Audiovisual Quality Estimation for Mobile Streaming Services

Abstract

3G mobile terminals supporting audio and video streaming services became reality although the perceptual quality for such low bit rates is limited. To select optimal codec parameters for audio and video, it is important to consider corresponding quality requirements based on human perception. The intention of this paper is to estimate perceptual audiovisual quality for low bit rate videos. We performed subjective perceptual experiments for audio, video and audiovisual quality for music and speech video contents. These experiments were carried out for different audio and video codecs supported by today's mobile terminals. Furthermore, we study the simultaneous influence of music and speech quality on subjective perceptual video quality and mutually compensation effect that occurs in this context. Finally, we propose audiovisual metric for an automated quality estimation.