This paper compares selected acoustic and perceptual properties of 12 MTNs including nine specially recorded MTNs that differ in semantic content and the gender of the talkers. Long-term temporal properties (amplitude density functions) of all MTNs were very similar. Long-term average spectra of most of the noises were also similar. However, the talkers' gender had an effect on low-and high-frequency portions of the spectrum. Despite relatively small physical differences between compared MTNs, perceptual data revealed significant differences in annoyance and smaller differences in masking produced by individual MTNs. The MTNs made up of both female and male voices seem to have the greatest masking power Effects of semantic content on acoustic and perceptual properties of MTNs were not observed. Presented conclusions are based on two pilot experiments. Further perceptual studies are needed to confirm reported observations and expand the scope of comparisons.
展开▼
