Emphatic Speech Generation with Conditioned Input Layer and Bidirectional LSTMS for Expressive Speech Synthesis | IEEE Conference Publication | IEEE Xplore