Figure 13.
Spectrogram of the Mandarin phrase “比麻黎伪善” /bǐ má lí wěi shàn/ [more hypocritical than Ma Li] [155], with broad phonetic transcriptions and pitch tracts (in blue). In both panels, C, V, and T stand for consonant, vowel, and tone, respectively. In (A), the segmentation is conventional [63,64]. The segmentation of /w/ is based on [65]. In (B) the segmentation is based on the synchronization hypothesis. In (C) the segmentation is based on the diphone principle [202,203].