Carlos Pérez-Sancho

Genre classification using chords and stochastic language models

Music genre meta-data is of paramount importance for the organisation of music repositories. People use genre in a natural way when entering a music store or looking into music collections. Automatic genre classification has become a popular topic in music information retrieval research both, with digital audio and symbolic data. This work focuses on the symbolic approach, bringing to music cognition some technologies, like the stochastic language models, already successfully applied to text categorisation. The representation chosen here is to model chord progressions as n-grams and strings and then apply perplexity and naïve Bayes classifiers, respectively, in order to assess how often those structures are found in the target genres. Some genres and sub-genres among popular, jazz, and academic music have been considered, trying to investigate how far can we reach using harmonic information with these models. The results at different levels of the genre hierarchy for the techniques employed are presented and discussed.

Style recognition through statistical event models

Abstract The automatic classification of music fragments into style classes is one challenging problem within the music information retrieval (MIR) domain and also for the understanding of music style perception. This has a number of applications, including the indexation and exploration of musical databases. Some technologies employed in text classification can be applied to this problem. The key point here is to establish the music equivalent to the words in texts. A number of works use the combination of intervals and duration ratios for this purpose. In this paper, different statistical text recognition algorithms are applied to style recognition using this kind of melody representation, exploring their performance for different word sizes.

Feature selection in a cartesian ensemble of feature subspace classifiers for music categorisation

In this paper, we evaluate the impact of feature selection on the classification accuracy and the achieved dimensionality reduction, which benefits the time needed on training classification models. Our classification scheme therein is a Cartesian ensemble classification system, based on the principle of late fusion and feature subspaces. These feature subspaces describe different aspects of the same data set. We use it for the ensemble classification of multiple feature sets from the audio and symbolic domains. We present an extensive set of experiments in the context of music genre classification, based on Music IR benchmark datasets. We show that while feature selection does not benefit classification accuracy, it greatly reduces the dimensionality of each feature subspace, and thus adds to great gains in the time needed to train the individual classification models that form the ensemble.

Stochastic Text Models for Music Categorization

Music genre meta-data is of paramount importance for the organization of music repositories. People use genre in a natural way when entering a music store or looking into music collections. Automatic genre classification has become a popular topic in music information retrieval research. This work brings to symbolic music recognition some technologies, like the stochastic language models, already successfully applied to text categorization. In this work we model chord progressions and melodies as n -grams and strings and then apply perplexity and naive Bayes classifiers, respectively, in order to assess how often those structures are found in the target genres. Also a combination of the different techniques as an ensemble of classifiers is proposed. Some genres and sub-genres among popular, jazz, and academic music have been considered. The results show that the ensemble is a good trade-off approach able to perform well without the risk of choosing the wrong classifier.

A Shallow Description Framework for Musical Style Recognition

In the field of computer music, pattern recognition algorithms are very relevant for music information retrieval (MIR). One challenging task within this area is the automatic recognition of musical style, that has a number of applications like indexing and selecting musical databases. In this paper, the classification of monophonic melodies of two different musical styles (jazz and classical) represented symbolically as MIDI files is studied, using different classification methods: Bayesian classifier and nearest neighbour classifier. From the music sequences, a number of melodic, harmonic, and rhythmic statistical descriptors are computed and used for style recognition. We present a performance analysis of such algorithms against different description models and parameters.

Modeling Musical Style with Language Models for Composer Recognition

In this paper we present an application of language modeling using n-grams to model the style of different composers. For this, we repeated the experiments performed in previous works by other authors using a corpus of 5 composers from the Baroque and Classical periods. In these experiments we found some signs that the results could be influenced by external factors other than the composers’ styles, such as the heterogeneity in the musical forms selected for the corpus. In order to assess the validity of the modeling techniques to capture the own personal style of the composers, a new experiment was performed with a corpus of fugues from Bach and Shostakovich. All these experiments show that language modeling is a suitable tool for modeling musical style, even when the styles of the different datasets are affected by several factors.

Harmonic and instrumental information fusion for musical genre classification

This paper presents a musical genre classification system based on the combination of two kinds of information of very different nature: the instrumentation information contained in a MIDI file (metadata) and the chords that provide the harmonic structure of the musical score stored in that file (content). The fusion of these two information sources gives a single feature vector that represents the file and to which classification techniques usually utilized for text categorization tasks are applied. The classification task is performed under a probabilistic approach that has improved the results previously obtained for the same data using the instrumental or the chord information independently.

A text categorization approach for music style recognition

The automatic classification of music files into styles is one challenging problem in music information retrieval and for music style perception understanding. It has a number of applications, like the indexation and exploration of musical databases. Some techniques used in text classification can be applied to this problem. The key point is to establish a music equivalent to the words in texts. A number of works use the combination of intervals and duration ratios for music description. In this paper, different statistical text recognition algorithms are applied to style recognition using this kind of melody representation, exploring their performance for different word sizes and statistical models.

Interactive multimodal music transcription

Automatic music transcription has usually been performed as an autonomous task and its evaluation has been made in terms of precision, recall, accuracy, etc. Nevertheless, in this work, assuming that the state of the art is far from being perfect, it is considered as an interactive one, where an expert user is assisted in its work by a transcription tool. In this context, the performance evaluation of the system turns into an assessment of how many user interactions are needed to complete the work. The strategy is that the user interactions can be used by the system to improve its performance in an adaptive way, thus minimizing the workload. Also, a multimodal approach has been implemented, in such a way that different sources of information, like onsets, beats, and meter, are used to detect notes in a musical audio excerpt. The system is focused on monotimbral polyphonic transcription.

Genre classification of music by tonal harmony

In this paper we present a genre classification framework for audio music based on a symbolic classification system. Audio signals are transformed into a symbolic representation of harmony using a chord transcription algorithm, based on the computation of harmonic pitch class profiles. Then, language models built from a ground truth of chord progressions for each genre are used to perform classification. We show that chord progressions are a suitable feature to represent musical genre, as they capture the harmonic rules relevant in each musical period or style. Finally, results using both pure symbolic information and chords transcribed from audio-from-MIDI are compared, in order to evaluate the effects of the transcription process in this task.