Feature Frame Stacking in RNN-Based Tandem ASR Systems - Learned vs. Predefined Context.

Martin Wllmer, Bjrn W. Schuller, Gerhard Rigoll

Browse the full Interspeech paper archive.