A streaming accent conversion model using an Emformer encoder and synthetic native TTS ground truth matches the quality of a non-streaming model at 0.8 seconds latency.
Our results show that the streaming model achieves performance comparable to its non- streaming counterpart
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Streaming Non-Autoregressive Model for Accent Conversion and Pronunciation Improvement
A streaming accent conversion model using an Emformer encoder and synthetic native TTS ground truth matches the quality of a non-streaming model at 0.8 seconds latency.