Perceptions of Diversity in Electronic Music: the Impact of Listener, Artist, and Track Characteristics

Abstract

Shared practices to assess the diversity of retrieval system results are still debated in the Information Retrieval community, partly because of the challenges of determining what diversity means in specific scenarios, and of understanding how diversity is perceived by end-users. The field of Music Information Retrieval is not exempt from this issue. Even if fields such as Musicology or Sociology of Music have a long tradition in questioning the representation and the impact of diversity in cultural environments, such knowledge has not been yet embedded into the design and development of music technologies. In this paper, focusing on electronic music, we investigate the characteristics of listeners, artists, and tracks that are influential in the perception of diversity. Specifically, we center our attention on 1) understanding the relationship between perceived diversity and computational methods to measure diversity, and 2) analyzing how listeners' domain knowledge and familiarity influence such perceived diversity. To accomplish this, we design a user-study in which listeners are asked to compare pairs of lists of tracks and artists, and to select the most diverse list from each pair. We compare participants' ratings with results obtained through computational models built using audio tracks' features and artist attributes. We find that such models are generally aligned with participants' choices when most of them agree that one list is more diverse than the other, while they present a mixed behaviour in cases where participants have little agreement. Moreover, we observe how differences in domain knowledge, familiarity, and demographics can influence the level of agreement among listeners, and between listeners and diversity metrics computed automatically.

TROMPA-MER: an open dataset for personalized music emotion recognition

“…Particularly, the proposed HC and MIX methods appear to significantly outperform the random baseline for SGD and XGB models. This finding leads us to promote the creation of subjectivity-aware machine learning methods which could have a high impact in novel applications of immersion in virtual reality (Warp et al, 2022) and emotion-based music recommendation (Grekow, 2021;Tarnowska, 2021)-several other tasks display low inter-rater agreement too: music auto-tagging (Bigand & Aucouturier, 2013), music similarity and diversity (Flexer et al, 2021;Porcaro et al, 2022), automatic chord estimation (Koops et al, 2019), and beat tracking (Holzapfel et al, 2012). In short, MER has been openly criticized due to the subjectivity issue (Gómez-Cañónet al, 2021) -however we advocate for "embracing subjectivity and potentially leveraging the opportunities it offers for better learning" (Rizos & Schuller, 2020).…”

TROMPA-MER: an open dataset for personalized Music Emotion Recognition

“…Particularly, the proposed HC and MIX methods appear to significantly outperform the random baseline for SGD and XGB models. This finding leads us to promote the creation of subjectivityaware machine learning methods -several other tasks have low inter-rater agreement too (music auto-tagging [65], music similarity and diversity [66,67], automatic chord estimation [68], and beat tracking [69]). In short, MER has been openly criticized due to the subjectivity issue [6] -however we advocate for "embracing subjectivity and potentially leveraging the opportunities it offers for better learning" [70].…”

Intra-list similarity and human diversity perceptions of recommendations: the details matter

“…While several works have addressed similarity perception, only a few have specifically addressed the diversity perception of lists. In the music domain (specifically, electronic music), for example, Porcaro et al (2022) found that instrument and samples, subgenre or sub-style, tempo, and mood strongly influence what track lists were considered diverse. On the artist level, the artists' origin and nationality, gender, and skin tone were considered key factors for diverse music lists.…”

“…First, we addressed whether the familiarity of items influences the perception of a list. Similar to Porcaro et al (2022), where domain knowledge played a role in diversity perception, participants without background knowledge regarding a specific set of items might perceive the list differently (e.g., more diverse) than participants with more knowledge. Second, we addressed whether different diversity levels impacted the participants' decision processes, e.g., in terms of their perceived choice difficulty or choice confidence, cf.…”

“…Lastly, we also investigated the effects of domain knowledge on the diversity perception of a list. Based on prior research conducted by Porcaro et al (2022), we might assume that participants with differing levels of background knowledge may perceive a list of items differently. On the one hand, users with more background knowledge might find more aspects in which items are different.…”