Music Similarity and Retrieval : An Introduction to Audio- and Web-Based Strategies.

Intro -- Foreword -- Preface -- Acknowledgments -- Contents -- 1 Introduction to Music Similarity and Retrieval -- 1.1 Music Information Retrieval -- 1.2 MIR from an Information Retrieval Perspective -- 1.2.1 Retrieval Tasks and Applications in MIR -- 1.2.2 Browsing Interfaces in MIR -- 1.2.3 Recommendation Tasks and Applications in MIR -- 1.2.4 MIR Beyond Retrieval, Browsing,and Recommendation -- 1.3 Music Similarity -- 1.3.1 Computational Factors of Music Similarity -- 1.3.2 Music Features -- 1.3.2.1 The Semantic Gap -- 1.3.2.2 A Critical View on the IR Perspective on Similarity -- 1.4 Contents of this Book -- 1.5 Evaluation of Music Similarity Algorithms -- 1.5.1 Evaluation Using Prelabeled Data -- 1.5.2 Evaluation Using Human Judgments -- 1.5.3 Evaluation Using Listening Histories -- 1.5.4 Music Collection and Evaluation in this Book -- 1.6 Further Reading -- Part I Content-Based MIR -- 2 Basic Methods of Audio Signal Processing -- 2.1 Categorization of Acoustic Music Features -- 2.2 Simplified Scheme of a Music Content Feature Extractor -- 2.2.1 Analog-Digital Conversion -- 2.2.2 Framing and Windowing -- 2.2.3 Fourier Transform -- 2.3 Common Low-Level Features -- 2.3.1 Time Domain Features -- 2.3.2 Frequency Domain Features -- 2.4 Summary -- 2.5 Further Reading -- 3 Audio Feature Extraction for Similarity Measurement -- 3.1 Psychoacoustic Processing -- 3.1.1 Physical Measurement of Sound Intensity -- 3.1.2 Perceptual Measurement of Loudness -- 3.1.3 Perception of Frequency -- 3.2 Frame-Level Features and Similarity -- 3.2.1 Mel Frequency Cepstral Coefficients -- 3.2.2 Statistical Summarization of Feature Vectors -- 3.2.3 Vector Quantization -- 3.2.4 Gaussian Mixture Models -- 3.2.5 Single Gaussian Model -- 3.3 Block-Level Features and Similarity -- 3.3.1 Fluctuation Pattern -- 3.3.2 Logarithmic Fluctuation Pattern -- 3.3.3 Spectral Pattern.

3.3.4 Correlation Pattern -- 3.3.5 Similarity in the Block-Level Feature Framework -- 3.4 Hubness and Distance Space Normalization -- 3.5 Summary -- 3.6 Further Reading -- 4 Semantic Labeling of Music -- 4.1 Genre Classification -- 4.2 Auto-tagging -- 4.2.1 Differences to Classification -- 4.2.2 Auto-Tagging Techniques -- 4.3 Mood Detection and Emotion Recognition -- 4.3.1 Models to Describe Human Emotion -- 4.3.1.1 Categorical Models -- 4.3.1.2 Dimensional Models -- 4.3.2 Emotion Recognition Techniques -- 4.3.2.1 Categorical Emotion Recognition Using Classification -- 4.3.2.2 Dimensional Emotion Recognition Using Regression Models -- 4.4 Summary -- 4.5 Further Reading -- Part II Music Context-Based MIR -- 5 Contextual Music Meta-data: Comparison and Sources -- 5.1 Web-Based Music Information Retrieval -- 5.1.1 The Web as Source for Music Features -- 5.1.2 Comparison with Content-Based Methods -- 5.1.3 Applications Using Web Data -- 5.2 Data Formats for Web-Based MIR -- 5.3 Tags and Annotations -- 5.3.1 Expert Annotations -- 5.3.2 Collaborative Tagging -- 5.3.3 Games with a Purpose -- 5.4 Web Texts -- 5.4.1 Web Pages Related to Music -- 5.4.1.1 Music-Focused Web Crawler -- 5.4.1.2 Page Retrieval Using a Web Search Engine -- 5.4.1.3 Web Content Filtering -- 5.4.1.4 Content Aggregation -- 5.4.2 Biographies, Product Reviews, and Audio Blogs -- 5.4.3 Microblogs -- 5.5 Lyrics -- 5.5.1 Analysis of Lyrics on the Web -- 5.5.2 Retrieval and Correction -- 5.6 Summary -- 5.7 Further Reading -- 6 Contextual Music Similarity, Indexing, and Retrieval -- 6.1 Text-Based Features and Similarity Measures -- 6.1.1 Vector Space Model -- 6.1.1.1 Similarity Calculation -- 6.1.2 Latent Semantic Indexing -- 6.1.3 Applications of Latent Factor Approaches -- 6.2 Text-Based Indexing and Retrieval -- 6.2.1 Pseudo Document Indexing -- 6.2.2 Document-Centered Rank-Based Scoring.

6.2.3 Auto-Tag Indexing -- 6.3 Similarity Based on Co-occurrences -- 6.4 Combination with Audio Content Information -- 6.4.1 Combined Similarity Measures -- 6.4.2 Contextual Filtering -- 6.4.3 Combined Tag Prediction -- 6.5 Stylistic Analysis and Similarity -- 6.6 Summary -- 6.7 Further Reading -- Part III User-Centric MIR -- 7 Listener-Centered Data Sources and Aspects: Traces of Music Interaction -- 7.1 Definition and Comparison of Listener-Centered Features -- 7.2 Personal Collections and Peer-to-Peer Network Folders -- 7.3 Listening Histories and Playlists -- 7.4 User Ratings -- 7.5 Modeling User Context -- 7.5.1 Sensor Data for Modeling User Context -- 7.5.2 Social Networks and User Connections -- 7.6 Factors of User Intentions -- 7.7 Summary -- 7.8 Further Reading -- 8 Collaborative Music Similarity and Recommendation -- 8.1 Similarity Estimation via Co-occurrence -- 8.2 Graph-Based and Distance-Based Similarity -- 8.3 Exploiting Latent Context from Listening Sessions -- 8.3.1 Latent Dirichlet Allocation -- 8.3.2 Case Study: Artist Clustering from Listening Events -- 8.3.3 Music Recommendation -- 8.4 Learning from Explicit and Implicit User Feedback -- 8.4.1 Memory-Based Collaborative Filtering -- 8.4.2 Model-Based Collaborative Filtering -- 8.5 Multimodal Combination -- 8.5.1 Hybrid Recommender Systems -- 8.5.1.1 Multimodal Extensions to Matrix Factorization -- 8.5.1.2 Classification and Regression -- 8.5.1.3 Probabilistic Combination -- 8.5.1.4 Graph-Based Combinations -- 8.5.2 Unified Metric Learning -- 8.5.2.1 Learning Similarity Metrics -- 8.5.2.2 Learning to Rank -- 8.6 Summary -- 8.7 Further Reading -- Part IV Current and Future Applications of MIR -- 9 Applications -- 9.1 Music Information Systems -- 9.1.1 Band Members and Their Roles -- 9.1.2 Artist's or Band's Country of Origin -- 9.1.3 Album Cover Artwork -- 9.1.4 Data Representation.

9.2 User Interfaces to Music Collections -- 9.2.1 Map-Based Interfaces -- 9.2.1.1 Self-organizing Map -- 9.2.1.2 SOM-Based Interfaces for Music Collections -- 9.2.1.3 Multidimensional Scaling -- 9.2.1.4 MDS-Based Interfaces for Music Collections -- 9.2.1.5 Map-Based Interfaces for Music Collections Based on Other Techniques -- 9.2.2 Other Intelligent Interfaces -- 9.3 Automatic Playlist Generation -- 9.4 Music Popularity Estimation -- 9.4.1 Popularity Estimation from Contextual Data Sources -- 9.4.2 Comparison of Data Sources -- 9.5 Summary -- 10 Grand Challenges and Outlook -- 10.1 Major Challenges -- 10.1.1 Methodological Challenges -- 10.1.2 Data-Related Challenges -- 10.1.3 User-Centric Challenges -- 10.1.4 General Challenges -- 10.2 Future Directions -- A Description of the Toy Music Data Set -- A.1 Electronic Music -- The Chemical Brothers -- Modeselektor -- Radiohead -- Trentemøller -- A.2 Classical Music -- Ludwig van Beethoven -- Johann Sebastian Bach -- Frédéric Chopin -- Dmitri Shostakovich -- A.3 Heavy Metal -- Black Sabbath -- Iron Maiden -- Kiss -- Metallica -- A.4 Rap -- Busta Rhymes -- Jay-Z -- Snoop Dogg -- Snow -- A.5 Pop -- Kylie Minogue -- Madonna -- Maroon 5 -- Lady Gaga -- References -- Index.

Description based on publisher supplied metadata and other sources.

Electronic reproduction. Ann Arbor, Michigan : ProQuest Ebook Central, 2024. Available via World Wide Web. Access may be limited to ProQuest Ebook Central affiliated libraries.