Previous studies examined feeling chills as peak emotional experiences when listening to music. However, it is said qualitatively that feeling tears are another peak emotion to music. We measured chills and tears in response to music, using a questionnaire survey, and the responses were classified into different groups quantitatively. Our results corroborated previous studies, by indicating that chills were related to acoustic features of music. Moreover, we found tears were related to the lyrical content of songs. Then, we developed the Empathy for Lyrics Scale (ELS) to assess the degree of a personʼs empathy with the lyrical content of songs and confirmed the validity of the ELS. Subsequently, we examined the relationship between ELS scores and chills and tears in response to music. Results indicated that empathy, when listeners project their thought to the lyrical content of songs, was significantly related to tears. These results suggest that feeling tears as a peak emotional response to music is a different from chills, and tears were strongly associated with the lyrical contents.

View full abstract

The dynamics plays an important role in the expression of music performance. The dynamics is expressed as the variation of loudness along the temporal stream in music performance. It is desirable to find objective metrics that show good correlation with loudness when the expression of music is examined. Temporally varying sounds audible in our daily life show complex patterns in frequency and sound pressure level. The authors have proposed “mean energy model” for the evaluation of the overall impression of the loudness of temporally varying sounds. It was found that there is a good correlation between the overall impression of loudness and the physical metrics based on the mean energy model. In this study, a method for calculating instantaneous loudness level has been developed and two experiments were conducted in order to examine the applicability of the “mean energy model”. As a result, it was found that instantaneous loudness level shows good correlation with instantaneous loudness impression and that the loudness level calculated on the mean energy model shows good correlation with overall impression of loudness. The results suggest that the mean energy model has the advantages in the correlation with loudness and in the fact that it can be easily measured and has wide applicability.

View full abstract