Active noise control for smartphone communication under sound masking

Abstract

Mobile phones are used in a variety of situations where speech privacy of users can be invaded by the surrounding people. When we have a phone conversation, we often care to keep conversation as private as possible from people around us whether it is a personal issue or a confidential business matter. This thesis attempts to design a mobile system for isolating speech from the unintended audience by achieving speech privacy of a near-end user when having a phone conversation with a far-end user. It initially demonstrates that sound masking technology can be implemented to provide speech privacy using built-in loudspeakers of smartphones. However, a generated masking sound has an adverse effect on the communication quality and interferes with the ability of the near-end user to communicate with the far-end user due to the feedback effect and Lombard effect. To overcome these effects, online feedback cancellation algorithm is implemented to preserve speech signal for the far-end user and active noise control with virtual sensing technology is utilized in a practical smartphone configuration to reduce the noise experienced by the near-end user. Through an experimental implementation, it is demonstrated that this integration of active noise control system with sound masking technology achieves speech privacy without hindering the communication between users over the phone.