Investigating Different Representations for Modeling and Controlling Multiple Emotions in DNN-based Speech Synthesis

Speech Communication - Netherlands

doi 10.1016/j.specom.2018.03.002