Towards a computational quality model for IP-based audio

This paper proposes a methodology for the development of a computational quality model for IP-based audio. The project mainly focuses on two questions: 1) choice of a quality measurement scale in the computational model and mapping of this scale with existing audio quality measurement metrics, and 2) quantification of encoding and packet loss impairments (only two quality degradation parameters are considered; others will be discussed in future work). The objective PEAQ algorithm in the SEAQ software implementation is used for the parameter quantification and the model accuracy estimation.