no code implementations • 27 Nov 2023 • Anna Leschanowsky, Ünal Ege Gaznepoglu, Nils Peters
This study investigates bias in voice anonymization systems within the context of the Voice Privacy Challenge.
no code implementations • 22 Aug 2023 • Ünal Ege Gaznepoglu, Nils Peters
Speaker anonymization systems continue to improve their ability to obfuscate the original speaker characteristics in a speech signal, but often create processing artifacts and unnatural sounding voices as a tradeoff.
no code implementations • 29 Jun 2023 • Ünal Ege Gaznepoglu, Nils Peters
In a deep learning setting, this is achieved by extracting multiple features from speech, altering the speaker identity, and waveform synthesis.
no code implementations • 31 Oct 2022 • Ünal Ege Gaznepoglu, Anna Leschanowsky, Nils Peters
We introduce a novel method to improve the performance of the VoicePrivacy Challenge 2022 baseline B1 variants.
no code implementations • 9 Dec 2021 • Nagashree K. S. Rao, Nils Peters
This paper explores the effect of perceptual audio coding on the classification performance using a DCASE 2020 challenge contribution [1].
no code implementations • 13 Oct 2021 • Ünal Ege Gaznepoglu, Nils Peters
Voice conversion for speaker anonymization is an emerging field in speech processing research.