Has anyone developed a system to embed redundant inaudible signals into an audio signal? Seems like you could support publishers 100x+ more than any other listener by spoofing these, while poisoning the advertiser's dataset.