Feasibility Study of a Parallel Audio and Video ECC for Watermarks in Short Audio and Video Frames
Kazuhiro Kondo, Yoshimi Yamaguchi, Toshio Koga · Multimedia technology · 2014
We investigated the feasibility of a parallel watermarking method for audio and video signals for multimedia streams with editing and/or switching capability. Since editing and switching are commonly done in short intervals, these watermarks need to be applied on a short frame basis. In this study, the watermark information with error correction coding (ECC) will be split and embedded into audio and video simultaneously. Information decoded from both audio and video signals is to be used collectively to recover the embedded information even if there is severe degradation in the decoded signal. We studied a scheme in which the watermark is embedded in the video stream with ECC, and the ECC is also redundantly embedded into the audio stream. We conducted a very basic simulation using a simple DCT‐based watermark for video, when the Bose‐Chaudhuri‐ Hocquenghem (BCH) code is placed in both video and audio streams. At an overall bit error rate of 5%, with BCH (63, 51) ECC, we estimated that the error correction capability can be improved by approximately 35% compared to when all information is embedded into only the video stream. A simple calculation‐based estimation assuming uniform error bit distribution also showed error correction capability improvement of approximately 48%, which is slightly higher than it was shown with the simulation.