Invisible Watermarks for Audio Leak Attribution
Embed a unique recipient ID into a master, stem, or screener. Then scan a copy to see whether that ID can be recovered after your workflow changes the audio.
Seamless Studio Ingestion
Embed invisible recipient signatures into master recordings, stems, and screeners during export with 0.0 dB perceptible coloration.
Distribution & Attack Resilience
Use the demo to evaluate detection after the transformations that matter to your distribution workflow.
Deterministic Attribution
Scan a suspected copy and compare the recovered ID with your distribution record.
CLI & Automated Pipeline Integration
Deploy as a standalone air-gapped binary or embed directly into Pro Tools, Reaper, batch mastering workflows, and automated leak monitoring clusters.
git clone https://github.com/KELLERBABG/Aureal-Watermark.git && cd Aureal-Watermark && npm test
Engineered for Major Studios, Record Labels & Rights Holders
Why traditional ID3 tags, ultrasonic chirps, and acoustic fingerprints fail — and how Aureal guarantees forensic chain of custody.
Pre-Release Screener & Promo Leak Attribution
When distributing unreleased music, film cues, stems, or review screeners to 200 journalists, radio stations, and mixing engineers, serialize every copy with a unique 32-bit recipient ID. If a track leaks to Telegram, Discord, or torrent trackers, recover the exact recipient within 350ms.
Lossy Transcoding & Social Media Resilience
Pirated leaks are never uncompressed WAV. Pirates compress to 64–128 kbps MP3, upload to TikTok / YouTube, or record audio through phone speakers. Aureal's DSSS spread-spectrum modulation embeds beneath biological psychoacoustic curves, surviving extreme lossy re-encoding and pitch shifts.
AI Voice Model & Training Set Ingestion Audit
Forensic watermarks can help you audit distributed audio. Validate results against your own files and keep the original master and distribution record as evidence.
100% Air-Gapped / Zero Cloud Leak Risk
Major studios will never upload high-profile unreleased masters to third-party cloud APIs. Aureal is written with zero npm dependencies and executes 100% locally on DAW workstations (Pro Tools, Reaper) or private on-premise clusters. Zero telemetry, zero external network requests.
Human Auditory System (HAS) Masking Curves
Metadata can be stripped during export or re-encoding. Aureal explores spread-spectrum embedding under a perceptual threshold; test the result with your own delivery pipeline.
Frequency Masking Thresholds
When a strong musical tone plays (e.g. 1 kHz snare transient), adjacent frequency bands are biologically rendered inaudible to the human cochlea:
Direct Sequence Spread Spectrum (DSSS)
The watermark payload is spread across thousands of frequency bins using an orthogonal cryptographic key sequence.
Because the signal energy in any single bin is orders of magnitude below ambient acoustic noise, lossy audio codecs cannot compress or filter it away without destroying the underlying music.
Audio Attribution Solutions Compared
Why ID3 metadata tags and acoustic fingerprinting fail to stop AI scraping and music theft.
| Watermarking Technique | ID3 Metadata Tags (MP3 / FLAC) | Acoustic Fingerprinting (Shazam) | Aureal Acoustic Spread-Spectrum |
|---|---|---|---|
| Lossy Compression Resistance | Stripped immediately on YouTube/TikTok re-encode | Database matching only | Tested in selected scenarios transcoding |
| Individual Recipient Tagging | Impossible (File hash is shared) | Cannot distinguish two identical tracks | Unique Invisible Serial ID per preview recipient |
| AI Voice Model Scrape Audit | Ignored by web scrapers | Fails on spliced synthetic voice training | Verifiable Provenance Signature inside training data |
| Offline Private Scanning | Trivial text inspection | Requires central corporate database | Client-Side Demo Scanner in browser |