Adapting General Disentanglement-Based Speaker Anonymization for Enhanced Emotion Preservation

TL;DR

This paper enhances speaker anonymization systems to better preserve emotional cues by integrating emotion embeddings and applying an emotion compensation post-processing step, balancing privacy and emotional fidelity.

Contribution

It introduces two novel strategies—using pre-trained emotion embeddings and an emotion compensation method—to improve emotion preservation in speaker anonymization.

Findings

Emotion embeddings aid in preserving emotional cues.

Emotion compensation improves emotional fidelity with slight privacy trade-offs.

Strategies are adaptable for other paralinguistic attributes.

Abstract

A general disentanglement-based speaker anonymization system typically separates speech into content, speaker, and prosody features using individual encoders. This paper explores how to adapt such a system when a new speech attribute, for example, emotion, needs to be preserved to a greater extent. While existing systems are good at anonymizing speaker embeddings, they are not designed to preserve emotion. Two strategies for this are examined. First, we show that integrating emotion embeddings from a pre-trained emotion encoder can help preserve emotional cues, even though this approach slightly compromises privacy protection. Alternatively, we propose an emotion compensation strategy as a post-processing step applied to anonymized speaker embeddings. This conceals the original speaker's identity and reintroduces the emotional traits lost during speaker embedding anonymization. Specifically, we model the emotion attribute using support vector machines to learn separate boundaries for each emotion. During inference, the original speaker embedding is processed in two ways: one, by an emotion indicator to predict emotion and select the emotion-matched SVM accurately; and two, by a speaker anonymizer to conceal speaker characteristics. The anonymized speaker embedding is then modified along the corresponding SVM boundary towards an enhanced emotional direction to save the emotional cues. The proposed strategies are also expected to be useful for adapting a general disentanglement-based speaker anonymization system to preserve other target paralinguistic attributes, with potential for a range of downstream tasks.