Abstract: This letter focuses on asynchronous voice anonymization, wherein machine-discernible speaker attributes in a speech utterance are obscured while human perception is preserved. We propose to ...