Loudness-Based Audio Signal Mixing in Automotive Environments

Abstract:
The presentation of navigation announcements and other information signals in noisy environments, such as that in vehicles, requires enhancement methods to achieve optimal speech intelligibility and information audibility. The capability of state-of-the-art vehicle infotainment systems to reproduce a multitude of entertainment signal sources, including a variety of input levels and compressions, makes it difficult to mix information and entertainment signals. This challenge is intensified by the dynamic, loudness dependent binaural intelligibility level difference of the human auditory system. This paper presents an approach to solve these issues by dynamically adapting the mixing levels of information and announcement signals to those of entertainment signals. Smooth adaptation avoids a modulation of the mixed signal, and speech intelligibility is significantly increased. Algorithm and method are depicted also in [1], [2] and [3]. This work was performed at the department "Center of Competence Media", Harman/Becker Automotive Systems, Karlsbad, Germany.