Combining unidirectional long short-term memory with convolutional output layer for high-performance speech synthesis | IEEE Conference Publication | IEEE Xplore