Self-Supervised Speech Representations Preserve Speech Characteristics while Anonymizing Voices | Publicación