Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2510.12780 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917246548312064 |
|---|---|
| author | Aggazzotti, Cristina Garg, Ashi Cai, Zexin Andrews, Nicholas |
| author_facet | Aggazzotti, Cristina Garg, Ashi Cai, Zexin Andrews, Nicholas |
| contents | Voice anonymization techniques have been found to successfully obscure a speaker's acoustic identity in short, isolated utterances in benchmarks such as the VoicePrivacy Challenge. In practice, however, utterances seldom occur in isolation: long-form audio is commonplace in domains such as interviews, phone calls, and meetings. In these cases, many utterances from the same speaker are available, which pose a significantly greater privacy risk: given multiple utterances from the same speaker, an attacker could exploit an individual's vocabulary, syntax, and turns of phrase to re-identify them, even when their voice is completely disguised. To address this risk, we propose a new approach that performs a contextual rewriting of the transcripts in an ASR-TTS pipeline to eliminate speaker-specific style while preserving meaning. We present results in a long-form telephone conversation setting demonstrating the effectiveness of a content-based attack on voice-anonymized speech. Then we show how the proposed content-based anonymization methods can mitigate this risk while preserving speech utility. Overall, we find that paraphrasing is an effective defense against content-based attacks and recommend that stakeholders adopt this step to ensure anonymity in long-form audio. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_12780 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Content Anonymization for Privacy in Long-form Audio Aggazzotti, Cristina Garg, Ashi Cai, Zexin Andrews, Nicholas Sound Computation and Language Voice anonymization techniques have been found to successfully obscure a speaker's acoustic identity in short, isolated utterances in benchmarks such as the VoicePrivacy Challenge. In practice, however, utterances seldom occur in isolation: long-form audio is commonplace in domains such as interviews, phone calls, and meetings. In these cases, many utterances from the same speaker are available, which pose a significantly greater privacy risk: given multiple utterances from the same speaker, an attacker could exploit an individual's vocabulary, syntax, and turns of phrase to re-identify them, even when their voice is completely disguised. To address this risk, we propose a new approach that performs a contextual rewriting of the transcripts in an ASR-TTS pipeline to eliminate speaker-specific style while preserving meaning. We present results in a long-form telephone conversation setting demonstrating the effectiveness of a content-based attack on voice-anonymized speech. Then we show how the proposed content-based anonymization methods can mitigate this risk while preserving speech utility. Overall, we find that paraphrasing is an effective defense against content-based attacks and recommend that stakeholders adopt this step to ensure anonymity in long-form audio. |
| title | Content Anonymization for Privacy in Long-form Audio |
| topic | Sound Computation and Language |
| url | https://arxiv.org/abs/2510.12780 |