ComfyUI Node: ◎ Radiance VAE Decode Pro

◎ Radiance VAE Decode Pro:

RadianceVAEDecode is a sophisticated node designed to transform VAE (Variational Autoencoder) latents back into high-resolution images, specifically up to 4K quality. This node is essential for AI artists who work with encoded image data and need to reconstruct the original visuals with high fidelity. It supports various color spaces and HDR (High Dynamic Range) modes, ensuring that the decoded images maintain their intended color accuracy and dynamic range. The node is particularly beneficial for workflows that require precise color transformations and high-quality image resampling, making it a valuable tool for both testing VAE reconstruction quality and preparing reference frames. By handling complex color space conversions and HDR adjustments, RadianceVAEDecode ensures that the output images are as close to the original as possible, preserving the artistic intent and visual details.

◎ Radiance VAE Decode Pro Input Parameters:

pixels

This parameter represents the input image data in the form of VAE latents. It is crucial for the node to have this data to perform the decoding process. The quality and resolution of the input latents directly impact the fidelity of the decoded image.

vae

The VAE model used for decoding the latents back into images. This parameter is essential as it defines the specific VAE architecture and weights that will be applied during the decoding process, influencing the final output quality.

source_space

This parameter specifies the color space of the input latents. It is important for ensuring that the color transformations during decoding are accurate. Options include various log and linear spaces, with "Linear" being the default. The choice of source space affects how colors are interpreted and transformed.

target_space

Defines the desired color space for the output image. This parameter ensures that the decoded image matches the intended color profile, which is crucial for maintaining visual consistency across different devices and platforms. The default is "Linear," but other options are available depending on the workflow requirements.

tile_size

Determines the size of the tiles used during the decoding process. Options range from "Auto" to specific sizes like "512," "768," "1024," "1280," and "1536," with "Auto" being the default. This parameter affects the processing speed and memory usage, with larger tiles potentially improving performance but requiring more VRAM.

overlap

Specifies the amount of overlap between tiles during decoding, with a default value of 128 and a range from 32 to 256. This parameter helps to minimize artifacts at tile boundaries, ensuring a seamless final image.

exposure

Adjusts the exposure level of the decoded image, with a default of 0.0 and a range from -10.0 to 10.0. This parameter allows for fine-tuning the brightness and contrast of the output, which can be crucial for achieving the desired visual effect.

◎ Radiance VAE Decode Pro Output Parameters:

img

The primary output of the node, representing the decoded image. This image is reconstructed from the VAE latents and should closely match the original input image in terms of resolution and color fidelity.

dec_meta

Metadata associated with the decoding process, providing information about the parameters and settings used. This output is useful for tracking and reproducing the decoding process, ensuring consistency in workflows.

_fmt

The format of the decoded image, indicating the specific settings and transformations applied during the process. This output helps in understanding the final image characteristics and ensuring compatibility with subsequent processing steps.

◎ Radiance VAE Decode Pro Usage Tips:

• Ensure that the vae model used for decoding matches the one used for encoding to maintain consistency and quality in the output image.
• Experiment with different tile_size and overlap settings to optimize performance and minimize artifacts, especially when working with high-resolution images.
• Use the exposure parameter to adjust the brightness and contrast of the output image, which can be particularly useful when dealing with HDR content.

◎ Radiance VAE Decode Pro Common Errors and Solutions:

"Log curves are unavailable — falling back to sRGB linearization"

• Explanation: This warning indicates that the specified log color space is not available, and the node is defaulting to sRGB linearization, which may result in incorrect color transformations.
• Solution: Ensure that the necessary log curves are installed and available. Consider installing color_utils.py to provide the required color space transformations.

"Insufficient VRAM for tile processing"

• Explanation: This error occurs when the selected tile_size is too large for the available VRAM, causing the decoding process to fail.
• Solution: Reduce the tile_size or increase the overlap to decrease VRAM usage. Alternatively, consider upgrading your hardware to support larger tile sizes.