facebookresearch / audioseal

Localized watermarking for AI-generated speech audios, with SOTA on robustness and very fast detector
MIT License
466 stars 56 forks source link

Add Evaluation Support for Multiple Audio Formats (AAC 64kbs and OBD) in AudioSeal #58

Open hastagAB opened 1 week ago

hastagAB commented 1 week ago

Description

Add support for evaluating the AudioSeal watermarking system across multiple audio formats, enhancing its applicability and robustness. The changes include the capability to handle different audio file types such as AAC 64kbs and OBD during the evaluation process.

Key Changes

  1. Support for Multiple Formats: Added functionality to evaluate audio signals from multiple audio formats (AAC 64kbs and OBD)
  2. Audio Conversion Logic: Integrated audio conversion to ensure compatibility with the AudioSeal processing pipeline.