Daniel Müllensiefen

The musicality of non-musicians: an index for assessing musical sophistication in the general population.

Musical skills and expertise vary greatly in Western societies. Individuals can differ in their repertoire of musical behaviours as well as in the level of skill they display for any single musical behaviour. The types of musical behaviours we refer to here are broad, ranging from performance on an instrument and listening expertise, to the ability to employ music in functional settings or to communicate about music. In this paper, we first describe the concept of ‘musical sophistication’ which can be used to describe the multi-faceted nature of musical expertise. Next, we develop a novel measurement instrument, the Goldsmiths Musical Sophistication Index (Gold-MSI) to assess self-reported musical skills and behaviours on multiple dimensions in the general population using a large Internet sample (n = 147,636). Thirdly, we report results from several lab studies, demonstrating that the Gold-MSI possesses good psychometric properties, and that self-reported musical sophistication is associated with performance on two listening tasks. Finally, we identify occupation, occupational status, age, gender, and wealth as the main socio-demographic factors associated with musical sophistication. Results are discussed in terms of theoretical accounts of implicit and statistical music learning and with regard to social conditions of sophisticated musical engagement.

The Role of Expectation and Probabilistic Learning in Auditory Boundary Perception: A Model Comparison

Grouping and boundary perception are central to many aspects of sensory processing in cognition. We present a comparative study of recently published computational models of boundary perception in music. In doing so, we make three contributions. First, we hypothesise a relationship between expectation and grouping in auditory perception, and introduce a novel information-theoretic model of perceptual segmentation to test the hypothesis. Although we apply the model to musical melody, it is applicable in principle to sequential grouping in other areas of cognition. Second, we address a methodological consideration in the analysis of ambiguous stimuli that produce different percepts between individuals. We propose and demonstrate a solution to this problem, based on clustering of participants prior to analysis. Third, we conduct the first comparative analysis of probabilistic-learning and rule-based models of perceptual grouping in music. In spite of having only unsupervised exposure to music, the model performs comparably to rule-based models based on expert musical knowledge, supporting a role for probabilistic learning in perceptual segmentation of music.

How do earworms start? Classifying the everyday circumstances of Involuntary Musical Imagery

Involuntary Musical Imagery (INMI) or “earworms” describes the experience whereby a tune comes into the mind and repeats without conscious control. The present article uses an inductive, generative, grounded theory-based qualitative analysis to classify reports of everyday INMI circumstances, and creates graphical models that determine their relative frequency within two population samples; listeners to the BBC radio station 6 Music and an online survey. Within the two models, four abstract categories were defined that described the characteristics of the circumstances surrounding the onset of INMI episodes; Music exposure, Memory triggers, Affective states, and Low attention states respectively. We also note the variety of musical media by which exposure to a tune results in an INMI episode and discuss the impact of musical engagement on INMI experiences. The findings of the present study are considered within a framework of involuntary retrieval theory from both the autobiographical and semantic memory literatures. In addition, the results highlight the potential facilitative effects of varying affective and attentional states on INMI episodes.

Effects of timbre and tempo change on memory for music

We investigated the effects of different encoding tasks and of manipulations of two supposedly surface parameters of music on implicit and explicit memory for tunes. In two experiments, participants were first asked to either categorize instrument or judge familiarity of 40 unfamiliar short tunes. Subsequently, participants were asked to give explicit and implicit memory ratings for a list of 80 tunes, which included 40 previously heard. Half of the 40 previously heard tunes differed in timbre (Experiment 1) or tempo (Experiment 2) in comparison with the first exposure. A third experiment compared similarity ratings of the tunes that varied in timbre or tempo. Analysis of variance (ANOVA) results suggest first that the encoding task made no difference for either memory mode. Secondly, timbre and tempo change both impaired explicit memory, whereas tempo change additionally made implicit tune recognition worse. Results are discussed in the context of implicit memory for nonsemantic materials and the possible differences in timbre and tempo in musical representations.

Classification in music research

Since a few years, classification in music research is a very broad and quickly growing field. Most important for adequate classification is the knowledge of adequate observable or deduced features on the basis of which meaningful groups or classes can be distinguished. Unsupervised classification additionally needs an adequate similarity or distance measure grouping is to be based upon. Evaluation of supervised learning is typically based on the error rates of the classification rules. In this paper we first discuss typical problems and possible influential features derived from signal analysis, mental mechanisms or concepts, and compositional structure. Then, we present typical solutions of such tasks related to music research, namely for organization of music collections, transcription of music signals, cognitive psychology of music, and compositional structure analysis.

Melodic Grouping in Music Information Retrieval: New Methods and Applications

We introduce the MIR task of segmenting melodies into phrases, summarise the musicological and psychological background to the task and review existing computational methods before presenting a new model, IDyOM, for melodic segmentation based on statistical learning and information-dynamic analysis. The performance of the model is compared to several existing algorithms in predicting the annotated phrase boundaries in a large corpus of folk music. The results indicate that four algorithms produce acceptable results: one of these is the IDyOM model which performs much better than naive statistical models and approaches the performance of the best-performing rule-based models. Further slight performance improvement can be obtained by combining the output of the four algorithms in a hybrid model, although the performance of this model is moderate at best, leaving a great deal of room for improvement on this task.

Absolute memory for pitch: A comparative replication of Levitin’s 1994 study in six European labs

In a widely cited study, Levitin (1994) suggested the existence of absolute pitch memory for music in the general population beyond the rare trait of genuine absolute pitch (AP). In his sample, a significant proportion of non-AP possessors were able to reproduce absolute pitch levels when asked to sing very familiar pop songs from memory. Forty-four percent of participants sang the correct pitch on at least one of two trials, and 12% were correct on both trials. However, until now, no replication of this study has ever been published. The current paper presents the results of a large replication endeavour across six different labs in Germany and the UK. All labs used the same methodology, carefully replicating Levitin’s original experiment. In each lab, between 40 and 50 participants were tested (N = 277). Participants were asked to sing two different pop songs of their choice. All sung productions were compared to the original songs. Twenty-five percent of the participants sang the exact pitch of at least one of the two chosen songs and 4% hit the right pitches for both songs. Our results generally confirm the findings of Levitin (1994). However, the results differ considerably across laboratories, and the estimated overall effect using meta-analysis techniques was significantly smaller than Levitin’s original result. This illustrates the variability of empirical findings derived from small sample sizes and corroborates the need for replication and meta-analytical studies in music psychology in general.

Der Gold-MSI: Replikation und Validierung eines Fragebogeninstrumentes zur Messung Musikalischer Erfahrenheit anhand einer deutschen Stichprobe

The present study introduces the German version of the Gold-MSI inventory, a tool for evaluating self-reported musical abilities and musical expertise. The Gold-MSI is based around the multidimensional construct of Musical Sophistication and builds on the idea that musical expertise cannot only be developed through musical training on an instrument but also through active engagement with music in its many facets. The questionnaire was developed with a very large English sample (Müllensiefen et al., 2014) and comprises musical expertise with five factors as well as the general factor Musical Sophistication. The English Gold-MSI questionnaire was translated into German and evaluated with a German sample (N = 641). Using confirmative factor analysis the underlying factor structure was confirmed. Furthermore, the results show high reliabilities of the five sub-factors as well as the general factor Musical Sophistication (Cronbach’s alpha between .72 and .91.). Additionally, relationships between variables of the socio-economic status and the sub-factors of the Gold-MSI of the German sample are investigated using a structural equation model. The statistical model reveals positive relationships between income and professional status on the one hand and musical training, perceptual abilities and emotional engagement with music on the other hand. The inventory is freely available and is designed to contribute to the refined investigation of musical sophistication and expertise in German speaking countries.

Modelling experts notions of melodic similarity

In this article we show that a subgroup of music experts has a reliable and consistent notion of melodic similarity, and that this notion can be measured with satisfactory precision. Our measurements enable us to model the similarity ratings of music experts by automated and algorithmic means. A large number of algorithmic similarity measure found in the literature were mathematically systematised and implemented. The best similarity algorithms compared to human experts were chosen and optimised by statistical means according to different contexts. A multidimensional scaling model of the algorithmic similarity measures is constructed to give an overiew over the different musical dimensions reflected by these measures. We show some examples where this optimised methods could be successfully applied to real world problems like folk song categorisation and analysis, and discuss further applications and implications.

On the non-existence of music: Why music theory is a figment of the imagination

We argue for an approach to the theory of music which starts from the position that music is primarily a construct of human minds (and secondarily a social construct) and contrast it with the approach implicit in the work of some music theorists, which treats music as though it were an externally defined quasi-Platonic absolute. We argue that a natural conclusion of this approach is that music theory, while already being a kind of folk psychology, can benefit from being more explicitly informed by music cognition studies. We give examples from work in the computational modelling of music cognition, following our approach, which attempts to place each musical phenomenon in an ecological context motivated by evolutionary considerations, and which aims to explain musical phenomena independently of the explicit intervention of the theorist. We argue that only thus can a theory be said veridically to explicate the phenomenology of music. We place our argument in context of the Generative Theory of Tonal Music (Lerdahl & Jackendoff, 1983), Generative Linguistics, and other papers in the current volume, and compare them all with results of modelling studies based on our espoused approach.