Sound Quality Evaluation Based on Attributes - Application to Binaural Contents
Abstract
The audio quality assessment is based on standards which mainly evaluate the overall quality to the detriment of more accurate sound criteria. On the other hand, a major problem of an assessment based on sound criteria is their meaning and their understanding that have to be the same for each listener. A previous study clustered a list of sound attributes in three main categories called 'timbre', 'space', 'defaults'. The work presented here is based on those previous results and aims at tuning a subjective test methodology of spatial audio quality. So the three families were included in a test dedicated to the assessment of spatial audio quality with binaural contents. The test was based on the MUSHRA method but using three anchors specifically to each attribute and without explicit reference. The original version was added as the hidden reference. The aim of the listening test described in this paper was to verify the relevance of those 3 attributes and their influence on the overall quality.