Melody Identification in Standard MIDI Files

2009

A Bayesian dynamic model is developed to model complex sequential data, with a focus on audio signals from music. The music is represented in terms of a sequence of discrete observations, and the sequence is modeled using a hidden Markov model (HMM) with time-evolving parameters. The model imposes the belief that observations that are temporally proximate are more likely to be drawn from HMMs with similar parameters, while also allowing for ldquoinnovationrdquo associated with abrupt changes in the music texture. Segmentation of a given musical piece is constituted via the model inference and the results are compared with other models and also to a conventional music-theoretic analysis.

2019

In this paper, we use Hidden Markov Models (HMM) and Mel-Frequency Cepstral Coefficients (MFCC) to build statistical models of classical music composers directly from the music datasets. Several musical pieces are divided by instruments (String, Piano, Chorus, Orchestra), and, for each instrument, statistical models of the composers are computed. We selected 19 different composers spanning four centuries by using a total number of 400 musical pieces. Each musical piece is classified as belonging to a composer if the corresponding HMM gives the highest likelihood for that piece. We show that the so-developed models can be used to obtain useful information on the correlation between the composers. Moreover, by using the maximum likelihood approach, we also classified the instrumentation used by the same composer. Besides as an analysis tool, the described approach has been used as a classifier. This overall originates an HMM-based framework for supporting accurate classification of mu...

Proc. ISMIR, 2005

We introduce a new model for extracting classified structural segments, such as intro, verse, chorus, break and so forth, from recorded music. Our approach is to classify signal frames on the basis of their audio properties and then to agglomerate contiguous runs of similarly classified frames into texturally homogenous (or 'self-similar') segments which inherit the classificaton of their consituent frames. Our work extends previous work on automatic structure extraction by addressing the classification problem using using an unsupervised Bayesian clustering model, the parameters of which are estimated using a variant of the expectation maximisation (EM) algorithm which includes deterministic annealing to help avoid local optima. The model identifies and classifies all the segments in a song, not just the chorus or longest segment. We discuss the theory, implementation, and evaluation of the model, and test its performance against a ground truth of human judgements. Using an analogue of a precisionrecall graph for segment boundaries, our results indicate an optimal trade-off point at approximately 80% precision for 80% recall.

The increasing availability of on-line music has motivated a growing interest for organizing, commercializing, and delivering this kind of multimedia content. For it, the use of metadata is of utmost importance. Metadata permit organization, indexing, and retrieval of music contents. They are, therefore, a subject of research both from the design and automatic extraction approaches. The present work focuses on this second issue, providing an open source tool for metadata extraction from standard MIDI files. The tool is presented, the utilized metadata are explained, and some applications and experiments are described as examples of its capabilities.

2013

This paper demonstrates that melodic structure fundamentally differentiates musical genres. We use two methods: the k-Means algorithm as an unsupervised learning method for understanding how to cluster unorganized music and a Markov Chain Model which determines relative probabilities for the next note’s most likely value, and evaluate our algorithms’ accuracy in predicting correct genre. Our experiments indicate that the k-Means approach is modestly successful for separating out most genres, whereas the Markov Chain Model tends to be very accurate for music classification.

EURASIP Journal on Applied …, 2007

We report our findings on using MIDI files and audio features from MIDI, separately and combined together, for MIDI music genre classification. We use McKay and Fujinaga's 3-root and 9-leaf genre data set. In order to compute distances between MIDI pieces, we use normalized compression distance (NCD). NCD uses the compressed length of a string as an approximation to its Kolmogorov complexity and has previously been used for music genre and composer clustering. We convert the MIDI pieces to audio and then use the audio features to train different classifiers. MIDI and audio from MIDI classifiers alone achieve much smaller accuracies than those reported by McKay and Fujinaga who used not NCD but a number of domain-based MIDI features for their classification. Combining MIDI and audio from MIDI classifiers improves accuracy and gets closer to, but still worse, accuracies than McKay and Fujinaga's. The best root genre accuracies achieved using MIDI, audio, and combination of them are 0.75, 0.86, and 0.93, respectively, compared to 0.98 of McKay and Fujinaga. Successful classifier combination requires diversity of the base classifiers. We achieve diversity through using certain number of seconds of the MIDI file, different sample rates and sizes for the audio file, and different classification algorithms.

2002

This paper describes a Hidden Markov Model (HMM)-based method of automatic transcription of MIDI (Musical Instrument Digital Interface) signals of performed music. The problem is formulated as recognition of a given sequence of fluctuating note durations to find the most likely intended note sequence utilizing the modern continuous speech recognition technique. Combining a stochastic model of deviating note durations and a stochastic grammar representing possible sequences of notes, the maximum likelihood estimate of the note sequence is searched in terms of Viterbi algorithm. The same principle is successfully applied to a joint problem of bar line allocation, time measure recognition, and tempo estimation. Finally, durations of consecutive ¤ notes are combined to form a "rhythm vector" representing tempo-free relative durations of the notes and treated in the same framework. Significant improvements compared with conventional "quantization" techniques are shown.

Future Directions of Music Cognition, 2021

A commonly-cited reason for the poor performance of automatic chord estimation (ACE) systems within music information retrieval (MIR) is that non-chord tones (i.e., notes outside the supporting harmony) contribute to error during the labeling process (e.g., . Despite the prevalence of machine-learning approaches in MIR, there are cases where alternative approaches provide a simpler alternative while allowing for insights into musicological practices. In this project, we present a statistical model for predicting chord tones based on music theory rules. Our model is currently focused on predicting chord tones in classical music, since composition in this style is highly constrained, theoretically making the placement of chord tones highly predictable. Indeed, music theorists have labeling systems for every variety of non-chord tone, primarily classified by the note's metric position and intervals of approach and departure. Using metric position, duration, and melodic intervals as predictors, we build a statistical model for predicting chord tones using the TAVERN dataset . While our probabilistic approach is similar to other efforts in the domain of automatic harmonic analysis (e.g., Temperley 1997; Temperley & Sleator 1999), our focus is on melodic reduction rather than predicting harmony. However, we hope to pursue applications for ACE in the future. Finally, we implement our melody-reduction model using an existing symbolic visualization tool, to assist with melody reduction and non-chord tone identification for computational musicology researchers and music theorists.

Pattern Recognition Techniques, Technology and Applications, 2008

Cognitive Science, 2008

This study presents a probabilistic model of melody perception, which infers the key of a melody and also judges the probability of the melody itself. (A "melody" is defined here as a sequence of pitches, without rhythmic information.) The model uses Bayesian reasoning. A generative probabilistic model is proposed, based on three principles: 1) melodies tend to remain within a narrow pitch range; 2) note-to-note intervals within a melody tend to be small; 3) notes tend to conform to a distribution (or "key-profile") that depends on the key. The model is tested in three ways: on a key-finding task, on a melodic expectation task, and on an error-detection task.