ComfyUI Node: Woosh Video Loader

Woosh Video Loader:

The WooshLoadVideo node is designed to facilitate the loading and processing of video files or image batches for video-to-audio generation tasks. This node is particularly useful for AI artists who wish to extract frames from video files or convert image batches into a format suitable for further processing with models like VFlow or DVFlow. By providing a seamless interface for handling video data, WooshLoadVideo enables users to efficiently prepare their media content for creative audio synthesis, ensuring that the video data is correctly formatted and within the specified duration limits. This node is essential for those looking to integrate video content into their AI-driven artistic workflows, offering a straightforward method to bridge the gap between visual and audio media.

Woosh Video Loader Input Parameters:

video_path

The video_path parameter is a string that specifies the path to the video file you wish to load. This can be any standard video format such as .mp4 or .avi. If you choose to use this parameter, ensure that the path is correct and the file exists, as it is crucial for the node to locate and process the video. The default value is an empty string, indicating that no video file is specified by default.

max_duration_s

The max_duration_s parameter is a float that determines the maximum duration of the video to be processed, measured in seconds. This parameter is important for controlling the length of the video segment that will be extracted and processed. The minimum value is 1.0 seconds, the maximum is 30.0 seconds, and the default is set to 8.0 seconds. This default aligns with the VFlow-8s model's limit, ensuring compatibility with specific model requirements.

image_batch

The image_batch parameter allows you to input a batch of images instead of a video file. This is useful if you have a sequence of images that you want to convert into a video format for processing. The images should be in the format [B, H, W, C] with float32 values ranging from 0 to 1. This parameter provides flexibility in input types, enabling the use of pre-existing image data for video-to-audio tasks.

Woosh Video Loader Output Parameters:

video

The video output parameter is a dictionary containing the processed video frames and the frame rate. The frames are extracted from the video file or converted from the image batch and are formatted as a tensor suitable for further processing. This output is crucial for subsequent steps in video-to-audio generation, as it provides the necessary video data in a compatible format for models like VFlow or DVFlow.

Woosh Video Loader Usage Tips:

• Ensure that the video_path is correctly specified and that the file exists to avoid errors during processing. Double-check the file path for typos or incorrect directory references.
• When using the image_batch parameter, make sure the images are correctly formatted and normalized to the expected range of [0, 1] to ensure accurate conversion and processing.
• Adjust the max_duration_s parameter according to the specific requirements of your project or the limitations of the model you are using. This can help optimize processing time and resource usage.

Woosh Video Loader Common Errors and Solutions:

Must provide either video_path or image_batch

• Explanation: This error occurs when neither a video path nor an image batch is provided to the node. The node requires at least one of these inputs to function.
• Solution: Ensure that you provide a valid video_path or image_batch input. Check that the video file exists or that the image batch is correctly formatted and passed to the node.

Video file missing: <video_path>

• Explanation: This error indicates that the specified video file could not be found at the given path. It may be due to an incorrect path or a missing file.
• Solution: Verify that the video_path is correct and that the video file exists at the specified location. Correct any typos in the path and ensure the file is accessible.