Describing Gender Equality in French Audiovisual Streams with a Deep Learning Approach

Abstract

A large-scale description of men and women speaking-time in media is presented, based on the analysis of about 700.000 hours of French audiovisual documents, broadcasted from 2001 to 2018 on 22 TV channels and 21 radio stations. Speaking-time is described using Women Speaking Time Percentage (WSTP), which is estimated using automatic speaker gender detection algorithms, based on acoustic machine learning models. WSTP variations are presented across channels, years, hours, and regions. Results show that men speak twice as much as women on TV and on radio in 2018, and that they used to speak three times longer than women in 2004. We also show only one radio station out of the 43 channels considered is associated to a WSTP larger than 50%. Lastly, we show that WSTP is lower during high-audience time-slots on private channels. This work constitutes a massive gender equality study based on the automatic analysis of audiovisual material and offers concrete perspectives for monitoring gender equality in media.The software used for the analysis has been released in open-source, and the detailed results obtained have been released in open-data.

An Open-Source Speaker Gender Detection Framework for Monitoring Gender Equality

“…This positive outcome allows considering the reliability of this gender detection system sufficient to perform large-scale gender equality descriptions, offering concrete perspectives for digital humanities. Detailed analyses realized using this framework goes beyond the scope of this paper and will be addressed in future works [24]. Early results obtained using 1 week of raw stream corresponding to 6 French TV channels and 8 radio stations highlighted several interesting tendencies: Speaking time was mainly attributed to male speaker, especially in radio material.…”

“…Feature extraction was realized on audio excerpts sampled at 16KHz using SIDEKIT [18]. 24 Mel-scaled filter-banks coefficients were computed on 25ms sliding windows with a 10ms shift. Those coefficients were fed directly to CNN models while a second set of acoustic features (MFCC) was obtained by applying Discrete Cosine Transform (DCT) to those filterbank coefficients to extract 19 Mel-Frequency Cepstrums coefficients (MFCC), which were augmented with log energy.…”

Gender bias in video game dialogue

“…Studies of the gender distribution of fictional dialogue show that male characters speak more than female characters in films, television shows, radio shows, plays and books [1][2][3][4][5][6][7][8][9]. This inequality in representation perpetuates stereotypes that can negatively affect audiences' conceptions of gender and gender roles [10,11].…”

Parole superposée et genre, étude des annotations pour les médias audiovisuels.

“…2 Parole superposée et genre : corpus et usages En informatique, la détection du genre est une tâche de classification qui consiste à évaluer si les caractéristiques acoustiques d'une voix sont plutôt proches de celles d'un homme ou d'une femme, ou éventuellement d'une troisième catégorie non identifiable (enfants, locuteurs non nommés). L'obtention des temps de parole en fonction des genres permet d'automatiser les analyses manuelles des sociologues et ainsi les étendre à des données massives (Doukhan et al, 2018). La détection de genre est également utile en prétraitement des systèmes de traitement automatique de la parole, pour permettre d'affiner les modèles au genre du locuteur et ainsi améliorer les performances.…”

“…Ces résultats montrent que la proportion de femmes et d'hommes dans les corpus étudiés n'est pas équilibrée. Cette observation est commune à plusieurs études(Doukhan et al, 2018) et induit un biais de donnée pour tout système entraîné sur celles-ci(Garnerin et al, 2019). Ce phénomène est amplifié par le fait que ces corpus précèdent les réglementations françaises sur l'égalité entre les femmes et les hommes 1 .On notera qu'un tiers des données des corpus ESTER2 et ETAPE ne disposent pas d'information sur le genre.…”