audio-forensics

AI audio watermark removal & forensic anti-attribution toolkit. Neutralizes Suno, Udio, and neural fingerprints with ISO 226 perceptual masking.

Python tool for aircraft Doppler audio geolocation using historical ADS-B flight data (Source-available)

Locate an audio or video recording from the Doppler curves of propeller aircraft, matched against archived ADS-B flight tracks. Rust, on frankenscipy and asupersync.

Locate where a recording was made from the Doppler curves of aircraft in its background, matched against ADS-B history. Agent skill + tested Python pipeline, English / 简体中文.

Find out where an audio recording was made. Planes flying past leave a faint engine hum whose pitch drops as they go by. AeroDoppler compares those pitch changes with public records of where planes actually flew, and works out where the microphone must have been.

Detect fake FLAC files — find MP3/AAC/Opus transcodes disguised as lossless audio. Fast Rust CLI + web UI.

Remove the sounds that give away where and when a recording was made (aircraft Doppler tones, mains hum, metadata) and keep the voices.

Ensemble speaker verification achieving 97% accuracy - Intelligent fusion of MFCC+DTW (92%) and Resemblyzer CNN (94%) for voice authentication

ArtifactBench — a reproducible, lineage-aware benchmark for AI-generated music detection. v2 frozen protocol (arXiv:2609.23550), plus the v1.1 8-way public model comparison.

Audio analysis toolkit

A robust AI-powered web app to detect deepfake audio using Wav2Vec2 and FastAPI. Deployed on Hugging Face Spaces for the CyberSecurity Hackathon.

Audio forensics analyzer for Android. Spot fake lossless and upsampled files, measure LUFS, true peak, dynamic range and clipping, and explore the spectrogram, all on-device, in a skeuomorphic rack-unit UI.

Desktop GUI tool for detecting audio deepfakes using hand-crafted features + XGBoost/LightGBM ensemble. Segment-level analysis, heatmaps, PDF reports. Zero deep learning required.

Text2Speech2Clone Data Set

Detecta manipulaciones en grabaciones de audio con IA: localiza el empalme y genera un informe pericial con hash SHA-256

Luma is an enterprise-grade deepfake detection and forensic analysis platform that uses a 3-stream ensemble architecture to analyze visual, biological, and acoustic signals, delivering explainable, human-readable evidence instead of just a probability score.

Master extractor: an exhaustive, reproducible measurement dump for lossless audio files. Measures; never interprets.

A web app that detects deepfake voices using an LLM forensic analyzer.

Free, open-source forensic audio workstation, non-destructive editing, spectral analysis, and a tamper-evident audit trail, the kind of tooling usually locked behind expensive proprietary software.

Free, local-first and explainable AI-origin music analysis for artists, labels, platforms, rights teams and researchers.