Study the effect of commonly used video compression techniques on sound recovery via negligible object vibrations for visual surveillance system

Closed circuit televisions (CCTVs) are becoming an inseparable part of our daily lives. The growth in surveillance technologies have provided a sense of security, leading to a lower crime rate in the monitored area. However, almost all the outdoor CCTVs are not equipped with a built-in microphone for sound recording due to some practical reasons. Recently, it has been shown that it is possible to recover sound via subtle vibrations of an object in a high speed video, leading to an overwhelming interest in remote sound acquisition. This study extends the literature further by investigating the feasibility of sound recovery from some commonly used CCTV compression techniques which can result in the compressed video file size to be more than 100 times smaller than the original high speed video. Qualitative and quantitative measures were used to evaluate the sounds recovered. It was found that high quality sounds could be recovered via video in the format of some commonly used compression techniques in CCTVs. These findings may pave the way to enable the CCTVs to have not only 'eyes' but also 'ears' to improve the surveillance system, despite not having any audio input.