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 > Difforum > Difforum · Flow Stabilize

ComfyUI Node: Difforum · Flow Stabilize

Class Name

Difforum_FlowStabilize

Category
Difforum/7 · Post
Author
chillithebillis (Account age: 0days)
Extension
Difforum
Latest Updated
2026-11-03
Github Stars
0.01K

How to Install Difforum

Install this extension via the ComfyUI Manager by searching for Difforum
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter Difforum 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

Difforum · Flow Stabilize Description

Stabilizes image sequences with motion-aligned blending to reduce flicker while limiting trails.

Difforum · Flow Stabilize:

Difforum_FlowStabilize is an anti-flicker node for image sequences. It blends information from earlier images along estimated motion, but only where that information appears to match the current image. This is intended to calm textures that seem to shimmer or “boil” from frame to frame while reducing the risk of trails behind moving subjects. It is most useful when a sequence has distracting changes in fine detail but you still want its motion to remain clear.

Difforum · Flow Stabilize Input Parameters:

frames

The IMAGE sequence to stabilize. The node processes these images together so it can use motion and image history to reduce flicker. This input has no default, minimum, or maximum specified in the interface. Connect the sequence you want to treat, rather than a single image when your goal is to improve consistency over time.

strength

Controls how strongly the node blends matching image history into the result. Higher values request more stabilization; lower values keep the result closer to the incoming images. Start with the default and adjust while watching both flickering textures and moving edges, since the best balance depends on the sequence. Minimum: 0.0; maximum: 0.95; default: 0.5; interface step: 0.05.

flow_scale

Sets the scale used for the optical-flow component that follows motion between images. The provided node source does not show the underlying flow calculation, so it does not establish a guaranteed speed or quality difference between settings. Treat this as a setting to compare visually on your own sequence, particularly around moving details. Minimum: 0.25; maximum: 1.0; default: 0.5; interface step: 0.25.

error_gate

Controls the match check used to decide where history can be blended. This matters when an earlier image no longer represents what is visible now, such as around a moving edge or a newly revealed area. The provided source does not define the gate’s exact calculation or establish how a numerical change translates into a particular visual result, so compare nearby settings rather than assuming that a larger value is always better. Minimum: 0.02; maximum: 0.5; default: 0.15; interface step: 0.01.

Difforum · Flow Stabilize Output Parameters:

frames

The resulting IMAGE sequence after stabilization. Use it in place of the incoming sequence for later processing or viewing. Compare it with the original sequence in motion: the intended improvement is less flicker in textures that remain consistent, without conspicuous trails around moving content. This is the node’s only output.

Difforum · Flow Stabilize Usage Tips:

  • Begin with the defaults: strength 0.5, flow_scale 0.5, and error_gate 0.15. Preview several consecutive images, since flicker and unwanted trails are easier to judge in motion than in a still image.
  • If textures still shimmer, raise strength gradually in 0.05 steps. Stop when additional blending no longer improves consistency or starts to make moving detail look less clear.
  • If you see trails or stale detail near motion, compare a lower strength first. You can then test nearby error_gate values to find a better balance between accepting matching history and rejecting mismatches.
  • Compare the available flow_scale settings on a short section with noticeable motion. The node interface offers values from 0.25 to 1.0, but the supplied source does not establish which setting will perform best for a particular scene.

Difforum · Flow Stabilize Common Errors and Solutions:

Missing required input: frames

  • Explanation: The node requires an IMAGE connection for the sequence it will process. This is an illustrative validation message; the exact wording depends on your host application.
  • Solution: Connect an IMAGE sequence to frames and run the node again.

Missing required input: strength, flow_scale, or error_gate

  • Explanation: All three numeric settings are required by the interface. This is an illustrative validation message, not a confirmed message from the node.
  • Solution: Supply values for every setting. To start, use strength 0.5, flow_scale 0.5, and error_gate 0.15.

Value out of range for strength, flow_scale, or error_gate

  • Explanation: A supplied value may be outside the range declared by the node interface. The exact validation message depends on your host application.
  • Solution: Keep strength between 0.0 and 0.95, flow_scale between 0.25 and 1.0, and error_gate between 0.02 and 0.5.

Input type mismatch: frames expects IMAGE

  • Explanation: A connection carrying a different data type may not be accepted at the frames socket. This is an illustrative host validation message.
  • Solution: Connect an IMAGE-producing node or convert the source to IMAGE before connecting it.

Runtime error inside stabilize

  • Explanation: The node passes its inputs to an imported stabilization function, but that function’s implementation and error messages were not provided. A runtime failure therefore cannot be diagnosed more precisely from this source alone.
  • Solution: Check the full error text in your host application, confirm that the connected images and numeric settings are valid, and test the same sequence with the default settings to narrow down the cause.

Difforum · Flow Stabilize Related Nodes

Go back to the extension to check out more related nodes.
Difforum
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.