On the Relevance of Phoneme Duration Variability of Synthesized Training Data for Automatic Speech Recognition.

Nick Rossenbach, Benedikt Hilmes, Ralf Schlter

Browse the full ASRU paper archive.