Forensic watermarking embeds an invisible, per-viewer (or per-session) identifier into the video so that if a copy leaks, it can be traced back to the account or device that leaked it. Unlike DRM, which tries to prevent unauthorized access, watermarking is about accountability after the fact — it deters and helps shut down piracy by making every stream uniquely attributable.
Implementations vary in where the mark is applied. Server-side or edge approaches generate a subtly different variant per session (A/B variant stitching) so the delivered stream carries the viewer's identity; client-side approaches apply the mark in the player. The mark must survive re-encoding, cropping, and screen capture — robustness against these attacks is the hard part — while staying imperceptible to the viewer.
Forensic watermarking is increasingly mandated for high-value live content (premium sports, early releases) and is the front line of anti-piracy operations: detection systems scan pirate streams, extract the watermark, identify the source, and trigger rapid takedown or account action.
