Preprocessed data for the SLTU experiments
There is one directory for each language and speaker we consider. Each directory contains two subdirectories: one with the wav file, the other with the transcriptions. Transcriptions are not segmnented (except for Na).
All data have been extracted from the Pangloss collection.