Publication detail

Depersonalization of Speech Using Speaker-Specific Transform Based on Long-Term Spectrum

RUJZL, M. SIGMUND, M.

Original Title

Depersonalization of Speech Using Speaker-Specific Transform Based on Long-Term Spectrum

Type

journal article in Web of Science

Language

English

Original Abstract

This paper introduces a novel approach for hiding personal information in speech signals. The proposed approach applied a transform warping function, which is obtained from a long-term linear prediction spectrum individually for each speaker. The depersonalized speech was compared with the often used technique based on vocal tract length normalization. The proposed approach performs wider manipulation of fundamental frequency and provides higher intelligibility by 5% in clean speech and by 8% for signal-to-noise ratio 5 dB. It also significantly alters the derived glottal pulses, making them difficult to use for personality analysis. Speech intelligibility index and glottal pulse distortion are new aspects in the field of voice depersonalization.

Keywords

Speech depersonalization, long-term spectrum, voice transformation, depersonalized speech evaluation

Authors

RUJZL, M.; SIGMUND, M.

Released

15. 12. 2023

Publisher

Czech Technical University in Prague

Location

Prague

ISBN

1210-2512

Periodical

Radioengineering

Year of study

32

Number

4

State

Czech Republic

Pages from

523

Pages to

530

Pages count

8

URL

https://www.radioeng.cz/fulltexts/2023/23_04_0523_0530.pdf

BibTex

@article{BUT186825,
  author="Miroslav {Rujzl} and Milan {Sigmund}",
  title="Depersonalization of Speech Using Speaker-Specific Transform Based on Long-Term Spectrum",
  journal="Radioengineering",
  year="2023",
  volume="32",
  number="4",
  pages="523--530",
  doi="10.13164/re.2023.0523",
  issn="1210-2512",
  url="https://www.radioeng.cz/fulltexts/2023/23_04_0523_0530.pdf"
}