The Anime Scripter ( TAS )

Table of Contents

• The Anime Scripter ( TAS )
  • Table of Contents
  • 🚀 Key Features
  • 🖥️ Graphical User Interfaces
  • 🛠️ Getting Started
    • Adobe After Effects Integration
    • Standalone / GUI:
      • Windows
      • Linux
    • CLI Interface:
      • Windows / Linux
  • 📚 Available Inputs
  • 📁 Available Models
    • Upscaling Models
    • Interpolation Models
    • Restore Models
  • 📈 Benchmarks
  • 🙏 Acknowledgements
  • 🌟 Star History
  • 🎥 Promo Video

🚀 Key Features

1. Smooth Motion Interpolation: Elevate video quality with seamless frame interpolation for fluid motion.

2. Crystal-Clear Upscaling: Immerse audiences in enhanced resolution, delivering sharper, more detailed visuals.

3. Optimal Video Size Deduplication: Streamline videos by intelligently removing redundant frames, optimizing file size.

4. Effortless Background-Foreground Segmentation: Simplify rotobrushing tasks with precision and ease.

5. 3D Wizardry - Depth Map Generation: Unlock advanced editing possibilities with detailed depth maps for immersive 3D effects.

6. Auto Clip Cutting with Scene Change Filter: Boost productivity by automatically cutting clips with a Scene Change Filter.

7. Seamless After Effects Integration: Enhance After Effects projects effortlessly with our seamless integration.

8. Multi-Effect Magic - Model Chaining: Combine features seamlessly within After Effects, running Deduplication, Upscaling, and Interpolation in one go.

9. Efficient In-Memory Processing: Experience swift transformations without additional frame extraction cycles.

10. Custom Model Support for Creativity: Unleash your creativity by incorporating your own trained models effortlessly.

11. Graphical User Interface: Navigate through our user-friendly interface designed for both beginners and professionals, ensuring a smooth editing workflow from start to finish.

🖥️ Graphical User Interfaces

• Windows and After Effects: Actively being reworked and improved!

  Adobe GUI

🛠️ Getting Started

Adobe After Effects Integration

1. You must have After Effects 2020 v17.5 or higher!
2. Download the latest -AdobeEdition release from here.
3. Extract the .zip file.
4. Follow this step-by-step tutorial

If installation fails, refer to the manual installation guide.

Video tutorial: https://youtu.be/JAdZ3z-os_A?si=fZQPmhMLtHfAktwn

Standalone / GUI:

Windows

• N/A - Under active development.

Linux

• N/A - Under active development.

CLI Interface:

Windows / Linux

• Stable: Download the latest release from here
• Nightly: Download nightly versions from here

📚 Available Inputs

All available parameters for interacting with the CLI or directly with main.py can be found in the Parameters guide.

📁 Available Models

Upscaling Models

Officially Supported:

• ShuffleCugan (CUDA, TensorRT, and NCNN)
• Span (CUDA, TensorRT, DirectML and NCNN versions)
• SRVGGNet (Available in Compact, UltraCompact, SuperUltraCompact, and their respective TensorRT and DirectML versions)
• OpenProteus ( Cuda, TensorRT and DirectML )
• AniScale 2 ( Cuda, TensorRT and DirectML )
• RTMOSR ( TensorRT and DirectML )

Unofficially Supported:

• Custom models compatible with Spandrel can be used via the --custom_model parameter.

Interpolation Models

• Rife CUDA (Versions: 4.6, 4.15, 4.15-lite, 4.16-lite, 4.17, 4.18, 4.20, 4.21, 4.22, 4.22-lite, 4.25, 4.25-heavy, Rife_Elexor ( modded 4.7 ) )
• Rife TensorRT (Versions: 4.6, 4.15, 4.15-lite, 4.17, 4.18, 4.20, 4.21, 4.22, 4.22-lite, 4.25, 4.25-heavy, Rife_Elexor ( modded 4.7 ) )
• Rife NCNN (Versions: 4.6, 4.15, 4.15-lite, 4.16-lite, 4.17, 4.18, 4.20, 4.21, 4.22, 4.22-lite )

Restore Models

• SCUNet ( Denoise ) ( Cuda )
• NAFNet ( Denoise ) ( Cuda )
• DPIR ( Denoise ) ( Cuda )
• Real-Plksr (DeJpeg) ( Cuda )
• Anime1080fixer ( Sharpen + Decompress ) ( Cuda & TensorRT )
• FastLineDarken ( Line darkening ) ( Cuda & TensorRT )

📈 Benchmarks

Both internal and user-generated benchmarks can be found here.

🙏 Acknowledgements

Name	For
SUDO	Shufflecugan and many more ONNX models
HZWER	Rife
the-database	Compact, UltraCompact, SuperUltraCompact models
SkyTNT	Anime Segmentation
DepthAnything	Depth Anything V2
FFmpeg Group	FFmpeg
YT-DLP	YT-DLP
Hongyuanyu	SPAN
Breakthrough	Automated Scene Detection
Chainner-org	Spandrel, easy to use arch implementations
TNTWise	Rife ONNX / NCNN and Spanimation
Hyperbrew	Bolt CEP
Sirosky	Open-Proteus and AniScale 2
Trentonom0r3	Helping with TAS Adobe Edition and Celux
Adegerard	Several ideas on how to further improve TAS
Elexor	Modded Rife Experiment(s)
Zarxrax	Anime1080Fixer model
sdaqo	Anipy-CLI
umzi	RTMOSR

If I forgot to mention anyone, please email: nilascontact@gmail.com

🌟 Star History

🎥 Promo Video