14th International Congress of Phonetic Sciences (ICPhS-14)
San Francisco, CA, USA
This paper describes an experiment aimed at determining whether automatic assessment of second language learners' fluency in spontaneous speech is feasible and whether it differs from automatic fluency assessment in read speech. Spontaneous speech of 60 learners of Dutch was scored for fluency by five raters and was analyzed by means of a continuous speech recognizer to calculate seven quantitative measures of speech quality known to be related to perceived fluency. The results show that automatic assessment of second language learners' fluency in spontaneous speech is feasible, although not all variables suitable for measuring fluency in read speech are as effective in spontaneous speech. In particular, measures that express the rate at which sounds are produced without taking pauses into account appear to be unsuitable for measuring fluency in spontaneous speech. Furthermore, the correlations between machine scores and human ratings are lower for spontaneous speech. Possible explanations are discussed.
Bibliographic reference. Strik, Helmer / Cucchiarini, Catia (1999): "Automatic assessment of second language learners' fluency", In ICPhS-14, 759-762.