PITCH SYNCHRONIZATION VOICE SYNTHESIS METHOD FOR DEEP LEARNING-BASED VOICE SYNTHESIS SYSTEM

摘要

Provided is a voice synthesis method which can improve a voice quality of a synthesis sound with deep learning-based voice synthesis by synchronizing a pitch according to section properties to set variable voice analysis/synthesis section and a moving section which are an integer of the pitch so as to improve the performance of the synthesis sound in a voice synthesis system using text-voice conversion of a small mobile device or the like.;COPYRIGHT KIPO 2017