Transform any 2D image into immersive stereoscopic 3D content in ComfyUI.
The SideBySide Stereoscope Node is a cutting-edge 2D-to-3D conversion tool designed for ComfyUI. Powered by Depth-Anything (V2 & V3), it automatically creates high-quality depth maps and generates side-by-side (SBS) 3D images and videos perfect for VR headsets and 3D displays.
- 🧠 AI-Powered Depth - Uses Depth-Anything (V2 & V3) for automatic, high-quality depth estimation.
- 👓 Stereoscopic Modes - Supports Parallel and Cross-eyed viewing techniques.
- 🎥 Video Support - Full workflow for converting videos to 3D (upload, process, combine).
- 🎛️ Fine Control - Adjust depth scale, blur radius, and depth inversion.
- ⚡ Integrated Workflow - No external tools needed; handles everything within ComfyUI.
- 🚀 Performance Mode - Optional "HighSodium Optimization" for up to 4x faster processing.
- ComfyUI installed
- Python >= 3.8
- FFmpeg (Recommended for video support)
- The node attempts to use your system's FFmpeg first.
- If missing, it will fallback to the
imageio-ffmpegbinary (included in dependencies). - Windows Users: For best performance, install FFmpeg manually from ffmpeg.org and ensure it is added to your System PATH.
- Navigate to your custom nodes directory:
cd ComfyUI/custom_nodes - Clone the repository:
git clone https://github.com/SamSeenX/ComfyUI_SSStereoscope.git
- Install dependencies:
pip install -r ComfyUI_SSStereoscope/requirements.txt
- Restart ComfyUI.
(Note: The AI model will download automatically on first use.)
- 👀 SBS V2: The main node. Input a 2D image, get a 3D SBS image + depth map.
depth_scale: Intensity of the 3D effect.mode: Cross-eyed vs Parallel.highsodium_optimization: Enable vectorized algorithm for faster processing (may produce slightly different results).
- 👀 SBS V2.1 (External Depth): New in v2.1! Uses the fast V2 engine (HighSodium) but accepts your custom Depth Maps.
- Fixes "reducing" artifacts found in the legacy node.
- Resolution-Relative Scaling:
depth_scale(0-100) now maps to 0-20% of image width. This gives you fine-grained control while preventing output-breaking values.- Scale 10 = 2% width separation.
- Scale 50 = 10% width separation.
- Scale 100 = 20% width separation (Max).
- 👀 SBS Video Uploader: Converts input video to image sequence.
- 👀 SBS Video Combiner: Merges processed frames back into a video.
- Video Input: Use
SBS Video Uploaderto load a video. - Process: Connect frames to
SBS V2node. - Output: Connect results to
SBS Video Combinerto save the 3D video.
ComfyUI_SSStereoscope/
├── sbs_v2.py # Main logic
├── requirements.txt # Dependencies
├── video_utils.py # Video processing helpers
└── README.md
Contributions are welcome!
- Fork the repository
- Create your feature branch
- Commit your changes
- Push to the branch
- Open a Pull Request
- SamSeen - Original Author
- HighSodium - Performance optimization (PR #16)
This project is licensed under the MIT License - see the LICENSE file for details.
If you find this project useful, please consider supporting me:
- ⭐ Starring this repository
- 🐛 Reporting issues
- ☕ Buy me a coffee
Created with ❤️ by SamSeen
