resemble-ai/Perth

Open Audio Watermarking Tool

What it solves

Perth is designed to protect audio content by embedding imperceptible watermarks into audio files. This allows creators to track and detect the presence of these watermarks even if the audio has been compressed, resampled, or otherwise manipulated.

How it works

The library provides a Python API and a command-line interface to apply and extract watermarks. It utilizes multiple techniques, most notably the Perth-Net Implicit approach, which uses a neural network to embed and retrieve watermarks while maintaining high audio quality.

Who it’s for

This tool is for developers and audio engineers who need to verify the audio origin or track content ownership through robust, invisible watermarking.

Highlights

  • Neural Network-based Watermarking: Uses the Perth-Net Implicit approach for high-quality, robust embedding.
  • Robustness: Watermarks are designed to survive common audio transformations like compression and resampling.
  • Integrated Evaluation Tools: Includes utilities to calculate quality metrics (such as SNR and PSNR) and visualize audio comparisons.
  • Flexible Interface: Offers both a comprehensive Python API for application integration and a CLI for quick tasks.

Related

  • Project
    Shpigford/nurb

    nurb is a locally‑run Python/macOS app that lets you describe a 3‑D part in plain language, has your existing LLM write real B‑Rep CAD code, checks the model against 13 print‑ability rules, shows live updates in a three.js viewer, and exports print‑ready 3MF (or STL/STEP/GLB) with slicer settings embedded.

  • Project
    mikeoliphant/neural-amp-modeler-lv2

    An LV2 plugin for playing back neural network machine learning amplifier models using the NeuralAudio engine.

  • Project
    BUTSpeechFIT/DiariZen

    A speaker diarization toolkit that uses self-supervised learning and structured pruning to accurately identify who spoke when in audio recordings.

  • Project
    bybit-exchange/svg-diagram

    svg‑diagram is a zero‑dependency Node skill that defines a strict visual style for SVG diagrams generated by AI coding agents and provides a command‑line linter (`svg‑lint`) to verify that a finished SVG follows the style.