Sound quality metric indicators of rotorcraft noise annoyance using multilevel regression analysis

It is hypothesized that sound quality metrics, particularly loudness, sharpness, tonality, impulsiveness, fluctuation strength, and roughness, could all be possible indicators of the reported annoyance to helicopter noise. To test this hypothesis, a psychoacoustic test was recently conducted in which subjects rated their annoyance levels to synthesized helicopter sounds [Krishnamurthy, InterNoise2018, Paper 1338]. After controlling for loudness, linear regression identified sharpness and tonality as important factors in predicting annoyance, followed by fluctuation strength. Current work focuses on multilevel regression techniques in which the regression slopes and intercepts are assumed to take on normal distributions across subjects. The importance of each metric is evaluated one-by-one, and the variation among subjects is evaluated using simple models. Then, more complete models are investigated, which include the combination of selected metrics and random effects. While the conclusions from linear regression analysis are affirmed by multilevel analysis, other important effects emerge. In particular, a random intercept is shown to be more important than a random slope. In this framework, the relative importance of sound quality metrics is re-examined, and the potential for the modeling of human annoyance to helicopter noise based on sound quality metrics is explored.