The IFCASL Corpus of French and German Non-native and Native Read Speech

The IFCASL Corpus of French and German Non-native and Native Read Speech Jürgen Trouvain1, Anne Bonneau2, Vincent Colotte2, Camille Fauth2,3, Dominique Fohr2, Denis Jouvet2, Jeanin Jügler1, Yves Laprie2, Odile Mella2, Bernd Möbius1, Frank Zimmerer1 1Computational Linguistics and Phonetics, Saarland University, Germany; 2Speech Group, LORIA (Inria; Université de Lorraine, LORIA, UMR 7503; CNRS, LORIA, UMR 7503) Villers-lès-Nancy, France; 3Phonetics, Strasbourg University, France Corresponding author: trouvain [at] coli.uni-saarland.de Abstract The IFCASL corpus is a French-German bilingual phonetic learner corpus designed, recorded and annotated in a project on individualized feedback in computer-assisted spoken language learning. The motivation for setting up this corpus was that there is no phonetically annotated and segmented corpus for this language pair of comparable of size and coverage. In contrast to most learner corpora, the IFCASL corpus incorporate data for a language pair in both directions, i.e. in our case French learners of German, and German learners of French. In addition, the corpus is complemented by two sub-corpora of native speech by the same speakers. The corpus provides spoken data by about 100 speakers with comparable productions, annotated and segmented on the word and the phone level, with more than 50% manually corrected data. The paper reports on inter-annotator agreement and the optimization of the acoustic models for forced speech-text alignment in exercises for computer-assisted pronunciation training. Example studies based on the corpus data with a phonetic focus include topics such as the realization of /h/ and glottal stop, final devoicing of obstruents, vowel quantity and quality, pitch range, and tempo. Keywords: learner corpus, phonetics, French, German, non-native speech, native speech 1. Introduction The IFCASL corpus is a French-German bilingual phonetic learner corpus designed, recorded and annotated in the project IFCASL (Individualized Feedback in Com- puter-Assisted Spoken Language Learning <www.ifcasl. org>) (Trouvain et al. 2013; Fauth et al. 2014). 1.1 Motivation and aims for the corpus The motivation for setting up this corpus was that there is no phonetically annotated and segmented corpus for this language pair of comparable of size and coverage. Generally speaking, most learner corpora exist for the written language and the majority of spoken learner corpora has English as the target language. In addition, only few learner corpora incorporate data for a language pair in both directions, i.e. in our case French learners of German, and German learners of French. To our knowledge existing spoken learner corpora for the French-German language pair are restricted to the HABLA Corpus (Hamburg Adult Bilingual LAnguage), with recordings of early French and German bilinguals (Kupisch al. 2012), and the German part of the phonological platform "Interphonologie du Français Contemporain" (IPFC-allemand), with well advanced German learners of French (Chervinski & Pustka 2010). The purpose of the IFCASL corpus is to have a reliable empirical foundation to investigate phonetic and phonological deviations of both learner groups. In the past these aspects were either based on personal and anecdotal experience or on purely theoretical assumptions based on contrastive comparisons of the phonological systems. The aims for constructing this corpus are as follows: …

Content truncated to 500 words