کد مقاله | کد نشریه | سال انتشار | مقاله انگلیسی | نسخه تمام متن |
---|---|---|---|---|
566142 | 875939 | 2009 | 15 صفحه PDF | دانلود رایگان |
Italian speakers tend to stress the second component of German morphologically complex words such as compounds and prefix verbs even if the first component is lexically stressed. To improve their prosodic phrasing an automatic pronunciation teaching method was developed based on auditory feedback of prosodically corrected utterances in the learners’ own voices. Basically, the method copies contours of F0, local speech rate, and intensity from reference utterances of a German native speaker to the learners’ speech signals. It also adds emphasis to the stress position in order to help the learners better recognise the correct pronunciation and identify their errors. A perception test with German native speakers revealed that manipulated utterances significantly better reflect lexical stress than the corresponding original utterances. Thus, two groups of Italian learners of German were provided with different feedback during a training session, one group with manipulated utterances in their individual voices and the other with correctly pronounced original utterances in the teacher’s voice. Afterwards, both groups produced the same sentences again and German native speakers judged the resulting utterances. Resynthesised stimuli, especially with emphasised stress, were found to be a more effective feedback than natural stimuli to learn the correct stress position. Since resynthesis was obtained without previous segmentation of the learners’ speech signals, this technology could be effectively included in Computer Assisted Language Learning software.
Journal: Speech Communication - Volume 51, Issue 10, October 2009, Pages 933–947