Save 4 hours! We auto-setup your workflow! Free!

Drop your workflow.json — we handle every dependency, custom node, and model. Just open the link and run.

Auto-Setup Workflow Json (Free) Now!
ComfyUI > Nodes > ComfyUI-MAINodes > H3 Temporal Insert (insert token-times, freeze originals) [experimental]

ComfyUI Node: H3 Temporal Insert (insert token-times, freeze originals) [experimental]

Class Name

H3TemporalInsert

Category
latent/minimax/motion
Author
matlowai (Account age: 1004days)
Extension
ComfyUI-MAINodes
Latest Updated
2026-08-26
Github Stars
0.11K

How to Install ComfyUI-MAINodes

Install this extension via the ComfyUI Manager by searching for ComfyUI-MAINodes
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter ComfyUI-MAINodes in the search bar
After installation, click the Restart button to restart ComfyUI. Then, manually refresh your browser to clear the cache and access the updated list of nodes.

Visit ComfyUI Online for ready-to-use ComfyUI environment

  • Free trial available
  • 16GB VRAM to 80GB VRAM GPU machines
  • 400+ preloaded models/nodes
  • Freedom to upload custom models/nodes
  • 200+ ready-to-run workflows
  • 100% private workspace with up to 200GB storage
  • Dedicated Support

Run ComfyUI Online

H3 Temporal Insert (insert token-times, freeze originals) [experimental] Description

Sophisticated node for manipulating temporal sequences in video data by inserting frames based on a temporal map for detailed analysis and creative effects.

H3 Temporal Insert (insert token-times, freeze originals) [experimental]:

H3TemporalInsert is a sophisticated node designed to manipulate temporal sequences in video data by inserting additional frames or tokens based on a specified temporal map. This node is particularly useful for expanding video sequences along the time dimension, allowing for more detailed temporal analysis or creative effects. By leveraging a temporal hold map, H3TemporalInsert can precisely control which frames are duplicated and which are interpolated, ensuring that the resulting sequence maintains a coherent flow. This capability is essential for applications that require temporal dilation, such as slow-motion effects or detailed frame-by-frame analysis. The node's ability to handle both video and audio data, while maintaining synchronization, makes it a versatile tool for multimedia processing.

H3 Temporal Insert (insert token-times, freeze originals) [experimental] Input Parameters:

samples

The samples parameter represents the input video data in the form of latent tensors. This data is the foundation upon which the temporal insertion operations are performed. The quality and structure of the input samples directly affect the output, as they determine the base content that will be expanded or interpolated. There are no explicit minimum or maximum values, but the input should be a well-formed tensor representing video frames.

hold_map

The hold_map parameter is a crucial input that dictates how the temporal insertion is executed. It is a map that specifies which frames should be held (duplicated) and which should be interpolated. This map allows for precise control over the temporal expansion process, ensuring that the resulting sequence aligns with the desired temporal structure. The hold map must be carefully crafted to achieve the intended effect.

init_mode

The init_mode parameter determines the method used to initialize the inserted frames. Options include "lerp" for linear interpolation between frames and "noise" for initializing with random noise. This choice affects the visual continuity and style of the inserted frames, with "lerp" providing smoother transitions and "noise" offering a more randomized appearance.

expand_to_end

The expand_to_end parameter is a boolean flag that, when set to true, ensures that the temporal expansion continues until the end of the sequence. This option is useful for ensuring that the entire sequence is uniformly expanded, rather than stopping at a predefined point. The default value is typically false, allowing for partial expansion based on the hold map.

noise_seed

The noise_seed parameter is used when the init_mode is set to "noise." It provides a seed for the random number generator, ensuring that the noise initialization is reproducible. By setting a specific seed, users can achieve consistent results across multiple runs, which is essential for debugging and iterative development.

H3 Temporal Insert (insert token-times, freeze originals) [experimental] Output Parameters:

samples

The samples output parameter contains the modified video data after the temporal insertion process. This output reflects the expanded sequence, with additional frames inserted according to the hold map and initialization mode. The structure and content of this output are crucial for subsequent processing or analysis, as it represents the final result of the node's operations.

noise_mask

The noise_mask output parameter is a tensor that indicates which frames in the output sequence were inserted and initialized with noise. This mask is essential for understanding the composition of the output sequence, as it differentiates between original and newly inserted frames. It is particularly useful for debugging and ensuring that the temporal insertion process has been executed as intended.

used

The used output parameter provides a report on the hold map's application, detailing how the temporal insertion was executed. This report includes information on the number of frames held, inserted, and the overall expansion achieved. It is a valuable tool for verifying that the node has performed as expected and for adjusting parameters in future runs.

report

The report output parameter offers a comprehensive summary of the temporal insertion process, including any notable events or deviations from the expected behavior. This report is essential for understanding the node's performance and for identifying any issues that may have arisen during execution.

H3 Temporal Insert (insert token-times, freeze originals) [experimental] Usage Tips:

  • Ensure that the hold_map is carefully crafted to achieve the desired temporal expansion, as it directly influences the output sequence's structure.
  • Use the init_mode parameter to control the style of the inserted frames, choosing "lerp" for smooth transitions or "noise" for a more randomized effect.
  • Set the noise_seed for reproducibility when using noise initialization, allowing for consistent results across multiple runs.

H3 Temporal Insert (insert token-times, freeze originals) [experimental] Common Errors and Solutions:

"Mismatch in input tensor dimensions"

  • Explanation: This error occurs when the input samples tensor does not match the expected dimensions required by the node.
  • Solution: Verify that the input tensor is correctly formatted and matches the expected dimensions for video data processing.

"Invalid hold_map format"

  • Explanation: The hold_map provided is not in the correct format or contains invalid values.
  • Solution: Ensure that the hold_map is a well-structured map with valid entries that specify the desired temporal expansion.

"Noise initialization failed due to missing seed"

  • Explanation: The init_mode is set to "noise," but no noise_seed is provided, leading to inconsistent initialization.
  • Solution: Provide a valid noise_seed to ensure reproducible noise initialization when using the "noise" mode.

H3 Temporal Insert (insert token-times, freeze originals) [experimental] Related Nodes

Go back to the extension to check out more related nodes.
ComfyUI-MAINodes
RunComfy
Copyright 2025 RunComfy. All Rights Reserved.

RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.

H3 Temporal Insert (insert token-times, freeze originals) [experimental]