-
1
-
-
85180525521
-
-
The distributions of particular vowels overlapped each other significantly for all three language corpora, reflecting the large perturbations in spectral structure contributed by phonetic and prosodic contextual influences. Variations in both F1 and F2 were large for many of the vowel categories, with the back vowels varying more in F2 values than the front vowels, and mid to low vowels varying more in F1 values than the high vowels
-
The distributions of particular vowels overlapped each other significantly for all three language corpora, reflecting the large perturbations in spectral structure contributed by phonetic and prosodic contextual influences. Variations in both F1 and F2 were large for many of the vowel categories, with the back vowels varying more in F2 values than the front vowels, and mid to low vowels varying more in F1 values than the high vowels.
-
-
-
-
2
-
-
85180449995
-
-
this study, the “fit index” was used to define non-native consonants as “good,” “fair,” or “poor” exemplars of native categories. However, this combined score affected the classification of only one of the consonants (/w/) differently than would have been the case if only the classification consistency were considered. Multiplication of these two kinds of scores assumes that goodness ratings are interval measures, an assumption that this author does not make in her own research
-
In this study, the “fit index” was used to define non-native consonants as “good,” “fair,” or “poor” exemplars of native categories. However, this combined score affected the classification of only one of the consonants (/w/) differently than would have been the case if only the classification consistency were considered. Multiplication of these two kinds of scores assumes that goodness ratings are interval measures, an assumption that this author does not make in her own research.
-
-
-
-
3
-
-
85180445502
-
-
About one-third of the native AE participants tested cannot pass the criterion within 50–100 training trials; their data are, thus, not included in the cross-language analysis. This demonstrates the considerable problems of orthographic representation of vowels in AE that will skew the results of a perceptual assimilation study that includes data from participants who do not demonstrate correct use of the response alternatives for native vowel utterances
-
About one-third of the native AE participants tested cannot pass the criterion within 50–100 training trials; their data are, thus, not included in the cross-language analysis. This demonstrates the considerable problems of orthographic representation of vowels in AE that will skew the results of a perceptual assimilation study that includes data from participants who do not demonstrate correct use of the response alternatives for native vowel utterances.
-
-
-
|
