Abstract. We present an algorithm that predicts musical genre and artist from an audio waveform. Our method uses the ensemble learner AdaBoost to select from a set of audio features that have been extracted from segmented audio and then aggregated. Our classifier proved to be the most effective method for genre classification at the recent MIREX 2005 international contests in music information extraction, and the second-best method for recognizing artists. This paper describes our method in detail, from feature extraction to song classification, and presents an evaluation of our method on three genre databases and two artist-recognition databases. Furthermore, we present evidence collected from a variety of popular features and classifiers that the technique of classifying features aggregated over segments of audio is better than classifying either entire songs or individual shorttimescale features.

A software system that automatically classifies MIDI files into hierarchically organized taxonomies of musical genres is presented. This extensible software includes an easy to use and flexible GUI. An extensive library of high-level musical features is compiled, including many original features. A novel hybrid classification system is used that makes use of hierarchical, flat and round robin classification. Both k-nearest neighbour and neural network-based classifiers are used, and feature selection and weighting are performed using genetic algorithms. A thorough review of previous research in automatic genre classification is presented, along with an overview of automatic feature selection and classification techniques. Also included is a discussion of the theoretical issues relating to musical genre, including but not limited to what mechanisms humans use to classify music by genre and how realistic genre taxonomies can be constructed.

Abstract—We present a new technique for audio signal comparison based on tonal subsequence alignment and its application to detect cover versions (i.e., different performances of the same underlying musical piece). Cover song identification is a task whose popularity has increased in the Music Information Retrieval (MIR) community along in the past, as it provides a direct and objective way to evaluate music similarity algorithms. This article first presents a series of experiments carried out with two state-of-the-art methods for cover song identification. We have studied several components of these (such as chroma resolution and similarity, transposition, beat tracking or Dynamic Time Warping constraints), in order to discover which characteristics would be desirable for a competitive cover song identifier. After analyzing many cross-validated results, the importance of these characteristics is discussed, and the best-performing ones are finally applied to the newly proposed method. Multiple evaluations of this one confirm a large increase in identification accuracy when comparing it with alternative state-of-the-art approaches.

For tasks like musical genre identification and similarity searches in audio databases, audio files have to be described by suitable feature sets. Since these feature sets usually try to capture diverse discriminative characteristics, it is interesting and desirable to create an acoustic representation of the feature set to support intuitive evaluation. In this paper, we present an approach for making a specific feature set, namely Rhythm Patterns, instantly human comprehensible by re-assembling sound from the numerical descriptors. The re-synthesized audio chunks represent clearly perceivable rhythmical characteristics on critical frequency bands of the original music.

Automatic annotation of music files is a key problem in multimedia information retrieval. In this paper we present a solution to this problem that addresses the issues of feature extraction, feature selection and design of classifier.

and beyond

Many if not most audio features used in MIR research are inspired by work done in speech recognition and are variations on the spectrogram. Recently, much attention has been given to new representations of audio that are sparse and time-relative. These representations are efficient and able to avoid the time-frequency trade-off of a spectrogram. Yet little work with music streams has been conducted and these features remain mostly unused in the MIR community. In this paper we further explore the use of these features for musical signals. In particular, we investigate their use on realistic music examples (i.e. released commercial music) and their use as input features for supervised learning. Furthermore, we identify three specific issues related to these features which will need to be further addressed in order to obtain the full benefit for MIR applications. 1

Abstract—This paper examines a query-by-example approach to spoken term detection in audio files. The approach is designed for low-resource situations in which limited or no in-domain training material is available and accurate word-based speech recognition capability is unavailable. Instead of using word or phone strings as search terms, the user presents the system with audio snippets of desired search terms to act as the queries. Query and test materials are represented using phonetic posteriorgrams obtained from a phonetic recognition system. Query matches in the test data are located using a modified dynamic time warping search between query templates and test utterances. Experiments using this approach are presented using data from the Fisher corpus. I.

Abstract. One of the biggest challenges facing current methods for classifying music by genre or artist is that features of the sound are computed on very small temporal scales (20 to 50 milliseconds), while the labels need to be assigned at relatively large temporal scales (3 to 5 minutes). We address this challenge by partitioning songs into smaller pieces and classifying each one separately. Our choice of features together with an AdaBoost.MH classifier proved to be the most effective method for genre classification at the recent MIREX 2005 international contests in music information extraction, and the second-best method for recognizing artists. This paper describes our method in detail, from feature extraction to song classification, and presents an evaluation of our method on three genre databases and two artist-recognition databases. Furthermore, we present evidence that the method of partitioning songs is better than classifying either entire songs or individual features, using a variety of popular features and classifiers.

This paper presents a genre classification algorithm for symbolic music data. The proposed methodology relies on note pitch and duration features, derived from the repeating patterns and duration histograms of a musical piece, respectively. Note-information histograms have a great capability in capturing a fair amount of information regarding harmonic as well as rhythmic features of different musical genres and pieces, while repeating patterns refer to segments of the piece that are semantically important. Detailed experimental results on intra-classical genres illustrate the significant performance gains due to the proposed features.