The comparison of world music cultures has been considered in musicological research since the end of the 19th century. Traditional methods from the field of comparative musicology typically involve the process of manual music annotation. While this provides expert knowledge, the manual input is timeconsuming and limits the potential for large-scale research. This thesis considers computational methods for the analysis and comparison of world music cultures. In particular, Music Information Retrieval (MIR) tools are developed for processing sound recordings, and data mining methods are considered to study similarity relationships in world music corpora. MIR tools have been widely used for the study of (mainly) Western music. The first part of this thesis focuses on assessing the suitability of audio descriptors for the study of similarity in world music corpora. An evaluation strategy is designed to capture challenges in the automatic processing of world music recordings and different state-of-the-art descriptors are assessed. Following this evaluation, three approaches to audio feature extraction are considered, each addressing a different research question. First, a study of singing style similarity is presented. Singing is one of the most common forms of musical expression and it has played an important role in the oral transmission of world music. Hand-designed pitch descriptors are used to model aspects of the singing voice and clustering methods reveal singing style similarities in world music. Second, a study on music dissimilarity is performed. While musical exchange is evident in the history of world music it might be possible that some music cultures have resisted external musical influence. Low-level audio features are combined with machine learning methods to find music examples that stand out in a world music corpus, and geographical patterns are examined. The last study models music similarity using descriptors learned automatically with deep neural networks. It focuses on identifying music examples that appear to be similar in their audio content but share no (obvious) geographical or cultural links in their metadata. Unexpected similarities modelled in this way uncover possible hidden links between world music cultures. This research investigates whether automatic computational analysis can uncover meaningful similarities between recordings of world music. Applications derive musicological insights from one of the largest world music corpora studied so far. Computational analysis as proposed in this thesis advances the state-of-the-art in the study of world music and expands the knowledge and understanding of musical exchange in the world.

Le XXème siècle marque le début de l’étude musicologique sur le chant basque, concept apparu dans le courant du XIXème siècle. La plupart des recherches insistent sur les spécificités de cette pratique. A cette même époque, l’enregistrement de ce répertoire permet la constitution d’archives sonores. Pourtant, musicologues et collecteurs de sources sonores ne se rencontrent pas. Or, l’examen approfondi des études musicologiques publiées sur le chant basque amène à penser que les caractéristiques musicales attribuées à cette pratique tout au long du XXème siècle ne sont pas toutes pertinentes, que ce soit dans les termes employés ou dans la conception même de la musique. Elles renseignent parfois plus sur les auteurs et le contexte d’écriture de l’ouvrage/ article que sur la musique elle-même. Ces descriptionspeuvent de ce fait être nuancées à la lumière de nouvelles analyses. Une comparaison diachronique et synchronique de sources écrites et musicales de la fin du XIXème siècle à la fin du XXème siècle permet d’une part d’observer le lien entre ce qui est « dit » sur le sujet du point de vue musicologique et ce que nous pouvons en entendre, et d’autre part de repérer les caractéristiques musicales communes qui se dégagent de l’ensemble. Nous formulons l’hypothèse qu’il existe dans le chant basque monodique, non des caractéristiques musicales intrinsèques, mais des éléments musicaux mobiles, variés, des « indices », notamment du point de vue de la hauteur, qui peuvent être observés grâce à des outils d’analyse modernes. / The XXth century marks the beginning of the musicological study of the Basque singing, a concept appeared during the course of the XIXth century. Most of these researchs insist on the specificities of such a practice. At the same time, the recording of this repertory allows the setting up of sound archives. Nevertheless, musicologists and collectors of the sound sources do not meet. However, a thorough examination of the musicology studies published about the Basque singing leads to think that the musical characteristics ascribed to this practice throughout the XXth century are not all relevant whether concerning the terms being used or the conception of the very music. They often inform more on the authors and the context of the writing of the work/article than on the music itself. These descriptions may then be qualified in the light of new analyses. A diachronic and synchronic comparison of the written and musical sources from the late XIXth century to the end of the XXth allows us, on the one hand, to observe the link between what is “said” on the topic from the musicology point of view and what we can hear, and on the other hand to spot the common musical characteristics emerging from the whole.We formulate the hypothesis that the monodic Basque singing does not present any intrinsic musical characteristics. Rather, we observe mobile and diverse musical elements, kinds of “clues”, especially as far as pitch is concerned, that can be observed thanks to modern analysis tools.

Cette recherche est consacrée à la découverte et à l’appréhension des musiques de l’Indochine au tournant du XXe siècle, dans un contexte où cohabitaient l’orientalisme – représentation de l’Ailleurs – et la colonisation – mis en présence d’une réalité. La recherche s’articule autour du musicologue et compositeur Gaston Knosp. Ce dernier joua le rôle de défricheur dans l’étude de musiques indochinoises en réalisant les premiers travaux in situ durant son séjour au Tonkin (1898 - 1904). L’appréhension des musiques indochinoises sera étudiée à partir d’une analyse synchronique et diachronique des discours francophones qu’elle a générés. Les écrits de Gaston Knosp constituent le noyau de cette recherche; ils seront mis en perspective avec des documents d’archives, des textes publiés en Indochine, dans le cadre de la documentation sur les colonies, et avec des articles musicologiques parus en France. Cela afin d’offrir un panorama plus complet du mode de perception sensible et intellectuel de ces musiques, lors de leur découverte, et d’analyser les mécanismes qui opèrent dans les discours. La découverte des musiques indochinoises représentant un moment de transition entre l’imaginaire et la connaissance, les principaux objectifs de cette recherche seront : d’observer et d’analyser comment se manifeste l’influence du bagage culturel de Knosp sur son appréhension et sa perception des musiques indochinoises; de déterminer ce qu’apportent les travaux d’un musicologue par rapport à ceux d’informateurs coloniaux, non spécialistes – certains ayant développé une connaissance des cultures autochtones qui manquait à Knosp. Cette recherche sera également l’occasion de développer les connaissances sur Gaston Knosp, un musicien peu étudié, et de présenter plusieurs écrits parus en Indochine, souvent restés dans l’ombre, qui contribuent cependant à alimenter les connaissances sur l’appréhension des musiques indochinoises durant la colonisation. / This research is devoted to the discovery and understanding of the music of Indochina at the turn of the 20th century, in a context where orientalism - the representation of Elsewhere - and colonization - were cohabiting in the presence of a reality. The research revolves around the musicologist and composer Gaston Knosp, who played the role of pioneer in the study of Indochinese music by carrying out the first works in situ during his stay in Tonkin (1898 - 1904). The apprehension of Indochinese music will be studied from a synchronic and diachronic analysis of the French-speaking discourses it generated. Gaston Knosp's writings constitute the core of this research; they will be put into perspective with archival documents, texts published in Indochina as part of the documentation on the colonies, and with musicological articles published in France. The aim is to offer a more complete panorama of the sensitive and intellectual perception of this music at the time of its discovery, and to analyze the mechanisms that operate in the discourses. Since the discovery of Indochinese music represents a moment of transition between imagination and knowledge, the main objectives of this research will be: to observe and analyze how Knosp's cultural background manifests itself on his apprehension and perception of Indochinese music; to determine what the work of a musicologist brings to the table in comparison with that of colonial, non-specialist informants - some of whom developed a knowledge of indigenous cultures that Knosp lacked. This research will also be an opportunity to develop knowledge about Gaston Knosp, a musician who was little studied, and to present several writings published in Indochina, which often remained in the shadows, but which nevertheless contribute to knowledge about the apprehension of Indochinese music during colonization.