Interaktioner - Open Seminars at MID, Fall 2017

Emma Frid
Sonification of Women in Sound and Music Computing - The Sound of Female Authorship in ICMC, SMC and NIME Proceedings
E Frid - KTH Royal Institute of technology, Stockholm, Sweden

Abstract
The primary goal of this study was to approximate the number of female authors in the academic field of Sound and Music Computing. This was done through gender prediction from author names for proceedings from the ICMC, SMC and NIME conferences, and by sonifying these results. Although gender classification by first name can only serve as an estimation of the actual number of female authors in the field, some conclusions could be drawn. The total percentage of author names classified as female was 10.3% for ICMC, 11.9% for SMC and 11.9% for NIME. When merging data from all three conferences for years 2004-2016, it could be concluded that names classified as female ranged from 9.5 to 14.3%. Changes in the ratio of female vs. male authors over time were further illustrated by sonifications, allowing the reader to explore, compare and reflect upon the results by listening to sonic representations of the data. The conclusion that can be drawn from this study is that the field of Sound and Music Computing is still far from being gender-balanced.

Anders Friberg
Cross-cultural aspects of perceptual features in K-Pop: A pilot study comparing Chinese and Swedish listeners
A Friberg, R Schön, A Elowsson - KTH Royal Institute of technology, Stockholm, Sweden
K Choi, J S Downie - University of Illinois Champaign, IL, 61820

Abstract
In previous studies it has been shown that perceptual features can be used as an intermediate representation in music processing to model higher-level semantic descriptions. In this pilot study, we focused on the cross-cultural aspect of such perceptual features, by asking both Chinese and Swedish listeners to rate a set of K-Pop samples using a web-based questionnaire. The music samples were selected from a larger set, previously rated in terms of different emotion labels. The selection procedure of the subset was carefully designed to maximize both the variation of emotion and genre. The listeners rated eight perceptual features: dissonance, speed, rhythmic complexity, rhythmic clarity, articulation, harmonic complexity, modality, and pitch. The results indicated a small but significant difference in the two groups, regarding the average speed and rhythmic complexity. In particular the perceived speed of hip hop was different for the two groups. We discuss the overall consistency of the ratings using this methodology in relation to the interface, selection and number of subjects.