ComfyUI Node: Split and concatenate sigmas

Split and concatenate sigmas:

The "Split and concatenate sigmas" node is designed to combine two sets of sigma values, which are often used in various sampling and denoising processes in AI art generation. This node allows you to split the first set of sigmas at a specified point and concatenate it with the remaining part of the second set of sigmas. This operation can be particularly useful when you want to merge different sampling strategies or blend different noise levels in a controlled manner. Additionally, the node offers an option to rescale the resulting sigma values to maintain a consistent sum, ensuring that the overall noise level remains balanced. This flexibility makes it a powerful tool for fine-tuning the sampling process and achieving desired artistic effects.

Split and concatenate sigmas Input Parameters:

sigmas_1

This parameter represents the first set of sigma values that you want to split and concatenate. It is a required input and must be provided for the node to function. The sigmas are typically used in the denoising process and influence the noise levels at different stages.

sigmas_2

This parameter represents the second set of sigma values that you want to concatenate with the first set. It is also a required input and must be provided. The combination of sigmas_1 and sigmas_2 allows for blending different noise characteristics.

sigmas_1_until

This integer parameter specifies the index at which to split the first set of sigmas (sigmas_1). The values from the start of sigmas_1 up to this index will be used, and the remaining values will be taken from sigmas_2. The default value is 10, with a minimum of 0 and a maximum of 1000, allowing for fine control over the split point.

rescale_sum

This boolean parameter determines whether the resulting concatenated sigma values should be rescaled to maintain the same sum as the original sigmas_1. If set to True, the sum of the resulting sigmas will be adjusted to match the sum of sigmas_1, ensuring consistent noise levels. The default value is False.

Split and concatenate sigmas Output Parameters:

SIGMAS

The output is a single set of sigma values resulting from the concatenation of the specified portions of sigmas_1 and sigmas_2. If the rescale_sum option is enabled, the output sigmas will be rescaled to maintain the same sum as the original sigmas_1. This output can be used in subsequent nodes for further processing or sampling.

Split and concatenate sigmas Usage Tips:

• To blend different noise levels effectively, experiment with the sigmas_1_until parameter to find the optimal split point that achieves the desired artistic effect.
• Use the rescale_sum option when you want to ensure that the overall noise level remains consistent after concatenation, which can be important for maintaining the quality of the generated art.

Split and concatenate sigmas Common Errors and Solutions:

"IndexError: index out of range"

• Explanation: This error occurs if the sigmas_1_until value is greater than the length of sigmas_1 or if the resulting indices exceed the bounds of the input sigma arrays.
• Solution: Ensure that the sigmas_1_until value is within the valid range of indices for sigmas_1 and sigmas_2.

"TypeError: expected Tensor as input"

• Explanation: This error occurs if the inputs sigmas_1 or sigmas_2 are not provided as tensors.
• Solution: Verify that the inputs are correctly formatted as tensors and are compatible with the node's requirements.

"ValueError: cannot concatenate empty tensors"

• Explanation: This error occurs if one or both of the input sigma sets are empty.
• Solution: Ensure that both sigmas_1 and sigmas_2 contain valid sigma values before attempting to concatenate them.