Parametric audio quality model for IPTV services - ITU-T P.1201.2 audio

This paper describes a parametric audio quality model for IPTV services. The model is standardized under the ITU-T P.1201.2 Recommendation. Subjective test databases used to develop and validate the model are also presented. The model is applicable to network planning and packet-layer quality monitoring of encrypted audio. Addressed degradations are audio compression (MPEG-1 Layer II & III, MPEG-2 AAC-LC, MPEG-4 HE-AACv2 and AC-3) and audio frame loss for various frame-loss-patterns. Two implementations (Dolby, Nero) have been tested for the AAC-LC and HE-AAC codecs. The proposed model takes as input parameters the audio codec, bitrate, frame loss percentage and a burst-length-related parameter. It shows high performance results on unknown data with a Pearson´s correlation coefficient of 0.92 with the subjective quality scores and an RMSE of 0.39 on the 5-point scale.