AnCoGen: Analysis, Control and Generation of Speech with a Masked Autoencoder
This article introduces AnCoGen, a novel method that leverages a masked autoencoder to unify the analysis, control, and generation of speech signals within a single model. AnCoGen can analyze speech by estimating key attributes, such as speaker identity, pitch, content, loudness, signal-to-noise ratio, and clarity index. In addition, it can generate speech from these attributes and allow precise control of the synthesized speech by modifying them. Extensive experiments demonstrated the effectiveness of AnCoGen across speech analysis-resynthesis, pitch estimation, pitch modification, and speech enhancement. Code and audio examples are available online<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup>.
Authors
- Laurent Girin (ORCID: https://orcid.org/0000-0002-9214-8760)
- Samir Sadok (ORCID: https://orcid.org/0009-0007-5956-4133)
- Simon Leglaive (ORCID: https://orcid.org/0000-0002-8219-1298)
- Gaël Richard (ORCID: https://orcid.org/0000-0002-4960-0010)
- Xavier Alameda-Pineda (ORCID: https://orcid.org/0000-0002-5354-1084)
Institutions
- Institut polytechnique de Grenoble (FR)
- Télécom Paris (FR)
- Centre National de la Recherche Scientifique (FR)
- Institut national de recherche en sciences et technologies du numérique (FR)
- Centre Inria de l'Université Grenoble Alpes (FR)
- CentraleSupélec (FR)
- GIPSA-Lab (FR)
- Laboratoire Traitement et Communication de l’Information (FR)
- Université Grenoble Alpes (FR)
Publication Details
- Published
- 2025-03-12
- DOI
- https://doi.org/10.1109/icassp49660.2025.10887856
- Citations
- 2
- Primary Topic
- Speech and Audio Processing
- Type
- article
- Field-Weighted Citation Impact
- 2.97
Funders
- Agence Nationale de la Recherche