Better VMAF

A native macOS application for calculating VMAF (Video Multi-Method Assessment Fusion) scores between two videos.

Features

• Native macOS interface
• Real-time VMAF calculation
• Detailed metrics including:
  • VMAF score
  • Score range (min/max)
  • Harmonic mean
• Support for common video containers and codecs

System Requirements

• macOS 13.0 or later
• Self-contained FFmpeg binary with libvmaf support (included in the app bundle)

Installation

1. Download the latest release from the Releases page
2. Open the downloaded Better_VMAF.dmg file
3. Drag the "Better VMAF" app to your Applications folder
4. The first time you run the app, you'll need to:
   • Open the app
   • Select "Done" on the security pop-up
   • Navigate to System Settings > Security and Privacy
   • Scroll down to the "Security" section
   • Press "Open Anyway" next to the Better VMAF warning
   • Provide your fingerprint or administrator password

This is necessary because the app is not signed with an Apple Developer ID. You only need to do this once.

Development Setup

1. Clone the repository:

   git clone https://github.com/oliverdougherC/BetterVMAF
   cd BetterVMAF

2. Open the project in Xcode:

   open VMAF.xcodeproj

3. Build and run the project in Xcode

Building for Distribution

To create a DMG for distribution:

1. Open Terminal and navigate to the project directory
2. Run the build script:

   ./create_dmg.sh

3. The script will create Better-VMAF.dmg in the project directory

How It Works

Better VMAF uses a self-contained FFmpeg binary with the libvmaf library to calculate video quality metrics. The app provides a simple interface for:

1. Selecting a reference video (original/high quality)
2. Selecting a comparison video (to be evaluated)
3. Calculating and displaying VMAF scores

The VMAF score ranges from 0 to 100, where:

• 100 represents perfect quality
• Scores above 93 indicate excellent quality
• Scores below 60 indicate significant quality issues

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• FFmpeg for video processing
• libvmaf for VMAF calculation
• Apple's SwiftUI framework for the user interface