Thirty subjects with equal numbers of experienced and naive subjects performed the repeatability test of annoyance judgments at two different sessions between 10 and 20 days. Ten stereophonic recorded sound stimuli with various speeds and loads were randomly presented to the subjects through a pair of loudspeakers in both sessions. The results showed insignificant difference regarding annoyance judgments of engine sounds between the listening tests. Annoyance ratings were higher for naive subjects than for experienced subjects. Principal Component Analysis and Partial Least Squares Regression were applied to analyze the data with respect to annoyance responses of sound stimuli and 27 psychoacoustic descriptors. Three descriptors were found to be significant in predicting annoyance response of engine sounds and they are known as loudness, sharpness (acum), and harmonic ratio (rumble noise). The result shows no significant difference in annoyance modeling for engine sounds in free field conditions due to different listening test sessions.
展开▼
