ComfyUI Node: Split 3 Bands

Split 3 Bands:

The SplitThreeBandsNode is designed to decompose an image into three distinct frequency components: low, medium, and high frequencies. This node is particularly useful for image processing tasks where different frequency bands need to be analyzed or manipulated separately. By applying Gaussian blurs with varying radii, the node effectively isolates these frequency bands, allowing you to focus on specific details or textures within an image. This capability is beneficial for tasks such as image enhancement, noise reduction, or artistic effects, where controlling the level of detail is crucial. The node ensures that the frequency components are within a proper value range, making it easier to work with them in subsequent processing steps.

Split 3 Bands Input Parameters:

image

The image parameter is the input image that you want to process. It serves as the source from which the frequency components will be extracted. This parameter is essential as it provides the data that the node will analyze and decompose into different frequency bands.

low_freq_radius

The low_freq_radius parameter determines the radius of the Gaussian blur applied to extract the low-frequency component of the image. A smaller radius will result in a less blurred image, retaining more details, while a larger radius will produce a smoother image by averaging out more details. The minimum value is 1, the maximum is 100, and the default is 5. Adjusting this parameter allows you to control the level of detail retained in the low-frequency band.

medium_freq_radius

The medium_freq_radius parameter specifies the radius of the Gaussian blur used to isolate the medium-frequency component. It must be larger than the low_freq_radius to ensure proper separation of frequency bands. The minimum value is 1, the maximum is 100, and the default is 10. This parameter helps in capturing the mid-level details of the image, which are not as fine as the high-frequency details but not as smooth as the low-frequency ones.

Split 3 Bands Output Parameters:

Low Frequency Image

The Low Frequency Image output represents the smoothed version of the input image, capturing the broad, general features and tones. This output is useful for tasks that require a focus on the overall structure and color distribution without the distraction of fine details.

Medium Frequency Image

The Medium Frequency Image output contains the mid-level details of the image, which are extracted by subtracting the low-frequency component from a slightly less blurred version of the image. This output is ideal for enhancing textures and patterns that are not too fine or too coarse.

High Frequency Image

The High Frequency Image output isolates the fine details and edges of the image by subtracting the low-frequency component from the original image. This output is particularly useful for sharpening and enhancing the intricate details and textures within the image.

Split 3 Bands Usage Tips:

• Ensure that the low_freq_radius is always smaller than the medium_freq_radius to avoid errors and achieve proper frequency separation.
• Experiment with different radius values to find the optimal balance between detail retention and smoothing for your specific image processing task.
• Use the Low Frequency Image for tasks that require a focus on overall color and tone, while the High Frequency Image can be used to enhance fine details and edges.

Split 3 Bands Common Errors and Solutions:

Low frequency radius must be smaller than medium frequency radius

• Explanation: This error occurs when the low_freq_radius is set to a value equal to or greater than the medium_freq_radius, which violates the requirement for proper frequency separation.
• Solution: Adjust the low_freq_radius to be smaller than the medium_freq_radius to ensure correct operation of the node.