Conference Papers Year : 2025

Reference-free Adversarial Sex Obfuscation in Speech

Michele Panariello
Massimiliano Todisco
  • Function : Author
  • PersonId : 1581623
Nicholas Evans

Abstract

Sex conversion in speech involves privacy risks from data collection and often leaves residual sex-specific cues in outputs, even when target speaker references are unavailable. We introduce RASO for Reference-free Adversarial Sex Obfuscation. Innovations include a sex-conditional adversarial learning framework to disentangle linguistic content from sex-related acoustic markers and explicit regularisation to align fundamental frequency distributions and formant trajectories with sex-neutral characteristics learned from sex-balanced training data. RASO preserves linguistic content and, even when assessed under a semi informed attack model, it significantly outperforms a competing approach to sex obfuscation.

Fichier principal
Vignette du fichier
publi-8317.pdf (288.5 Ko) Télécharger le fichier
Origin Files produced by the author(s)
Licence

Dates and versions

hal-05251512 , version 1 (12-09-2025)

Licence

Identifiers

  • HAL Id : hal-05251512 , version 1

Cite

Yangyang Qu, Michele Panariello, Massimiliano Todisco, Nicholas Evans. Reference-free Adversarial Sex Obfuscation in Speech. APSIPA 2025, 17th Asia Pacific Signal and Information Processing Association Annual Summit and Conference, Oct 2025, Shangri-la, Singapore. ⟨hal-05251512⟩

Collections

954 View
70 Download

Share

  • More