Unlock Stunning 60FPS Videos: A Comprehensive Workflow Guide

CN
ComfyUI.org
2025-04-04 08:04:55

1. Workflow Overview

m92i0nfqq2eaxhv5z5e07e9db0d72925ab69e470208cd61bfd085f3ce0855bf287c666ec3583bc8a4e.gif

This is a Video Generation & HD Enhancement workflow with:

  • First/Last frame-based video interpolation (Image-to-Video)

  • 4K Super-Resolution Upscaling

  • Smart Frame Interpolation (60FPS output)

  • Video Style Consistency Control

2. Core Models

Model Name

Function

Wan2_1-I2V-14B-480P_fp8_e4m3fn

Main video generation model (480P output)

open-clip-xlm-roberta-large-vit-huge-14

Multilingual CLIP for image feature encoding

umt5-xxl-enc-bf16

Text encoder for negative prompts

rife47.pth

RIFE model for 60FPS interpolation

3. Key Nodes

Node Name

Function

Installation

WanVideoSampler

Core video generation sampler

Install ComfyUI-WanVideoWrapper

WanVideoClipVisionEncode

Image feature encoding (style control)

Same as above

RIFE VFI

Frame interpolation (FPS boost)

Install ComfyUI-VideoHelperSuite

VHS_VideoCombine

Video composition & export

Same as above

4. Workflow Structure

  • Group 1: Frame Processing

    • Input: Start/End frames (Auto-resized to 512x512)

    • Output: Normalized images + dimension parameters

  • Group 2: Video Generation

    • Input: Frames + Text prompts (Auto negative prompts)

    • Output: 480P raw video sequence (20 frames)

  • Group 3: Post-Processing

    • Input: Raw video + Interpolation params

    • Output: 60FPS HD video (H.264 MP4)

5. Inputs & Outputs

  • Required Inputs:

    1. Start frame (JPG/PNG)

    2. End frame (JPG/PNG)

    3. Output resolution (Default 832x480)

  • Output:

    • Format: MP4 (H.264)

    • Frame Rate: 60FPS

    • Example: WanVideoWrapper_I2V_endframe_00009.mp4

6. Notes

  • VRAM: β‰₯16GB GPU (For 480P generation)

  • Common Errors:

    • Frame size mismatch β†’ Use ImageScaleByAspectRatio node

    • Interpolation lag β†’ Reduce Upscale Factor in RIFE VFI (Default 10 β†’ 5)

  • Optimization:

    • Enable torch.compile (Requires NVIDIA RTX 30+)

    • Use bf16 precision to save VRAM

Recommend