But how would this steganograohic ID fair throughout the multiple encoding/transcoding that happens during scene distribution?