ComfyUI Node: Blur (mtb)

Blur (mtb):

The Blur (mtb) node is designed to apply a Gaussian blur to an image, which is a common technique used in image processing to reduce noise and detail. This node is particularly useful for AI artists who want to create smooth transitions and soften edges in their images. By leveraging Gaussian filters, the node ensures that the blurring effect is evenly distributed, resulting in a natural and aesthetically pleasing outcome. The main goal of this node is to provide a flexible and efficient way to blur images, enhancing the visual quality and artistic appeal of your work.

Blur (mtb) Input Parameters:

image

The image parameter is the input image that you want to apply the Gaussian blur to. It should be provided in the form of a tensor. This parameter is required for the node to function.

sigmaX

The sigmaX parameter controls the standard deviation of the Gaussian filter in the horizontal direction. A higher value results in a stronger blur effect. The default value is 3.0, with a minimum of 0.0 and a maximum of 200.0. Adjusting this parameter allows you to fine-tune the blurring effect to your preference.

sigmaY

The sigmaY parameter controls the standard deviation of the Gaussian filter in the vertical direction. Similar to sigmaX, a higher value results in a stronger blur effect. The default value is 3.0, with a minimum of 0.0 and a maximum of 200.0. This parameter provides additional control over the blurring effect, allowing for anisotropic blurring if desired.

sigmasX

The sigmasX parameter is an optional list of standard deviations for the Gaussian filter in the horizontal direction, applied to each image in a batch individually. If provided, it overrides the sigmaX parameter. This allows for more granular control over the blurring effect for each image in a batch.

sigmasY

The sigmasY parameter is an optional list of standard deviations for the Gaussian filter in the vertical direction, applied to each image in a batch individually. If provided, it overrides the sigmaY parameter. This allows for more granular control over the blurring effect for each image in a batch.

Blur (mtb) Output Parameters:

IMAGE

The output parameter IMAGE is the blurred version of the input image. This output retains the same dimensions and format as the input image but with the applied Gaussian blur effect. The result is a smoother, less detailed image that can be used for various artistic and processing purposes.

Blur (mtb) Usage Tips:

• To achieve a subtle blurring effect, start with lower values for sigmaX and sigmaY and gradually increase them until you reach the desired level of smoothness.
• For images that require different levels of blurring in different directions, use the sigmasX and sigmasY parameters to specify individual standard deviations for each image in a batch.
• Experiment with different combinations of sigmaX and sigmaY to create unique artistic effects, such as directional blurring.

Blur (mtb) Common Errors and Solutions:

ValueError: SigmasX must have same length as image, sigmasX is <length> but the batch size is <batch_size>

• Explanation: This error occurs when the length of the sigmasX list does not match the number of images in the batch.
• Solution: Ensure that the sigmasX list has the same number of elements as the batch size of the input image.

ValueError: SigmasY must have same length as image, sigmasY is <length> but the batch size is <batch_size>

• Explanation: This error occurs when the length of the sigmasY list does not match the number of images in the batch.
• Solution: Ensure that the sigmasY list has the same number of elements as the batch size of the input image.

TypeError: image must be a torch.Tensor

• Explanation: This error occurs when the input image is not provided as a tensor.
• Solution: Convert your input image to a tensor format before passing it to the node.

RuntimeError: Expected 4-dimensional input for 4-dimensional weight [<dimensions>]

• Explanation: This error occurs when the input image does not have the expected dimensions.
• Solution: Ensure that your input image is in the correct format, typically (batch_size, height, width, channels).