A linguistic and psychoacoustic study for questionnaire analysis: first results

This research is inspired by the consideration that language represents a significant dimension of perception and description of the soundscape. A language is a strategic tool that can consistently describe the qualities of physical acoustic environments, introducing an efficient metric to describe the sound perception by people. In this paper, we explore the possibility of using sentiment analysis for extracting the emotional impact of noise from lists of adjectives that describe sounds. To address this, an investigation campaign was conducted to identify the Italian adjectives that best describe the noise sources associated with port and retro-port infrastructures. 402 Italian university students responded to a listening experiment. They were asked to associate adjectives with audios. The association of these adjectives with the sources of sound formed the basis for sentiment analysis. Psychoacoustic parameters were calculated following measurements and processed for each of the sounds administered to look for possible correlations between the different perceptual aspects. The first results show how industrial or traffic sounds are as expected associated with negative sentiment whereas natural sounds evoke positive emotions. The proposed sentiment analysis, in connection with the psychoacoustic investigations developed, provides a framework for future research in the investigation of sound perception.