WavCube is a compressed continuous latent from SSL speech encoders that unifies understanding, reconstruction, and generation through a two-stage semantic bottleneck and acoustic injection process.
Emilia: An extensive, multilingual, and diverse speech dataset for large-scale speech generation
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
eess.AS 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
WavCube: Unifying Speech Representation for Understanding and Generation via Semantic-Acoustic Joint Modeling
WavCube is a compressed continuous latent from SSL speech encoders that unifies understanding, reconstruction, and generation through a two-stage semantic bottleneck and acoustic injection process.