One aim of the IFCASL project is to construct a corpus of native and non-native speech
for the language pair French-German. Our survey shows that no suitable corpus,
in terms of size and coverage, is currently available for the target language pair.
Our design of such a corpus has three aims:
We will bring to the research community two non-native corpora for the
French-German language pair, whereas most studies have focused on English.
The corpora will be informed by in-depth phonetic knowledge to predict
the types of errors made by French and German learners.
The corpora can further be used by the research community
for the recognition of non-native speech.
The following samples are provided as a preview of the IFCASL corpus, and represent
all combinations of the French-German language pair. Click on a sample below to listen to
the audio, view a screenshot of the annotation (PRAAT textgrid), and learn more about
the phenomena that can be observed in the utterance.