Automatic identification of a speaker’s native language background may have forensic applications. This paper explores the feasibility of automatic identification of the native language background... Show moreAutomatic identification of a speaker’s native language background may have forensic applications. This paper explores the feasibility of automatic identification of the native language background of a foreign speaker of English, using phonetically interpretable measurements. The production of the ten monophthongs of (American) English by Dutch, Mandarin Chinese and American speakers was used as a test case. Vowel formants F1 (corresponding to articulatory vowel height), F2 (capturing vowel backness and lip rounding) and vowel duration were extracted. Clearly different duration and patterning of the vowels in the vowel space were seen. Automatic classification of the speaker’s native language was 90 percent correct when all acoustic parameters were used as predictors. Language identification was slightly poorer when only formant data were used (85% correct) and substantially poorer – but much better than chance – when only vowel duration was used (60% correct). We conclude that vowel duration provides a weaker cue to foreign-accent identification in English than the spectral properties but that the combination of both information sources yields the best results. Show less
Automatic identification of a speaker’s native language background may have forensic applications. This paper explores the feasibility of automatic identification of the native language background... Show moreAutomatic identification of a speaker’s native language background may have forensic applications. This paper explores the feasibility of automatic identification of the native language background of a foreign speaker of English, using phonetically interpretable measurements. The production of the ten monophthongs of (American) English by Dutch, Mandarin Chinese and American speakers was used as a test case. Vowel formants F1 (corresponding to articulatory vowel height), F2 (capturing vowel backness and lip rounding) and vowel duration were extracted. Clearly different duration and patterning of the vowels in the vowel space were seen. Automatic classification of the speaker’s native language was 90 percent correct when all acoustic parameters were used as predictors. Language identification was slightly poorer when only formant data were used (85% correct) and substantially poorer – but much better than chance – when only vowel duration was used (60% correct). We conclude that vowel duration provides a weaker cue to foreign-accent identification in English than the spectral properties but that the combination of both information sources yields the best results. Show less
We measured F1, F2 and duration of ten English monophthongs produced by American native speakers and by Danish, Norwegian, Swedish, Dutch, Hungarian and Chinese L2 speakers. We hypothesized that (i... Show moreWe measured F1, F2 and duration of ten English monophthongs produced by American native speakers and by Danish, Norwegian, Swedish, Dutch, Hungarian and Chinese L2 speakers. We hypothesized that (i) L2 speakers would approximate the English vowels more closely as the phonological distance between the L2 and English is smaller, and (ii) English vowels of L2 speaker groups will be more similar as the L2s are closer to one another. Comparison of acoustic vowel diagrams and Linear Discriminant Analyses (LDA) confirm the hypotheses, with one exception: Dutch speakers deviate more from L1 English than the Scandinavian groups. The Interlanguage Speech Intelligibility Benefit was convincingly simulated by the LDA. Show less
