作者: Tino Haderlein , Michael Döllinger , Václav Matoušek , Elmar Nöth
DOI: 10.3109/14015439.2015.1019563
关键词:
摘要: Automatic voice assessment is often performed using sustained vowels. In contrast, speech analysis of read-out texts can be applied to and assessment. recognition prosodic were used find regression formulae between automatic perceptual four criteria. The was trained with 21 men 62 women (average age 49.2 years) tested another set 24 49 (48.3 years), all suffering from chronic hoarseness. They read the text ‘Der Nordwind und die Sonne’ (‘The North Wind Sun’). Five therapists evaluated data on 5-point Likert scales. Ten accuracy measures (features) identified which describe examined Inter-rater correlation within expert group r = 0.63 for criterion ‘match breath sense units’ 0.87 overall quality. Human–machine 0.40 match of...