They say in their answer that they use audio fingerprinting for ACR instead of screenshots. Are you certain they are lying about this?