Politeness and Voice Quality – The Alternative Method to Measure Aspiration Noise

This paper discusses some problems regarding the measurement of breathiness directly from the acoustic waveform, especially the estimation of aspiration noise found in the high frequency region, which is a prominent feature of breathiness. Klatt and Klatt (1990) suggested the noise rating method for this, which employed the subjective ratings of visual observation of the irregularity of waveforms after they were band-pass filtered around the third formant frequency (F3), so as to quantify the aspiration noise of higher frequency region. However this method heavily depends on individual raters’ subjective observation of the waveform. It is therefore questionable if the ratings are reliable. Since the interest of this study is to examine the correlation between breathiness and politeness, the technical problem of this noise rating method needs to be remedied. This paper proposes an improved technique of quantifying the aspiration noise in the framework of direct waveform measurement. First, jitter and shimmer of band-pass filtered waveforms around the F3 region were measured. Noise ratings, in which raters observe irregularity of the waveforms, were found to be highly correlated with aspiration noise. Therefore, it is natural to assume that either jitter, shimmer, or both, which are ratios of the irregularity of frequency and amplitude, also reflect aspiration noise. Second, in order to consider the interference of harmonics on the waveforms extracted above, the original waveform’s jitter and shimmer were measured as references. Finally, measurements of jitter and shimmer were employed for the comparison of the judgement of politeness and breathiness, as the latter has been suggested to express politeness and to show care. Listeners showed that there is a significant difference of shimmer around the F3 region, between utterances directed to people of both superior and inferior status. From this result, it is reasonable to say that the shimmer in the F3 region is a possible cue when judging politeness.