The Timbre Toolbox: Extracting audio descriptors from musical signals

Abstract

The analysis of musical signals to extract audio descriptors that can potentially characterize their timbre has been disparate and often too focused on a particular small set of sounds. The Timbre Toolbox provides a comprehensive set of descriptors that can be useful in perceptual research, as well as in music information retrieval and machine-learning approaches to content-based retrieval in large sound databases. Sound events are first analyzed in terms of various input representations (short-term Fourier transform, harmonic sinusoidal components, an auditory model based on the equivalent rectangular bandwidth concept, the energy envelope). A large number of audio descriptors are then derived from each of these representations to capture temporal, spectral, spectrotemporal, and energetic properties of the sound events. Some descriptors are global, providing a single value for the whole sound event, whereas others are time-varying. Robust descriptive statistics are used to characterize the time-varying descriptors. To examine the information redundancy across audio descriptors, correlational analysis followed by hierarchical clustering is performed. This analysis suggests ten classes of relatively independent audio descriptors, showing that the Timbre Toolbox is a multidimensional instrument for the measurement of the acoustical structure of complex sound signals.

Perceptual Dimensions of Short Audio Clips and Corresponding Timbre Features

“…Indeed, because timbral perception does not require formalised musical knowledge, individuals could be expected to vary in the information they can access for this task purely on the basis of what they have previously listened to, and to what extent. We will look at three other datasets-including shorter, 400ms clips-and explore other features, for example those provided by Peeters and colleagues' recently published toolbox [28], as well as standard MFCC coefficients and spectral centroid-based measures.…”

Organizing a sonic space through vocal imitations

“…In the area of music information retrieval a lot of research has been devoted to extract audio descriptors (or features) that could concisely represent sound and music [24]. Several software libraries are available to easily extract brightness, spectral flux, and other descriptors from a given soundfile and to collect statistical descriptors from them.…”

Feature-based characterization of violin timbre

“…We extract features presented in [7] and others typically used in the Music Information Retrieval field: the four spectral moments (Centroid, Spread, Skewness, Kurtosis), other spectral indicators (Brightness, Rolloff, Flux, Irregularity, Flatness), features related to the distribution of harmonics (Tristimulus coefficients, Odd-Even ratio), two vectorial features describing the spectrum shape (Mel-Frequency Cepstral Coefficients, MFCC, and Spectral Contrast [12]) and some temporal features (Attack time, Attack slope, RMS energy and Zero-crossing rate). We refer to [13], [14], [15], [16], [17] for a detailed explanation of these features.…”

“…Low-level features are objective descriptors devoted to capture specific aspects of the sound. Since the timbre is the combination of many factors ranging from acoustics to perception, feature-based analysis resulted particularly suitable for musical instruments characterization [6] [7] [8]. In [9] the authors take advantage of feature-based analysis for a musical instruments recognition scenario.…”