We formulate tempo tracking in a Bayesian framework where a tempo tracker is modeled as a stochastic dynamical system. The tempo is modeled as a hidden state variable of the system and is estimated by a Kalman filter. The Kalman filter operates on a Tempogram, a wavelet-like multiscale expansion of a real performance. An important advantage of our approach is that it is possible to formulate both off-line or real-time algorithms. The simulation results on a systematically collected set of MIDI piano performances of Yesterday and Michelle by the Beatles shows accurate tracking of approximately %90 of the beats.

We present a probabilistic generarive model for timing deviations in expressive music performance. The structure of the proposed model is equivalent to a switching state space model. The switch variables correspond to discrete note locations as in a musical score. The continuous hidden variables denote the tempo. We formulate two well known music recognition problems, namely tempo tracking and automatic transcription (rhythm quantization) as filtering and maximum a posteriori (MAP) state estimation tasks. Ex- act computation of posterior features such as the MAP state is intractable in this model class, so we introduce Monte Carlo methods for integration and optimization. We compare Markov Chain Monte Carlo (MCMC) methods (such as Gibbs sampling, simulated annealing and iterative improvement) and sequential Monte Carlo methods (particle filters). Our simulation results suggest better results with sequential methods. The methods can be applied in both online and batch scenarios such as tempo tracking and transcription and are thus potentially useful in a number of music applications such as adaptive automatic accompaniment, score typesetting and music information retrieval.

Abstract not found

The thesis proposes an algorithm for the recognition of musical meter from acoustic signals of music. Musical meter is a part of rhythm that is constantly present in music, as it spans the musical time base. The proposed model is capable of finding metrical levels, including the beat and the tatum, in real time from a musical audio signal. The model comprises four main components: an onset detector, a tatum estimator, a phenomenal accent model, and a beat estimator. The onset detector finds distinct sound onsets from an acoustic signal, using multiband signal processing. After this, the tatum, which is the lowest metrical level, is computed from onset times. Phenomenal accents are computed from a set of 16 acoustic signal features using Bayesian pattern recognition. The tatum and the accents then yield the beat. The proposed model operates causally and is able to respond to tempo changes. The design of the model aims at generality in regard to musical genres, and thus the model is trained and tested using 330 music excerpts from multiple genres. The model performance varies according to the rhythmic difficulty of the input signal. Most pop/rock music poses no problems for the algorithm, while classical music and expressive jazz pieces are intractable. The model produces more errors than Eric Scheirer's beat tracker, but at the same time it follows more metrical levels than Scheirer's model. The results of this thesis are directly applicable in music production and post-processing. The access to musical time enables new levels of productivity and automation in both music software and hardware. Meter-synchronized comparison, mixing, and editing of pieces of music is possible. Robust meter recognition is a vital component of music information retrieval applications.

Abstract not found

model is to see whether it shows a good fit with the empirical data, recent literature on theory testing and model selection criticizes the assumption that this is actually strong evidence for the validity of a model. This article presents a case study from music cognition (modeling the ritardandi in music performance) and compares two families of computational models (kinematic and perceptual) using three different model selection criteria: goodness-of-fit, model simplicity, and the degree of surprise in the predictions. In the light of what counts as strong evidence for a model’s validity— namely that it makes limited range, nonsmooth, and relatively surprising predictions—the perception-based model is preferred over the kinematic model.

Automated musical accompaniment of human performers often requires an agent be able to follow a musical score with similar facility to that of a human performer. Systems described in the literature represent musical scores in a way that assumes no large-scale structural variation of the piece during performance. If the performer deviates from the expected path by skipping or repeating a section, the system may become lost. We describe a way to automatically generate a Markov model from a written score that models the score form, and an on-line algorithm to align a performance to a score. The resulting system can follow performances that take alternate paths through the score without losing its place. We compare the performance of our system to that of sequence-based score followers on a melodic corpus of 98 Jazz melodies. Results show that explicitly representing the branching structure of a score significantly improves score following when the branch a performer may take is unknown beforehand.

Live performance using a combination of musicians and electronics has so far been problematic art. Getting the two components to create a coherent sound is a major problem. Automatic conductor following is one way to do this. In this case, the conductor can control the electronics as one controls the musicians. Conductor following requires multiple information technology tools. Pattern recognition techniques are used to recognize the gestures (movements) conductor makes. The identified gestures need to be matched with general knowledge of the conducting technique and the knowledge of the piece being conducted. Finally the conductor follower must create musical, expressive output. Little research has been done on this field. Most systems only track the tempo of the piece. Few can track dynamics or finer nuances like staccato. Our main research goal is interpreting the tempo of the conductor and reacting musically in real-time. A multi-layer perceptron network is used to estimate the phase ...

Understanding and following musical improvisations with a computer requires methods different from those needed when following score-based musical performances. This paper discusses questions related to interactive music systems for the recognition and accompaniment of tonal improvisations. This includes discussion about rhythmic parsing and harmonic analysis with a recursive Bayesian classifier. Finally, a real-time MIDI system for identifying and following tonal improvisations is described.

The performance of autocorrelation-based metre induction was tested with two large collections of folk melodies, consisting of approximately 13,000 melodies in MIDI file format, for which the correct metres were available. The analysis included a number of melodic accents assumed to contribute to metric structure. The performance was measured by the proportion of melodies whose metre was correctly classified by Multiple Discriminant Analysis. Overall, the method predicted notated metre with an accuracy of 75 % for classification into nine categories of metre. The most frequent confusions were made within the groups of duple and triple/compound metres, whereas confusions across these groups where significantly less frequent. In addition to note onset locations and note durations, Thomassen's melodic accent was found to be an important predictor of notated metre.