Abstract not found

Abstract — In a situation where multiple sound sources are concurrently active, the signals of the individual sources often overlap in time and in frequency. This is particularly likely for voiced instruments where the frequencies of some of the partials of one single note coincide with the frequencies of some of the partials of another instrument playing a harmonically related note. A source separation algorithm suitable for musical applications must address the problem of overlapping partials. A method is proposed for the separation of overlapping narrow-band partials in multi-channel mixtures. The method is based on the observation that, for many instruments, all the partials of a single note have similar temporal envelopes. For narrow band partials these similarities can be exploited in order to estimate demixing matrices in the frequency domain. Effectively, one can recover estimates of the original partials from a multi-channel mixture where they overlap. The method is computationally efficient in that it works on highly downsampled narrow frequency bands. It performs well for closely spaced and colliding partials, and (to some extent) also for frequency modulations such as vibrato effects. Index Terms — source separation, harmonic instruments, overlapping partials. I.

This paper contains a review of the issues surrounding sinusoidal parameter estimation which is a vital part of many audio manipulation algorithms. A number of algorithms which use the phase of the Fourier transform for estimation (e.g. [1]) are explored and shown to be identical. Their performance against a classical interpolation estimator [2] and comparison with the Cramer Rao Bound (CRB) is presented. Component detection is also considered and various methods of improving these algorithms are discussed.

A modification to the hybrid sinusoidal model is proposed for the purpose of high-quality audio coding. In our proposal the amplitude envelope of each harmonic partial is modeled by a narrowband complex signal. Such representation incorporates most of the signal energy associated with sinusoidal components, including that related to frequency estimation and quantization errors. It also takes into account the natural width of each spectral line. The advantages of such model extension are a more straightforward and robust representation of the deterministic component and a clean stochastic residual without ghost sinusoids. The reconstructed signal is virtually free from harmonic artifacts and more natural sounding. We propose to encode the complex envelopes by the means of MCLT transform coefficients with coefficient interleave across partials within an MPEG-like coding scheme. We show some experimental results with high compression efficiency achieved. 1.

deste artigo sem autorização expressa da AES Brasil.

Peak picking algorithms play an important role in audio analysis and synthesis methods based on sinusoidal modelling. Peak detection aims at selecting only those peaks corresponding to genuine resonant components present in the signal, while rejecting noise-induced ones. This paper investigates the performance of four threshold-based schemes for peak detection: two-pass split window, low-order autoregressive model, a nonlinear recursive filter and a stochastic spectrum estimator. The results of performance comparisons measured under a common metric are reported for a set of specific experimental setups. 0

This paper proposes an algorithm for studying spectral contents of pitched sounds in real-world recordings. We assume that the 2 nd-order difference, w.r.t. partial index, of a pitched sound is bounded by some small positive value, rather than equal to 0 in a perfect harmonic case. Given a spectrum and a fundamental frequency f0, the algorithm searches the spectrum for partials that can be associated with f0 by dynamic programming. In section 3 a background-foreground model is plugged into the algorithm to make it work with reverberant background, such as in a piano recording. In section 4 we illustrate an application of the algorithm in which a multipitch scoring machine, which involves special processing for close or shared partials, is coupled with a tree searching method for polyphonic transcription task. Results are evaluated on the traditional note level, as well as on a partial-based sub-note level.

The analysis stage of spectral models based on phase vocoder techniques uses short-time spectra of the sound studied. This article introduces an original approach for determining which peaks of these spectra correspond to sinusoids in order to prevent bad connections between partials in successive frames. This approach differs from the existing ones because it is based on the statistical analysis of the intensity uctuations and not on the results of the Fourier transform. Experiments show that the methods proposed greatly improve the selection of peaks, especially in noisy conditions. Moreover, the measure computed is normalized and thus does not depend on the volume of the sound analyzed.

When instruments play together, different partials will often overlap in time and frequency. This is particularly likely for harmonic instruments. We present a new method for the separation of overlapping partials in multi-channel mixtures. The method is based on the obsewation that when a harmonic instrument plays a note. all the partials have similar shapes, i.e. common onset. offset. anplitude and frequency modulation. For nmow band partidls we devise a method to estimate a demixing matrix that can recover the original Source partials from the multi-channel mixture where they overlap. The method is computationally efficient in that it works on highly downsampled narrow frequency bands and it performs equally well for closely spaced partials as for crossing partials, e.g. due to frequency modulations such as vibrato effects. It is able to separate partials in mixtures with a high number of or,erlapping panials, such as two instruments playing notes where the fundamental frequencies are in fifth (32) or octave (2:l) relation. 1.

October, 2007“I certify that this thesis and the research to which it refers are the product of my own work, and that any ideas or quotations from the work of other people, published or otherwise, are fully acknowledged in accordance with the standard referencing practices of the discipline. This thesis presents the theory, implementation and applications of the harmonic sinusoid modeling of pitched audio events. Harmonic sinusoid modeling is a parametric model that expresses an audio signal, or part of an audio signal, as the linear combination of concurrent slow-varying sinusoids, grouped together under harmonic frequency constraints. The harmonic sinusoid modeling is an extension of the sinusoid modeling, with the additional frequency constraints so that it is capable to directly model tonal sounds. This enables applications such as object-oriented audio manipulations, polyphonic transcription, instrument/singer recognition with background music, etc. The modeling system consists of an analyzer and a synthesizer. The analyzer