This study investigates the influence of listener native language with respect to talker native language on perception of degree of foreign accent in English. Listeners from native English, Finnish, German and Mandarin backgrounds rated the accentedness of native English, Finnish, German and Mandarin talkers producing a controlled set of English sentences. Results indicate that non-native listeners, like native listeners, are able to classify non-native talkers as foreign-accented, and native talkers as unaccented. However, while non-native talkers received higher accentedness ratings than native talkers from all listener groups, non-native listeners judged talkers with non-native accents less harshly than did native English listeners. Similarly, non-native listeners assigned higher degrees of foreign accent to native English talkers than did native English listeners. It seems that non-native listeners give accentedness ratings that are less extreme, or closer to the centre of the rating scale in both directions, than those used by native listeners.

No abstract

No abstract

No abstract

Grapheme-based models have been proposed for both ASR and TTS as a way of circumventing the lack of expert-compiled pronunciation lexicons in under-resourced languages. It is a common observation that this should work well in languages employing orthographies with a transparent letter-to-phoneme relationship, such as Spanish. Our experience has shown, however, that there is still a significant difference in intelligibility between grapheme-based systems and conventional ones for this language. This paper explores the contribution of different levels of linguistic annotation to system intelligibility, and the trade-off between those levels and the quantity of data used for training. Ten systems spaced across these two continua of knowledge and data were subjectively evaluated for intelligibility.