The Perceptual Objective Listening Quality Assessment algorithm in telecommunication: Introduction of ITU-T new metrics POLQA

摘要

In order to accomplish speech quality evaluation tasks in an efficient and economical way, objective computational models which simulate human hearing characteristics of speech perception have been intensively researched in the past decades. Several models have already been developed and standardized in telecommunication industry to assess the listening quality of speech signals transmitted through telecommunication network or generated by voice-communication terminals. Among those standardized models, POLQA, short for Perceptual Objective Listening Quality Assessment, is the newest generation objective metrics designed to address the shortcoming and weakness existed in previous published models such as PESQ and PSQM. This paper gives an overview of POLQA characteristics and describes the functional modules utilized in POLQA at the algorithm level. Furthermore the performance of this new model evaluated by standardization group, which reflects the correlation between objective scores outputted by this model to subjective scores, is also shown.