Perceived Singing Voice Femininity Survey Data

This repository contains the survey's results collected in the paper "Perceived femininity in singing voice: analysis and prediction", accepted at CMMR 2025, London.

The audio segments used for the survey can be found and downloaded from STraDa. The segments used are the 1200 segments from the test set of STraDa.

Files

participant_info.csv

This file contains information about the participants. The columns are:

• username: Unique user ID. x_y represents the participant answered the x-th survey and was numbered as the y-th participant of this survey.
• female/male: Gender of the participant.
• english/french/mandarin/spanish/german: 1 indicates the participant speaks this language, 0 means the opposite.
• 0/20/35/50/65: Age range. 0 represents 0-19, 20 represents 20-34, 35 represents 35-49, etc.
• instruments: Indicates whether the participant plays an instrument or not.

all_answer.csv

This file contains the survey answers. The columns are:

• segment_name: Segment ID, corresponding to the ID used in the test set of STraDa.
• gender/age/language: Gender, age group, and language of the segment's singer. The age groups are divided the same as in participant_info.csv.
• x_y: The user ID, corresponding to the username in participant_info.csv.

Empty bin means the participant knows the singer/song.

Tiny Issue

We reported 126 valid answers in the paper. However, I accidently deleted the answers of the participant 1_1, and could not recover it in any way :(

Citation

If you use this survey data, please cite:

@inproceedings{kong2025perceived,
  title={Perceived femininity in singing voice: analysis and prediction},
  author={Kong, Yuexuan and Tran, Viet-Anh and Hennequin, Romain},
  booktitle={The 17th International Symposium on Computer Music Multidisciplinary Research (CMMR 2025)},
  year={2025}
}