The notion of perceptual features is introduced for describing general music properties based on human perception. This is an attempt at rethinking the concept of features, aiming to approach the underlying human perception mechanisms. Instead of using concepts from music theory such as tones, pitches, and chords, a set of nine features describing overall properties of the music was selected. They were chosen from qualitative measures used in psychology studies and motivated from an ecological approach. The perceptual features were rated in two listening experiments using two different data sets. They were modeled both from symbolic and audio data using different sets of computational features. Ratings of emotional expression were predicted using the perceptual features. The results indicate that (1) at least some of the perceptual features are reliable estimates; (2) emotion ratings could be predicted by a small combination of perceptual features with an explained variance from 75% to 93% for the emotional dimensions activity and valence; (3) the perceptual features could only to a limited extent be modeled using existing audio features. Results clearly indicated that a small number of dedicated features were superior to a "brute force" model using a large number of general audio features.

This article presents a polyphonic pitch tracking system that is able to extract both framewise and note-based estimates from audio. The system uses several artificial neural networks trained individually in a deep layered learning setup. First, cascading networks are applied to a spectrogram for framewise fundamental frequency (f0) estimation. A sparse receptive field is learned by the first network and then used as a filter kernel for parameter sharing throughout the system. The f0 activations are connected across time to extract pitch contours. These contours define a framework within which subsequent networks perform onset and offset detection, operating across both time and smaller pitch fluctuations at the same time. As input, the networks use, e.g., variations of latent representations from the f0 estimation network. Finally, erroneous tentative notes are removed one by one in an iterative procedure that allows a network to classify notes within a correct context. The system was evaluated on four public test sets: MAPS, Bach10, TRIOS, and the MIREX Woodwind quintet and achieved state-of-the-art results for all four datasets. It performs well across all subtasks f0, pitched onset, and pitched offset tracking.

A system is proposed in which rhythmic representations are used to model the perception of tempo in music. The system can be understood as a five-layered model, where representations are transformed into higher-level abstractions in each layer. First, source separation is applied (Audio Level), onsets are detected (Onset Level), and interonset relationships are analyzed (Interonset Level). Then, several high-level representations of rhythm are computed (Rhythm Level). The periodicity of the music is modeled by the cepstroid vector-the periodicity of an interonset interval (IOI)-histogram. The pulse strength for plausible beat length candidates is defined by computing the magnitudes in different IOI histograms. The speed of the music is modeled as a continuous function on the basis of the idea that such a function corresponds to the underlying perceptual phenomena, and it seems to effectively reduce octave errors. By combining the rhythmic representations in a logistic regression framework, the tempo of the music is finally computed (Tempo Level). The results are the highest reported in a formal benchmarking test (2006-2013), with a P-Score of 0.857. Furthermore, the highest results so far are reported for two widely adopted test sets, with an Acc1 of 77.3% and 93.0% for the Songs and Ballroom datasets.

Vocal sound imitations provide a new challenge for understanding the coupling between articulatory mechanisms and the resulting audio. In this study, the classification of three articulatory categories, phonation, supraglottal myoelastic vibrations, and turbulence, have been modeled from audio recordings. Two data sets were assembled, consisting of different vocal imitations by four professional imitators and four non-professional speakers in two different experiments. The audio data were manually annotated by two experienced phoneticians using a detailed articulatory description scheme. A separate set of audio features was developed specifically for each category using both time-domain and spectral methods. For all time-frequency transformations, and for some secondary processing, the recently developed Auditory Receptive Fields Toolbox was used. Three different machine learning methods were applied for predicting the final articulatory categories. The result with the best generalization was found using an ensemble of multilayer perceptrons. The cross-validated classification accuracy was 96.8% for phonation, 90.8% for supraglottal myoelastic vibrations, and 89.0% for turbulence using all the 84 developed features. A final feature reduction to 22 features yielded similar results. V

By varying the dynamics in a musical performance, the musician can convey structure and different expressions. Spectral properties of most musical instruments change in a complex way with the performed dynamics, but dedicated audio features for modeling the parameter are lacking. In this study, feature extraction methods were developed to capture relevant attributes related to spectral characteristics and spectral fluctuations, the latter through a sectional spectral flux. Previously, ground truths ratings of performed dynamics had been collected by asking listeners to rate how soft/loud the musicians played in a set of audio files. The ratings, averaged over subjects, were used to train three different machine learning models, using the audio features developed for the study as input. The highest result was produced from an ensemble of multilayer perceptrons with an R 2 of 0.84. This result seems to be close to the upper bound, given the estimated uncertainty of the ground truth data. The result is well above that of individual human listeners of the previous listening experiment, and on par with the performance achieved from the average rating of six listeners. Features were analyzed with a factorial design, which highlighted the importance of source separation in the feature extraction.