result parameter is the output tensor containing the high-frequency components of the input image. This output highlights the fine details and textures by subtracting the blurred Value channel from the original Value channel and normalizing the result. The high-frequency image is stacked to match the RGB channels for consistency.

low_freq_result

The low_freq_result parameter is the output tensor containing the low-frequency components of the input image. This output represents the broader color and lighting variations by replacing the original Value channel with the blurred Value channel. The low-frequency image is then converted back to the RGB color space.

Frequency Separation HSV Node Usage Tips:

• To enhance fine details in an image, use a smaller blur_radius to retain more high-frequency information.
• For noise reduction, use a larger blur_radius to smooth out the high-frequency noise while preserving the overall structure of the image.
• Experiment with different blur_radius values to find the optimal balance between detail preservation and noise reduction for your specific image.

Frequency Separation HSV Node Common Errors and Solutions:

Image at index {i} does not have 3 channels

• Explanation: This error occurs when the input image does not have exactly 3 color channels (RGB).
• Solution: Ensure that the input image is a 3-channel RGB image before passing it to the node.

ValueError: Image at index {i} does not have 3 channels

• Explanation: This error indicates that one of the images in the batch does not have the required 3 channels.
• Solution: Check the batch of images to ensure all images are in RGB format with 3 channels.

Invalid blur radius

• Explanation: The blur_radius parameter must be an odd integer.
• Solution: Provide an odd integer value for the blur_radius parameter, such as 3, 5, 7, etc.

High frequency values out of range

• Explanation: The high-frequency values must be clipped to the range [0, 1].
• Solution: Ensure that the high-frequency values are properly normalized and clipped within the range [0, 1] during processing.