We introduce a new neurofeedback approach, which allows users to manipulate expressive parameters in music performances using their emotional state, and we present the results of a pilot clinical experiment applying the approach to alleviate depression in elderly people. Ten adults (9 female and 1 male, mean = 84, SD = 5.8) with normal hearing participated in the neurofeedback study consisting of 10 sessions (2 sessions per week) of 15 min each. EEG data was acquired using the Emotiv EPOC EEG device. In all sessions, subjects were asked to sit in a comfortable chair facing two loudspeakers, to close their eyes, and to avoid moving during the experiment. Participants listened to music pieces preselected according to their music preferences, and were encouraged to increase the loudness and tempo of the pieces, based on their arousal and valence levels. The neurofeedback system was tuned so that increased arousal, computed as beta to alpha activity ratio in the frontal cortex corresponded to increased loudness, and increased valence, computed as relative frontal alpha activity in the right lobe compared to the left lobe, corresponded to increased tempo. Pre and post evaluation of six participants was performed using the BDI depression test, showing an average improvement of 17.2% (1.3) in their BDI scores at the end of the study. In addition, an analysis of the collected EEG data of the participants showed a significant decrease of relative alpha activity in their left frontal lobe (p = 0.00008), which may be interpreted as an improvement of their depression condition.

The automatic assessment of music performance has become an area of increasing interest due to the growing number of technology-enhanced music learning systems. In most of these systems, the assessment of musical performance is based on pitch and onset accuracy, but very few pay attention to other important aspects of performance, such as sound quality or timbre. This is particularly true in violin education, where the quality of timbre plays a significant role in the assessment of musical performances. However, obtaining quantifiable criteria for the assessment of timbre quality is challenging, as it relies on consensus among the subjective interpretations of experts. We present an approach to assess the quality of timbre in violin performances using machine learning techniques. We collected audio recordings of several tone qualities and performed perceptual tests to find correlations among different timbre dimensions. We processed the audio recordings to extract acoustic features for training tone-quality models. Correlations among the extracted features were analyzed and feature information for discriminating different timbre qualities were investigated. A real-time feedback system designed for pedagogical use was implemented in which users can train their own timbre models to assess and receive feedback on their performances.

We present a machine learning approach to automatically generate expressive (ornamented) jazz performances from un-expressive music scores. Features extracted from the scores and the corresponding audio recordings performed by a professional guitarist were used to train computational models for predicting melody ornamentation. As a first step, several machine learning techniques were explored to induce regression models for timing, onset, and dynamics (i.e. note duration and energy) transformations, and an ornamentation model for classifying notes as ornamented or non-ornamented. In a second step, the most suitable ornament for predicted ornamented notes was selected based on note context similarity. Finally, concatenative synthesis was used to automatically synthesize expressive performances of new pieces using the induced models. Supplemental online material for this article containing musical examples of the automatically generated ornamented pieces can be accessed at

No abstract

Expert musicians introduce expression in their performances by manipulating sound properties such as timing, energy, pitch, and timbre. Here, we present a data driven computational approach to induce expressive performance rule models for note duration, onset, energy, and ornamentation transformations in jazz guitar music. We extract high-level features from a set of 16 commercial audio recordings (and corresponding music scores) of jazz guitarist Grant Green in order to characterize the expression in the pieces. We apply machine learning techniques to the resulting features to learn expressive performance rule models. We (1) quantitatively evaluate the accuracy of the induced models, (2) analyse the relative importance of the considered musical features, (3) discuss some of the learnt expressive performance rules in the context of previous work, and (4) assess their generailty. The accuracies of the induced predictive models is significantly above base-line levels indicating that the audio performances and the musical features extracted contain sufficient information to automatically learn informative expressive performance patterns. Feature analysis shows that the most important musical features for predicting expressive transformations are note duration, pitch, metrical strength, phrase position, Narmour structure, and tempo and key of the piece. Similarities and differences between the induced expressive rules and the rules reported in the literature were found. Differences may be due to the fact that most previously studied performance data has consisted of classical music recordings. Finally, the rules' performer specificity/generality is assessed by applying the induced rules to performances of the same pieces performed by two other professional jazz guitar players. Results show a consistency in the ornamentation patterns between Grant Green and the other two musicians, which may be interpreted as a good indicator for generality of the ornamentation rules.

In this paper, we propose a system that enables nonmusicians to enjoy improvisation just by drawing a melodic outline on the piano-roll display. Once the user draws a melodic outline, the system immediately generates a melody using a genetic algorithm, in which the fitness function is calculated based on the similarity to the outline, an N-gram probability, and entropy. Experimental results show that generated melodies have quality similar to performances by non-expert human performers.

No abstract

The automatic assessment of music performance has become an area of special interest due to the increasing amount of technology-enhanced music learning systems. However, in most of these systems the assessment of the musical performance is based on the accuracy of onsets and pitch, paying little attention to other relevant aspects of performance. In this paper we present a preliminary study to assess the quality of violin performance using machine learning techniques. We collect recording examples of selected violin exercises varying from expert to amateur performances. We process the audio signal to extract features to train models using clustering based on Dynamic Time Warping distance. The quality of new performances is evaluated based on the level of match/miss-match to each of the recorded training examples.