ComfyUI Node: (deforum) Add Advanced Noise

(deforum) Add Advanced Noise:

The AddAdvancedNoiseNode is designed to introduce various types of advanced noise into images using PyTorch. This node is particularly useful for AI artists looking to add texture, complexity, or specific noise patterns to their images, enhancing the visual appeal or simulating real-world imperfections. By leveraging different noise types, such as wavelet, value, flow, turbulence, and more, you can achieve a wide range of artistic effects. The node's primary goal is to provide a flexible and powerful tool for noise addition, allowing for fine-tuned control over the amount and type of noise applied, thus enabling the creation of unique and visually interesting images.

(deforum) Add Advanced Noise Input Parameters:

images

This parameter expects an image or a batch of images to which the noise will be added. The images should be in a format compatible with PyTorch tensors, typically with dimensions representing batch size, height, width, and channels.

noise_type

This parameter specifies the type of noise to be added to the images. The available options include "wavelet", "value", "flow", "turbulence", "ridged_multifractal", "reaction_diffusion", "voronoi", and "simplex". Each noise type offers a different pattern and texture, allowing for a variety of artistic effects.

amount

This parameter controls the intensity of the noise applied to the images. It is a floating-point value with a default of 0.1, a minimum of 0.0, and a maximum of 100.0. Adjusting this value will increase or decrease the visibility and impact of the noise on the image.

seed

This optional parameter allows you to set a specific seed for the noise generation process. Using the same seed will produce consistent noise patterns across different runs, which can be useful for reproducibility.

kill_rate

This parameter is specific to certain noise types and controls the rate at which certain elements of the noise pattern are "killed" or reduced. It is a floating-point value with a default of 0.06, a minimum of 0.01, and a maximum of 0.1. This parameter is typically adjusted using a slider.

(deforum) Add Advanced Noise Output Parameters:

NOISED_IMAGE

This output parameter provides the image or batch of images with the applied noise. The resulting images will have the same dimensions as the input images but with the added noise patterns, clamped to ensure pixel values remain within a valid range.

NOISE

This output parameter provides the noise pattern that was generated and applied to the images. This can be useful for further analysis or for applying the same noise pattern to other images.

(deforum) Add Advanced Noise Usage Tips:

• Experiment with different noise_type options to discover unique textures and patterns that enhance your images.
• Adjust the amount parameter incrementally to find the perfect balance between subtle and pronounced noise effects.
• Use the seed parameter to ensure consistency in noise patterns across multiple images or projects.
• For noise types that support it, fine-tune the kill_rate parameter to achieve the desired level of detail and complexity in the noise pattern.

(deforum) Add Advanced Noise Common Errors and Solutions:

"RuntimeError: Expected tensor for images parameter"

• Explanation: This error occurs when the input images are not provided in the correct tensor format.
• Solution: Ensure that your input images are converted to PyTorch tensors before passing them to the node.

"ValueError: Invalid noise_type specified"

• Explanation: This error occurs when an unsupported noise type is specified.
• Solution: Check the available options for the noise_type parameter and ensure you are using one of the supported types: "wavelet", "value", "flow", "turbulence", "ridged_multifractal", "reaction_diffusion", "voronoi", or "simplex".

"TypeError: amount must be a float"

• Explanation: This error occurs when the amount parameter is not provided as a floating-point value.
• Solution: Ensure that the amount parameter is specified as a float, within the range of 0.0 to 100.0.

"RuntimeError: Mismatched tensor device"

• Explanation: This error occurs when the noise tensor and the images tensor are on different devices (e.g., one on CPU and the other on GPU).
• Solution: Ensure that both the noise tensor and the images tensor are on the same device by using the .to() method to match their devices.