Voice Quality Modelling for Expressive Speech Synthesis

Table 2

Average confusion matrix (%) for the subjective test (the maximum correct classification value is in bold).

(%)	NEU	HAP	SEN	AGG	SAD	Dk/A
NEU	𝟖𝟔.𝟒	1.3	3.6	5.3	0.7	2.7
HAP	1.9	𝟖𝟏.𝟎	0.2	15.6	0.1	1.2
SEN	4.7	0.1	𝟖𝟔.𝟖	0.0	5.7	2.6
AGG	1.8	14.2	0.1	𝟖𝟐.𝟕	0.1	1.1
SAD	0.5	0.0	0.6	0.0	𝟗𝟖.𝟖	0.1