Richard Kronland-Martinet

ANALYSIS OF SOUND PATTERNS THROUGH WAVELET TRANSFORMS

This paper starts with a brief discussion of so-called wavelet transforms, i.e., decompositions of arbitrary signals into localized contributions labelled by a scale parameter. The main features of the method are first illustrated through simple mathematical examples. Then we present the first applications of the method to the recognition and visualisation of characteristic features of speech and of musical sounds.

Asymptotic wavelet and Gabor analysis: extraction of instantaneous frequencies

The behavior of the continuous wavelet and Gabor coefficients in the asymptotic limit using stationary phase approximations are investigated. In particular, it is shown how, under some additional assumptions, these coefficients allow the extraction of some characteristics of the analyzed signal, such as frequency and amplitude modulation laws. Applications to spectral line estimations and matched filtering are briefly discussed. >

A Real-Time Algorithm for Signal Analysis with the Help of the Wavelet Transform

The purpose of this paper is to present a real-time algorithm for the analysis of time-varying signals with the help of the wavelet transform. We shall briefly describe this transformation in the following. For more details, we refer to the literature [1].

Reading and Understanding Continuous Wavelet Transforms

One of the aims of wavelet transforms is to provide an easily interpretable visual representation of signals. This is a prerequisite for applications such as selective modifications of signals or pattern recognition.

Characterization of acoustic signals through continuous linear time-frequency representations

One important field in the framework of computer music concerns the modeling of sounds. In order to design digital models mirroring as closely as possible a real sound and permitting in addition intimate transformations by altering the synthesis parameters, we look for a signal model based on additive synthesis whose parameters are estimated by the analysis of real sounds. This model is relevant from both the physical and perceptual points of view, especially when the sound to be analyzed comes from a musical instrument. We present some techniques, mostly unpublished, based on time-frequency representations which make possible the estimation of relevant parameters such as frequency and amplitude modulation laws corresponding to each spectral component of the sound. The techniques described extend the results presented by Delprat et al. (see IEEE Trans. Inform. Theory, vol.38, p.644-65, March 1992). These methods are then transposed to broadband signals, allowing the characterization of transients.

The simulation of piano string vibration: From physical models to finite difference schemes and digital waveguides

A model of transverse piano string vibration, second order in time, which models frequency-dependent loss and dispersion effects is presented here. This model has many desirable properties, in particular that it can be written as a well-posed initial-boundary value problem (permitting stable finite difference schemes) and that it may be directly related to a digital waveguide model, a digital filter-based algorithm which can be used for musical sound synthesis. Techniques for the extraction of model parameters from experimental data over the full range of the grand piano are discussed, as is the link between the model parameters and the filter responses in a digital waveguide. Simulations are performed. Finally, the waveguide model is extended to the case of several coupled strings.

A 3-D Immersive Synthesizer for Environmental Sounds

Nowadays, interactive 3-D environments tend to include both synthesis and spatialization processes to increase the realism of virtual scenes. In typical systems, audio generation is created in two stages: first, a monophonic sound is synthesized (generation of the intrinsic timbre properties) and then it is spatialized (positioned in its environment). In this paper, we present the design of a 3-D immersive synthesizer dedicated to environmental sounds, and intended to be used in the framework of interactive virtual reality applications. The system is based on a physical categorization of environmental sounds (vibrating solids, liquids, aerodynamics). The synthesis engine has a novel architecture combining an additive synthesis model and 3-D audio modules at the prime level of sound generation. An original approach exploiting the synthesis capabilities for simulating the spatial extension of sound sources is also presented. The subjective results, evaluated with a formal listening test, are discussed. Finally, new control strategies based on a global manipulation of timbre and spatial attributes of sound sources are introduced.

Analysis-synthesis of impact sounds by real-time dynamic filtering

This paper presents a sound synthesis model that reproduces impact sounds by taking into account both the perceptual and the physical aspects of the sound. For that, we used a subtractive method based on dynamic filtering of noisy input signals that simulates the damping of spectral components. The resulting sound contains the perceptual characteristics of an impact on a given material. Further, the addition of very few modal contributions-using additive or banded digital waveguide synthesis-together with a bandpass filtering taking into account the interaction with the excitator, allows realistic impact sounds to be synthesized. The synthesis parameters can be linked to a perceptual notion of material and geometry of the sounding object. To determine the synthesis parameters, we further address the problem of analysis-synthesis aiming at reconstructing a given impact sound. The physical parameters are extracted through a time-scale analysis of natural sounds. Examples are presented for sounds generated by impacting plates made of different materials and a piano soundboard.

Controlling the Perceived Material in an Impact Sound Synthesizer

In this paper, we focused on the identification of the perceptual properties of impacted materials to provide an intuitive control of an impact sound synthesizer. To investigate such properties, impact sounds from everyday life objects, made of different materials (wood, metal and glass), were recorded and analyzed. These sounds were synthesized using an analysis-synthesis technique and tuned to the same chroma. Sound continua were created to simulate progressive transitions between materials. Sounds from these continua were then used in a categorization experiment to determine sound categories representative of each material (called typical sounds). We also examined changes in electrical brain activity (using event related potentials (ERPs) method) associated with the categorization of these typical sounds. Moreover, acoustic analysis was conducted to investigate the relevance of acoustic descriptors known to be relevant for both timbre perception and material identification. Both acoustic and electrophysiological data confirmed the importance of damping and highlighted the relevance of spectral content for material perception. Based on these findings, controls for damping and spectral shaping were tested in synthesis applications. A global control strategy, with a three-layer architecture, was proposed for the synthesizer allowing the user to intuitively navigate in a “material space” and defining impact sounds directly from the material label. A formal perceptual evaluation was finally conducted to validate the proposed control strategy.

The evocative power of sounds: Conceptual priming between words and nonverbal sounds

Two experiments were conducted to examine the conceptual relation between words and nonmeaningful sounds. In order to reduce the role of linguistic mediation, sounds were recorded in such a way that it was highly unlikely to identify the source that produced them. Related and unrelated sound–word pairs were presented in Experiment 1 and the order of presentation was reversed in Experiment 2 (word–sound). Results showed that, in both experiments, participants were sensitive to the conceptual relation between the two items. They were able to correctly categorize items as related or unrelated with good accuracy. Moreover, a relatedness effect developed in the event-related brain potentials between 250 and 600 msec, although with a slightly different scalp topography for word and sound targets. Results are discussed in terms of similar conceptual processing networks and we propose a tentative model of the semiotics of sounds.