Sound quality estimation using hearing sensation
Soon-Hyob Kim · The Journal of the Acoustical Society of America · 1998
This paper is on a study of objective sound quality estimation by using human hearing sensation. Until now, the sound quality estimation has been done by subjective methods by people. Unfortunately, it depends on environment and condition of the human ear. So it was difficult to tell exactly how sound quality was. This study was performed to make objective estimation of sounds, regardless of the factors. In this paper, six parameters were defined, which are related to the sound quality: loudness, sharpness, fluctuation strength, roughness, pleasantness, and annoyance, and signal processing methods for them were developed. In this way, a simulator was implemented that measures human hearing sensation. To prove objectiveness of the simulator, the simulator results were compared with subjective test results which were given by people. As a result, the average error rates were 8.1% in loudness, 12% in sharpness, 20% in fluctuation strength, 18.5% in roughness, 5.5% in pleasantness, and 19% in annoyance. The overall average error rate was 13.7%, which showed that the simulator is very useful.