The impulsive components of printer sounds can affect users’ comfort. Thus, to develop a psychoacoustic model to evaluate printer noise quality, assessment of printer impulsive noise needs to be addressed. Previously, a method to generate synthetic printer noises having varying levels of different sound attributes, such as loudness and sharpness, was developed. In that way, 233 sounds were generated based on measurements taken from three different printers. Each simulated sound contained broadband noise components and one impulse event, while audible tones and other events were removed. The simulated sounds were divided into six groups: within each group, the sound level, spectral balance and rise time of the impulse were varied from sound-to-sound while the background noise was maintained constant. The sounds were presented to 35 subjects who rated their annoyance given a description of the environment in which they would hear the sounds. The relationship between the subjects’ annoyance scores and sound quality metric values for each of the sounds was examined. The results showed that a loudness metric based on near maximum loudness levels predicted most of the response variation, but inclusion of a sharpness metric in the annoyance model can significantly improve the prediction accuracy.
Yu et al. (Wed,) studied this question.