What does the DFR (Diffusion Fidelity Rendering) path actually do?

DFR runs a half‑resolution generation that establishes motion and produces detailing keyframes, then applies a spatial ×2 pass with a pixel‑spatial IC‑LoRA to recover high‑frequency detail. An optional temporal ×2 pass can further densify motion for smoother playback. This sequence improves fidelity and temporal coherence compared to a single‑pass render.

When should I enable the optional Temporal ×2 pass?

Enable Temporal ×2 when motion feels choppy or you need extra smoothness for character beats and slow camera moves. It increases compute time but reduces temporal gaps between frames. For quick drafts or very short clips, leaving it off is faster.

Will this work with any aspect ratio?

Yes—use ResolutionSelector to target 16:9, 9:16, 1:1, etc. For best results, keep the chosen aspect close to your source image to avoid heavy cropping or padding that can reduce detail or introduce stretching.

How is audio handled in the final video?

If the subgraph produces an audio track, SaveVideo will mux it into the output so motion and sound are synchronized. If no audio branch is active, the result will be a silent clip—you can add music or VO later in an editor.

LTX-2.5 DFR: Image to Video - ComfyUI Workflow

This workflow turns a single image into a short, cinematic video using LTX‑2.5’s DFR (Diffusion Fidelity Rendering) path. The subgraph “Image to Video (LTX‑2.5 DFR)” runs a half‑resolution generation to establish motion and generate detailing keyframes, then applies a spatial ×2 detailing pass with a pixel‑spatial IC‑LoRA to recover fine texture. For projects that need extra smoothness, you can optionally add a temporal ×2 pass to densify motion between frames. The result is a cohesive clip that preserves your image’s look while adding natural camera movement and subject animation, with synchronized audio when provided by the subgraph.

Technically, the pipeline is compact and easy to drive in ComfyUI: LoadImage brings in your source frame, ResolutionSelector sets aspect and target size, and the DFR subgraph node (97d5d35b-bb0d-4212-b2f2-97ab13cbfb1f) encapsulates the LTX‑2 family model and its IC‑LoRA detailing passes. PreviewImage helps you quickly inspect a representative frame before final encoding, and SaveVideo writes the finished clip (including audio if the subgraph outputs it). MarkdownNote nodes in the template annotate key choices and link to ComfyUI’s Subgraph feature, so you can open the subgraph and tweak options without getting lost.

API

Use this workflow from code

Every Comfy workflow is a JSON graph. The payload below is this workflow, exactly as ComfyUI runs it — fetch it from the URL, keep it in version control, or load it in ComfyUI and run it node by node.

999b68cf27cb.json

Fetching workflow JSON…

Run it from TypeScript or Python with the Comfy SDK. The same code targets Comfy Cloud or a ComfyUI you host yourself — only the base URL changes.

// Install (beta)
npm i @comfyorg/sdk

// Run this workflow (TypeScript)
import { Comfy } from "@comfyorg/sdk";

const client = new Comfy({ apiKey: "comfyui-..." });
const wf = await client.workflows.fromFile("workflow_api.json");
const job = await client.run(wf);
await job.getOutputs("<output-node-id>")[0].toFile("output.png");

The SDK takes a workflow in API format: open this workflow in ComfyUI and use File → Export Workflow (API).

Comfy Cloud API access requires a plan with an API key.

SDK docs Get an API key

FAQ

Frequently Asked Questions