H3 MCtx Load (obvpm):
The H3LoadMCtx node is designed to efficiently load a clip's motion-context bundle, which includes latents and lineage, from a verified sidecar file without decoding the video itself. This node is particularly useful for workflows that require only the latent data for further processing, such as extending graphs, without the need to handle the actual video frames or audio. By focusing on the sidecar file, H3LoadMCtx ensures a fast and resource-efficient operation, as it bypasses the need to decode the video, making it ideal for scenarios where only the latent information is needed. This approach not only saves time but also preserves the quality of the data by avoiding unnecessary conversions.
H3 MCtx Load (obvpm) Input Parameters:
clip
The clip parameter specifies the video file located in the output folder that has an associated motion-context (mctx) sidecar. This sidecar is crucial as it contains the verified bundle of latents and lineage data. The parameter expects a string representing the file name of the video. If the clip is not specified or the picker is empty, the node will raise an error. This parameter is essential for the node to function, as it identifies the specific video and its corresponding sidecar to be processed.
create_pins
The create_pins parameter determines how the node should handle pin creation for the loaded clip. It offers options such as "none," "extend (pin tail)," and "prepend (pin head)." These options dictate whether and how the node should create pin specifications for the clip, which can be used in further processing. The default value is "none," meaning no pins are created unless specified otherwise. This parameter allows users to customize the handling of the clip's data within their workflow.
pin_window
The pin_window parameter specifies the window size for pin creation, which is relevant when the create_pins option is set to either "extend (pin tail)" or "prepend (pin head)." It defines the number of frames to consider when creating pins, allowing users to control the granularity of the pinning process. The default value is "22," but users can adjust this to suit their specific needs, depending on the desired level of detail in the pinning process.
H3 MCtx Load (obvpm) Output Parameters:
mctx
The mctx output provides the motion-context bundle, which includes the latents and lineage data extracted from the verified sidecar. This output is crucial for workflows that require latent data for further processing, as it contains all the necessary information without the need to decode the video itself.
pin_specs
The pin_specs output delivers a ready-made pin specification based on the create_pins parameter. If no pins are created, this output will be empty. This output is useful for users who need to integrate the clip's data into a larger workflow, as it provides a structured way to handle the clip's information.
latent
The latent output contains the stored audio-visual latent data, exactly as it was produced by the sampler. This output is essential for users who need to work with the latent representation of the video and audio, as it allows for further processing without any loss of quality due to decoding or encoding.
H3 MCtx Load (obvpm) Usage Tips:
- Ensure that the video file has a corresponding verified sidecar in the output folder to utilize the node effectively.
- Use the
create_pinsparameter to customize how the clip's data is integrated into your workflow, especially if you need specific pin configurations.
H3 MCtx Load (obvpm) Common Errors and Solutions:
obvpm.h3: H3 MCtx Load (obvpm) has no clip -- the picker is empty.
- Explanation: This error occurs when the
clipparameter is not specified or is empty, meaning the node does not have a video file to process. - Solution: Ensure that you provide a valid video file name in the
clipparameter that corresponds to a file in the output folder with an associated mctx sidecar.
H3MCtxPinSpec: mctx is None -- the loaded clip has no verified sidecar.
- Explanation: This error indicates that the specified clip does not have a verified sidecar, which is necessary for the node to function.
- Solution: Verify that the video file has a corresponding sidecar file in the output folder. If not, consider using the pixel route with
H3MCtxFromFramesor re-save the take withH3SaveVideoWithMCtx.
