This book provides a summary of the manifold audio- and web-based approaches to music information retrieval (MIR) research. In contrast to other books dealing solely with music signal processing, it addresses additional cultural and listener-centric aspects and thus provides a more holistic view. Consequently, the text includes methods operating on features extracted directly from the audio signal, as well as methods operating on features extracted from contextual information, either the cultural context of music as represented on the web or the user and usage context of music. Following the prevalent document-centered paradigm of information retrieval, the book addresses models of music similarity that extract computational features to describe an entity that represents music on any level (e.g., song, album, or artist), and methods to calculate the similarity between them. While this perspective and the representations discussed cannot describe all musical dimensions, they enable us to effectively find music of similar qualities by providing abstract summarizations of musical artifacts from different modalities. The text at hand provides a comprehensive and accessible introduction to the topics of music search, retrieval, and recommendation from an academic perspective. It will not only allow those new to the field to quickly access MIR from an information retrieval point of view but also raise awareness for the developments of the music domain within the greater IR community. In this regard, Part I deals with content-based MIR, in particular the extraction of features from the music signal and similarity calculation for content-based retrieval. Part II subsequently addresses MIR methods that make use of the digitally accessible cultural context of music. Part III addresses methods of collaborative filtering and user-aware and multi-modal retrieval, while Part IV explores current and future applications of music retrieval and recommendation.>

Content-based multimedia retrieval is a challenging research field with many unsolved problems. This monograph details concepts and algorithms for robust and efficient information retrieval of two different types of multimedia data: waveform-based music data and human motion data. It first examines several approaches in music information retrieval, in particular general strategies as well as efficient algorithms. The book then introduces a general and unified framework for motion analysis, retrieval, and classification, highlighting the design of suitable features, the notion of similarity used to compare data streams, and data organization.

Sound waves propagate through various media, and allow communication or entertainment for us, humans. Music we hear or create can be perceived in such aspects as rhythm, melody, harmony, timbre, or mood. All these elements of music can be of interest for users of music information retrieval systems. Since vast music repositories are available for everyone in everyday use (both in private collections, and in the Internet), it is desirable and becomes necessary to browse music collections by contents. Therefore, music information retrieval can be potentially of interest for every user of computers and the Internet. There is a lot of research performed in music information retrieval domain, and the outcomes, as well as trends in this research, are certainly worth popularizing. This idea motivated us to prepare the book on Advances in Music Information Retrieval. It is divided into four sections: MIR Methods and Platforms, Harmony, Music Similarity, and Content Based Identification and Retrieval. Glossary of basic terms is given at the end of the book, to familiarize readers with vocabulary referring to music information retrieval.

Music Information Retrieval: Recent Developments and Applications surveys the young but established field of research that is Music Information Retrieval (MIR). In doing so, it pays particular attention to the latest developments in MIR, such as semantic auto-tagging and user-centric retrieval and recommendation approaches. Music Information Retrieval: Recent Developments and Applications starts by reviewing the well-established and proven methods for feature extraction and music indexing, from both the audio signal and contextual data sources about music items, such as web pages or collaborative tags. These in turn enable a wide variety of music retrieval tasks, such as semantic music search or music identification ("query by example"). Subsequently, it elaborates on the current work on user analysis and modeling in the context of music recommendation and retrieval, addressing the recent trend towards user-centric and adaptive approaches and systems. A discussion follows about the important aspect of how various MIR approaches to different problems are evaluated and compared. It concludes with a discussion about the major open challenges facing MIR.

This thesis shows how to develop an automatic, large-scale music recommendation system. To achieve this goal we solve three problems preventing the currently top-performing class of content-based music similarity algorithms from being used as recommendation engine in huge databases with millions of songs: First, we show how to correctly use the non-vectorial music similarity features with their non-metric divergences in centroid-computing algorithms. All previous approaches had to artificially vectorize the data before they were able to work with the features. Second, we show how the problem of 'hubs' can be alleviated. Hubs are objects in a recommendation system which are unwontedly often retrieved as nearest neighbors. The examined music recommendation methods are especially prone to hubs, significantly decreasing their retrieval quality. We also identify hubs as a problem of machine learning and show the beneficial effects of our method on a large number of general public machine learning collections. Third, we present a new method to speed up music recommendation queries. The method uses a filter-and-refine systems layout. It achieves a very high retrieval accuracy and speeds up queries by a factor of 10--40 compared to a linear scan. The method enables us to use the music similarity methods with very large databases. We finally merge all three introduced methods in a large-scale, high-quality music recommendation prototype: the system computes (i) a natural clustering of the music similarity features to (ii) apply the introduced hub-reducing method and (iii) use the filter-and-refine method to allow for fast retrieval. The prototype is called 'Wolperdinger', it operates on a collection of 2.3 million songs and it is able to answer recommendation queries in a fraction of a second. It is the largest content-based music recommendation system published to date.