Annoyance of low-level tonal sounds – A penalty model. (February 2019)

• Record Type: Journal Article
• Authors: Hongisto, Valtteri; Saarinen, Pekka; Oliva, David
• Is Part Of: Applied acoustics. Volume 145(2019)
• Page Start: 358
• Page End: 361
• Publication Date: 2019-02
• Subjects: Environmental noise -- Tonal noise -- Narrow-band noise -- Noise penalty -- Psychoacoustics -- Noise annoyance; Acoustical engineering -- Periodicals; Periodicals; 620.2
• DOI: 10.1016/j.apacoust.2018.09.023
• Languages: English
• ISSNs: 0003-682X

Abstract

Highlights: Tonal sounds can be more annoying than non-tonal sounds. Penalty is applied in many countries to consider the annoyance of tonal sounds. Penalty means that a certain dB value is added over the measured dB value. Data from a recent study was used to develop a penalty prediction model. The model needs further validation from future psychoacoustic experiments. Abstract: Environmental noise and building service noise involving tonal components, i.e. distinguishable narrow-band frequencies, can be more annoying than broad-band noise having the same A-weighted level. Noise regulations of many countries involve a fixed penalty which is added to the measured or predicted A-weighted level, if the sound is tonal. Oliva et al. (2017) found that the annoyance penalty due to tonality depends strongly on the tonal frequency and tonal audibility, i.e. the height of the peak. The purpose of our study is to present a mathematical formulation to predict the penalty of tonal sounds when the tonal frequency and tonal audibility are known. The model was based on the psychoacoustic experiment of Oliva et al. They recruited forty subjects to rate the annoyance of 20 tonal sounds being combinations of five tonal frequencies and four levels of tonal audibility. A simple mathematical model was developed to predict, or actually retrodict, the experimental observations of Oliva et al. The correlation coefficient between retrodicted and observed penalty was 0.98. The model can be applied for low-level tonal sounds within 25–4000 Hz. Further psychoacoustic experiments are needed to validate the prediction model.