System and Method for Sound Translation Based on Generative Adversarial Networks

摘要

Disclosed is an acoustic modification system and method. In this example,When transforming between mutually unpaired sounds, a self-attention module and training based on a characteristic loss function are applied to facilitate the transformation that simulates the target tone, while the pitch, length, silence, and vibration It is an object of the present invention to provide an acoustic transformation system and method in which the reality of the deformed sound is increased by maintaining the same elements.