Response variations in psychoacoustic tests focused on the assessment of tonal office noise

Tonal noise from rotating machinery is known to be a problem in buildings. To guide the design of a comfortable office environment, a proper understanding of people’s annoyance due to exposure to tonal office noise is valuable. Earlier, authors have proposed and verified two annoyance prediction models based on the psychoacoustic subjective tests conducted in an office mock-up, and further developed a software to implement the metric models. In the present work, the focus is shifted to some unexpected response variations observed while analyzing the responses from subjective tests: (1) the learning pattern that occurs at the beginning of the test; (2) the significant differences in the averaged annoyance ratings with respect to the gender and the Noise-Sensitivity-Questionnaire (NoiSeQ) score; (3) the acclimation to sounds: i.e., the annoyance ratings for the 2-min sounds were lower than the ratings for the corresponding 5-s sounds; and (4) the inconsistencies in ratings: i.e., subjects tend to use the annoyance scale differently in different parts of the test and to rate sounds relative to the background level in the room. In the end, several approaches to deal with these response variations and a few recommendations on the psychoacoustic test design are summarized.