PhaseAug: A Differentiable Augmentation for Speech Synthesis to Simulate One-to-Many Mapping.

Junhyeok Lee Seungu Han Hyunjae Cho Wonbin Jung

Published in: ICASSP (2023)

Keyphrases

speech synthesis
speech recognition
text to speech
vocal tract
prosodic features
speech corpus
data sets
objective function
data mining
real time
information systems
hidden markov models
language model
non stationary
loss function