Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Evolution of Voices in French Audiovisual Media Across Genders and Age in a Diachronic Perspective

Albert Rilliard, David Doukhan, Rémi Uro, Simon Devauchelle

TL;DR

This work investigates diachronic changes in French broadcast voice across gender and age by building a large corpus of 1023 speakers from four time periods. It estimates base-$F_0$ and a vocal tract length proxy from the first four formants using a robust, multi-estimator pipeline on both raw and separated signals, then analyzes long-term measures with linear mixed models that control for age. Key findings show a period effect toward lower voices that is largely independent of gender, and an age-related decrease in base-$F_0$ for female voices only; VTL increases modestly in later periods and is strongly influenced by gender. The study advances understanding of sociophonetic dynamics in media while highlighting methodological challenges in extracting robust acoustic measures from heterogeneous archival data.

Abstract

We present a diachronic acoustic analysis of the voice of 1023 speakers from French media archives. The speakers are spread across 32 categories based on four periods (years 1955/56, 1975/76, 1995/96, 2015/16), four age groups (20-35; 36-50; 51-65, >65), and two genders. The fundamental frequency ($F_0$) and the first four formants (F1-4) were estimated. Procedures used to ensure the quality of these estimations on heterogeneous data are described. From each speaker's $F_0$ distribution, the base-$F_0$ value was calculated to estimate the register. Average vocal tract length was estimated from formant frequencies. Base-$F_0$ and vocal tract length were fit by linear mixed models to evaluate how they may have changed across time periods and genders, corrected for age effects. Results show an effect of the period with a tendency to lower voices, independently of gender. A lowering of pitch is observed with age for female but not male speakers.

Evolution of Voices in French Audiovisual Media Across Genders and Age in a Diachronic Perspective

TL;DR

This work investigates diachronic changes in French broadcast voice across gender and age by building a large corpus of 1023 speakers from four time periods. It estimates base- and a vocal tract length proxy from the first four formants using a robust, multi-estimator pipeline on both raw and separated signals, then analyzes long-term measures with linear mixed models that control for age. Key findings show a period effect toward lower voices that is largely independent of gender, and an age-related decrease in base- for female voices only; VTL increases modestly in later periods and is strongly influenced by gender. The study advances understanding of sociophonetic dynamics in media while highlighting methodological challenges in extracting robust acoustic measures from heterogeneous archival data.

Abstract

We present a diachronic acoustic analysis of the voice of 1023 speakers from French media archives. The speakers are spread across 32 categories based on four periods (years 1955/56, 1975/76, 1995/96, 2015/16), four age groups (20-35; 36-50; 51-65, >65), and two genders. The fundamental frequency () and the first four formants (F1-4) were estimated. Procedures used to ensure the quality of these estimations on heterogeneous data are described. From each speaker's distribution, the base- value was calculated to estimate the register. Average vocal tract length was estimated from formant frequencies. Base- and vocal tract length were fit by linear mixed models to evaluate how they may have changed across time periods and genders, corrected for age effects. Results show an effect of the period with a tendency to lower voices, independently of gender. A lowering of pitch is observed with age for female but not male speakers.
Paper Structure (12 sections, 2 figures, 1 table)

This paper contains 12 sections, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. Methods
  3. INA's diachronic corpus
  4. Voicing, $F_0$ and Formant estimation
  5. Acoustic features filtering
  6. Long-term features
  7. Statistical analysis
  8. Results
  9. Base-$F_0$
  10. Vocal tract length
  11. Discussion & Conclusion
  12. Acknowledgements

Figures (2)

  • Figure 1: Fit of base-$F_0$ (in semitones) from the Speaker Age and Gender.
  • Figure 2: Fit of base-$F_0$ (semitones) for speaker Age by Period and separated by Gender (males: dashed lines) to show the relative effect of the Period slopes on voices of each gender.