Synthetic speaker models using VTLN to improve the performance of children in mismatched speaker conditions for ASR.

D. Rama Sanand, Torbjrn Svendsen

Browse the full Interspeech paper archive.