No description available.

This thesis introduces a music visualization system for stereophonic source separated music. Music visualization systems are a popular way to represent information from audio signals through computer graphics. Visualization can help people better understand music and its complex and interacting elements. This music visualization system extracts pitch, panning, and loudness features from source separated audio files to create the visual. Most state-of-the art visualization systems develop their visual representation of the music from either the fully mixed final song recording, where all of the instruments and vocals are combined into one file, or from the digital audio workstation (DAW) data containing multiple independent recordings of individual audio sources. Original source recordings are not always readily available to the public so music source separation (MSS) can be used to obtain estimated versions of the audio source files. This thesis surveys different approaches to MSS and music visualization as well as introduces a new music visualization system specifically for source separated music.

Beat tracking is a common research area within music information retrieval (MIR). Jazz is a musical genre that is commonly rich in rhythmical complexity, which makes beat tracking challenging. The aim of this study is to analyze how well beat tracking algorithms detect the beats of jazz instruments. In this study, drums and double-bass were recorded in an ensemble but on single tracks. Three modern beat tracking algorithms were examined: LibRosa Dynamic, Essentia Multi-Feature, and madmom RNN. The algorithms’ beat trackings were evaluated using common metrics: the F-measure, P-score and Cemgil accuracy. The results showed that bass tracks generally got consistent results from all algorithms. However, all algorithms struggled with octave errors (the detected number of beats is off by a factor of two) and off-beats. When music was played without restrictions to the beat rhythm, madmom RNN generally performed the best, which suggests that machine learning with RNN (recurrent neural networks) is a good approach for beat tracking on rhythmically complex tracks. / Beat tracking är ett vanligt område inom music information retrieval (MIR). Jazz är en musikgenre som ofta är rytmisk komplex, vilket kan göra beat tracking utmanande. Denna studies syfte är att analysera hur väl beat tracking-algoritmer detekterar taktslagen hos jazzinstrument. I denna studie spelades trummor och kontrabas in samtidigt, men på separata spår. Tre moderna algoritmer för beat tracking undersöktes: LibRosa Dynamic, Essentia Multi-Feature, samt madmom RNN. Algoritmernas taktslagsestimeringar utvärderades utifrån vanligen tillämpade mätetal: F-measure, P-score och Cemgil accuracy. Resultaten visade att basspåren generellt sett gav konsekventa resultat från alla algoritmer. Däremot fick alla algoritmer oktavfel (antalet detekterade beats är fel med en faktor två) och off-beats. När musiken framfördes utan restriktioner i spelade taktslag presterade madmom RNN generellt bäst. Detta pekar mot att maskininlärning med RNN (recurrent neural networks) är en bra metod för beat tracking av rytmiskt komplexa spår.

Smart musical instruments are an emerging class of digital musical instruments designed for music creation in an interconnected Internet of Musical Things scenario. These instruments aim to integrate embedded computation, real-time feature extraction, gesture acquisition, and networked communication technologies. As embedded computers become more capable and new embedded audio platforms are developed, new avenues for real-time embedded gesture acquisition open up. Expressive guitar technique recognition is the task of detecting notes and classifying the playing techniques used by the musician on the instrument. Real-time recognition of expressive guitar techniques in a smart guitar would allow players to control sound synthesis or to wirelessly interact with a wide range of interconnected devices and stage equipment during performance. Despite expressive guitar technique recognition being a well-researched topic in the field of Music Information Retrieval, the creation of a lightweight real-time recognition system that can be deployed on an embedded platform still remains an open problem. In this thesis, expressive guitar technique recognition is investigated by focusing on real-time execution, and the execution of deep learning inference on resource-constrained embedded computers. Initial efforts have focused on clearly defining the challenges of embedded real-time music information retrieval, and on the creation of a first, fully embedded, real-time expressive guitar technique recognition system. The insight gained, led to the refinement of the various steps of the proposed recognition pipeline. As a first refinement step, a novel procedure for the optimization of onset detectors was developed. The proposed procedure adopts an evolutionary algorithm to find parameter configurations that are optimal both in terms of detection accuracy and latency. A subsequent study is devoted to shedding light on the performance of generic deep learning inference engines for embedded real-time audio classification. This consisted of a comparison of four common inferencing libraries, which focus on the applicability of each library to real-time audio inference, and their performance in terms of execution time and several additional metrics. Different insights from these studies supported the development of a new expressive guitar technique classifier, which is accompanied by an in-depth analysis of different aspects of the recognition problem. Finally, the experience collected during these studies culminated in the definition of a procedure to deploy deep learning inference to a prominent embedded platform. These investigations have been shown to improve the state-of-the-art by proposing approaches that surpass previous alternatives and providing new knowledge on problems and tools that can aid the creation of a smart guitar. The new knowledge provided was also adopted for embedded audio tasks that differ from real-time expressive guitar technique recognition.

No description available.

No description available.

Cette thèse rend compte de travaux portant sur l’inférence de structures répétitives à partir du signal audio à l’aide d’algorithmes du texte. Son objectif principal est de proposer et d’évaluer des algorithmes d’inférence à partir d’une étude formelle des notions de similarité et de répétition musicale.Nous présentons d’abord une méthode permettant d’obtenir une représentation séquentielle à partir du signal audio. Nous introduisons des outils d’alignement permettant d’estimer la similarité entre de telles séquences musicales, et évaluons l’application de ces outils pour l’identification automatique de reprises. Nous adaptons alors une technique d’indexation de séquences biologiques permettant une estimation efficace de la similarité musicale au sein de bases de données conséquentes.Nous introduisons ensuite plusieurs répétitions musicales caractéristiques et employons les outils d’alignement pour identifier ces répétitions. Une première structure, la répétition d’un segment choisi, est analysée et évaluée dans le cadre dela reconstruction de données manquantes. Une deuxième structure, la répétition majeure, est définie, analysée et évaluée par rapport à un ensemble d’annotations d’experts, puis en tant qu’alternative d’indexation pour l’identification de reprises.Nous présentons enfin la problématique d’inférence de structures répétitives telle qu’elle est traitée dans la littérature, et proposons notre propre formalisation du problème. Nous exposons alors notre modélisation et proposons un algorithme permettant d’identifier une hiérarchie de répétitions. Nous montrons la pertinence de notre méthode à travers plusieurs exemples et en l’évaluant par rapport à l’état de l’art. / The work presented in this thesis deals with repetitive structure inference from audio signal using string matching techniques. It aims at proposing and evaluating inference algorithms from a formal study of notions of similarity and repetition in music.We first present a method for representing audio signals by symbolic strings. We introduce alignment tools enabling similarity estimation between such musical strings, and evaluate the application of these tools for automatic cover song identification. We further adapt a bioinformatics indexing technique to allow efficient assessments of music similarity in large-scale datasets. We then introduce several specific repetitive structures and use alignment tools to analyse these repetitions. A first structure, namely the repetition of a chosen segment, is retrieved and evaluated in the context of automatic assignment of missingaudio data. A second structure, namely the major repetition, is defined, retrieved and evaluated regarding expert annotations, and as an alternative indexing method for cover song identification.We finally present the problem of repetitive structure inference as addressed in literature, and propose our own problem statement. We further describe our model and propose an algorithm enabling the identification of a hierarchical music structure. We emphasize the relevance of our method through several examples and by comparing it to the state of the art.

Cover song identification is a field of music information retrieval where the task is to determine whether two different audio tracks represent different versions of the same underlying song. Since covers might differ in tempo, key, instrumentation and other characteristics, many clever features have been developed over the years. We perform a rigorous analysis of 32 features used in related works while distinguishing between exact and scalable features. The former are based on a harmonic descriptor time series (typically chroma vectors) and offer better performance at the cost of computation time. The latter have a small constant size and only capture global phenomena in the track, making them fast to compute and suitable for use with large datasets. We then select 7 scalable and 3 exact features to build our own two-level system, with the scalable features used on the first level to prune the dataset and the exact on the second level to refine the results. Two distinct machine learning models are used to combine the scalable resp. exact features. We perform the analysis and the evaluation of our system on the Million Song Dataset. The experiments show the exact features being outperformed by the scalable ones, which lead us to a decision to only use the 7 scalable features in our system. The...

Les systèmes actuels d’improvisation musicales sont capables de générer des séquences musicales unidimensionnelles par recombinaison du matériel musical. Cependant, la prise en compte de plusieurs dimensions (mélodie, harmonie...) et la modélisation de plusieurs niveaux temporels sont des problèmes difficiles. Dans cette thèse, nous proposons de combiner des approches probabilistes et des méthodes issues de la théorie des langages formels afin de mieux apprécier la complexité du discours musical à la fois d’un point de vue multidimensionnel et multi-niveaux dans le cadre de l’improvisation où la quantité de données est limitée. Dans un premier temps, nous présentons un système capable de suivre la logique contextuelle d’une improvisation représentée par un oracle des facteurs tout en enrichissant son discours musical à l’aide de connaissances multidimensionnelles représentées par des modèles probabilistes interpolés. Ensuite, ces travaux sont étendus pour modéliser l’interaction entre plusieurs musiciens ou entre plusieurs dimensions par un algorithme de propagation de croyance afin de générer des improvisations multidimensionnelles. Enfin, nous proposons un système capable d’improviser sur un scénario temporel avec des informations multi-niveaux représenté par une grammaire hiérarchique. Nous proposons également une méthode d’apprentissage pour l’analyse automatique de structures temporelles hiérarchiques. Tous les systèmes sont évalués par des musiciens et improvisateurs experts lors de sessions d’écoute / Current musical improvisation systems are able to generate unidimensional musical sequences by recombining their musical contents. However, considering several dimensions (melody, harmony...) and several temporal levels are difficult issues. In this thesis, we propose to combine probabilistic approaches with formal language theory in order to better assess the complexity of a musical discourse, both from a multidimensional and multi-level point of view in the context of improvisation where the amount of data is limited. First, we present a system able to follow the contextual logic of an improvisation modelled by a factor oracle whilst enriching its musical discourse with multidimensional knowledge represented by interpolated probabilistic models. Then, this work is extended to create another system using a belief propagation algorithm representing the interaction between several musicians, or between several dimensions, in order to generate multidimensional improvisations. Finally, we propose a system able to improvise on a temporal scenario with multi-level information modelled with a hierarchical grammar. We also propose a learning method for the automatic analysis of hierarchical temporal structures. Every system is evaluated by professional musicians and improvisers during listening sessions

Harmony is the aspect of music concerned with the structure, progression, and relation of chords. In Western tonal music each period had different rules and practices of harmony. Similarly some composers and musicians are recognised for their characteristic harmonic patterns which differ from the chord sequences used by other musicians of the same period or genre. This thesis is concerned with the automatic induction of the harmony rules and patterns underlying a genre, a composer, or more generally a 'style'. Many of the existing approaches for music classification or pattern extraction make use of statistical methods which present several limitations. Typically they are black boxes, can not be fed with background knowledge, do not take into account the intricate temporal dimension of the musical data, and ignore rare but informative events. To overcome these limitations we adopt first-order logic representations of chord sequences and Inductive Logic Programming techniques to infer models of style. We introduce a fixed length representation of chord sequences similar to n-grams but based on first-order logic, and use it to characterise symbolic corpora of pop and jazz music. We extend our knowledge representation scheme using context-free definite-clause grammars, which support chord sequences of any length and allow to skip ornamental chords, and test it on genre classification problems, on both symbolic and audio data. Through these experiments we also compare various chord and harmony characteristics such as degree, root note, intervals between root notes, chord labels and assess their characterisation and classification accuracy, expressiveness, and computational cost. Moreover we extend a state- of-the-art genre classifier based on low-level audio features with such harmony-based models and prove that it can lead to statistically significant classification improvements. We show our logic-based modelling approach can not only compete with and improve on statistical approaches but also provides expressive, transparent and musicologically meaningful models of harmony which makes it suitable for knowledge discovery purposes.