FLSampler (BSS):
FLS_SamplerV4, part of the ComfyUI-BSS_FLSampler node pack, is designed to enhance image generation by intelligently adjusting latent variables during the denoising process. This node employs Foveated Latent Sampling (FLS) to analyze changes in latent space across denoising steps, creating a dynamic focus map known as the "Fovea." By selectively applying high-frequency local contrast boosts and injecting stochastic noise into these high-activity areas, the node enhances micro-details, sharpens edges, and controls texture. This results in images with crisp boundaries and rich textures, making it a powerful tool for artists seeking to add depth and detail to their AI-generated artworks.
FLSampler (BSS) Input Parameters:
model
The model parameter specifies the AI model used for image generation. It is crucial as it determines the style and quality of the output image. The model must be compatible with the ComfyUI framework.
seed
The seed parameter is an integer that initializes the random number generator, ensuring reproducibility of results. It ranges from 0 to 18,446,744,073,709,551,615, with a default value of 0. Using the same seed with identical settings will produce the same output, allowing for consistent experimentation.
steps
This integer parameter defines the number of denoising steps, ranging from 1 to 10,000, with a default of 20. More steps generally lead to higher quality images but increase computation time.
cfg
The cfg parameter is a float that controls the guidance scale, influencing how closely the generated image follows the input conditions. It ranges from 0.0 to 100.0, with a default of 8.0. Higher values make the output more faithful to the input conditions but may reduce diversity.
sampler_name
This parameter selects the sampling algorithm from available options in the ComfyUI framework. The choice of sampler can affect the style and quality of the generated image.
scheduler
The scheduler parameter determines the scheduling strategy for the denoising process, affecting the progression and final quality of the image. It must be chosen from the available schedulers in the ComfyUI framework.
positive
This conditioning parameter provides positive guidance to the model, influencing the features and style that should be emphasized in the generated image.
negative
The negative parameter offers negative guidance, helping to suppress unwanted features or styles in the generated image, thus refining the output.
latent_image
This parameter represents the initial latent space representation of the image, serving as the starting point for the denoising process.
denoise
The denoise parameter is a float that controls the amount of noise reduction applied during the denoising process. It ranges from 0.0 to 1.0, with a default of 1.0. Lower values retain more noise, potentially preserving more detail.
fovea_strength
This float parameter, ranging from 0.0 to 10.0 with a default of 3.0, controls the intensity of the foveated adjustments. Higher values increase the focus on high-activity areas, enhancing detail and texture.
FLSampler (BSS) Output Parameters:
latent
The latent output is the final latent space representation after the denoising process. It encapsulates the refined features and details that will be used to generate the final image.
fovea_mask
The fovea_mask output is an image that visualizes the focus areas identified during the FLS process. It highlights the regions where enhancements like contrast boosts and noise injections were applied, providing insight into the node's adjustments.
FLSampler (BSS) Usage Tips:
- Experiment with different seed values to explore a variety of outputs while maintaining reproducibility.
- Adjust the fovea_strength parameter to balance between subtle and pronounced enhancements, depending on the desired level of detail and texture.
- Use a higher number of steps for more detailed and refined images, especially when working with complex scenes or high-resolution outputs.
FLSampler (BSS) Common Errors and Solutions:
Failed to load FLSampler node
- Explanation: This error indicates that the FLSampler node could not be loaded, possibly due to missing dependencies or incorrect installation.
- Solution: Ensure that all required dependencies are installed and that the node is correctly placed within the ComfyUI framework. Check the logs for specific error messages and address any missing components.
Partial load due to error
- Explanation: This warning suggests that the node pack was only partially loaded, which may affect functionality.
- Solution: Review the error logs to identify the specific issue causing the partial load. Resolve any missing or incompatible components and restart the ComfyUI framework to ensure full functionality.