Search

In this paper we introduce the multiple-description ℓ1-compression problem: minimize kz1k1+λkz2k1 subject to the three distortion constraints kA1z1 −yk2 ≤ δ1, kA2z2 −yk2 ≤ δ2, and k1/2(A1z1 + A2z2) − yk2 ≤ γ. This problem formulation is interesting in, e.g., ad-hoc networks where packets can be lost. If a description (z2) is lost in the network and only one description is received (z1), it is still possible to decode and obtain a signal quality equal or better than described by the parameter δ1 (and vice versa). If both descriptions are received, the quality is determined by the parameter γ. This problem is difficult to solve using first-order projection methods due to the intersecting second-order cones. However, we show that by recasting the problem into its dual form, one can avoid the difficulties due to conflicting fidelity constraints. We then show that efficient first-order ℓ1-compression methods are applicable, which makes it possible to solve large scale problems, e.g., multiple-description ℓ1-compression of video.

This thesis addresses the fundamental question in room acoustic signal processing concerning the appropriateness of different parametric models for room acoustics. In order to improve the perceived sound quality, room acoustic signal processing algorithms require the acoustic response of the room to be represented by means of parametric models and to be identified from the input and output signals of the room acoustic system. In particular, a good model should be both accurate, thus capturing those features of room acoustics that are physically and perceptually most relevant, and efficient, so that it can be implemented as a digital filter and used in practical tasks. The main goal of this thesis is then to develop accurate yet efficient parametric models for room acoustics, based on recent advances in system identification and numerical optimization, and on the physical understanding of room acoustics. Orthonormal basis functions (OBF) models are investigated, whose properties, such as orthogonality and scalability, are exploited in the development of iterative flexible algorithms, providing a significant reduction in the number of parameters compared to conventional parametric models. Improvements were obtained also in the room acoustic system identification framework using OBF adaptive filters, which present interesting properties in terms of error performance and convergence of the filter coefficients. A scalable iterative approach is adopted for the pole estimation, providing a reduction in the filter order, thus helping in addressing some of the issues encountered in RASE applications, such as echo path undermodeling in acoustic echo cancellation, or frequency allocation in inverse filtering for digital equalization. Particular attention is addressed to the low-frequency region of modal resonances, where the acoustics of small rooms is typically more problematic. In this regard, the issues of measuring RIRs at low frequencies, mostly related to high ambient noise and to the nonlinear distortions produced by the subwoofer, are addressed and a novel procedure for estimating the frequency-dependent reverberation time is suggested. Finally, an effective automatic procedure for the design of a low-complexity parametric equalizer for loudspeaker/room response equalization is proposed, and guidelines for the implementation of an existing solution for nonminimum-phase multi-channel equalization of car cabin acoustics are given. status: published

Systems built using deep learning neural networks trained on low-level spectral periodicity features (DeSPerF) reproduced the most “ground truth” of the systems submitted to the MIREX 2013 task, “Audio Latin Genre Classification.” To answer why this was the case, we take a closer look at the behavior of a DeSPerF system we create and evaluate using the benchmark dataset BALLROOM. We find through time stretching that this DeSPerF system appears to obtain a high figure of merit on the task of music genre recognition because of a confounding of tempo with “ground truth” in BALLROOM. This observation motivates several predictions.

Multimedia consumer electronics are nowadays everywhere from teleconferencing, hands-free communications, in-car communications to smart TV applications and more. We are living in a world of telecommunication where ideal scenarios for implementing these applications are hard to find. Instead, practical implementations typically bring many problems associated to each real-life scenario. This thesis mainly focuses on two of these problems, namely, acoustic echo and acoustic feedback. On the one hand, acoustic echo cancellation (AEC) is widely used in mobile and hands-free telephony where the existence of echoes degrades the intelligibility and listening comfort. On the other hand, acoustic feedback limits the maximum amplification that can be applied in, e.g., in-car communications or in conferencing systems, before howling due to instability, appears. Even though AEC and acoustic feedback cancellation (AFC) are functional in many applications, there are still open issues. This means that many of the issues associated to practical AEC and AFC are overlooked. In this thesis, we contribute to the development of a number of algorithms to tackle the main issues associated to AEC and AFC namely, (1) that very long room impulse responses (RIRs) make standard adaptive filters converge slowly and lead to a high computational complexity, (2) that double-talk (DT) situations in AEC and model mismatch in AFC distort the near-end signal , and (3) that the nonlinear response of some of the elements forming part of the audio chain makes linear adaptive filters fail. In the view of computational complexity reduction, we consider introducing the concept of common-acoustical-pole room modeling into AEC. To this end, we perform room transfer function (RTF) equalization by compensating for the main acoustic resonances common to multiple RTFs in the room. We discuss the utilization of different norms (i.e., 2-norm and 1-norm) and models (i.e., all-pole and pole-zero) for RTF modeling and equalization. A computationally cheap extension from single-microphone AEC to multi-microphone AEC is then presented for the case of a single loudspeaker. The RTF models used for multi-microphone AEC share a fixed common denominator polynomial, which is calculated off-line by means of a multi-channel warped linear prediction. This allows high computational savings. In the context of acoustic feedback control, we develop a method for acoustic howling suppression based on adaptive notch filters (ANF) with regularization (RANF). This method achieves frequency tracking, howling suppression and improved howling detection in a one-stage scheme. The RANF approach to howling suppression introduces minimal processing delay and minimal complexity, in contrast to non-parametric block-based methods that feature a non-parametric frequency analysis in a two-stage scheme.To tackle the issue of robustness to double-talk (DT) in AEC and robustness to model mismatch in AFC, a new adaptive filtering framework is proposed. It is based on the frequency-domain adaptive filtering (FDAF) implementation of the so-called PEM-AFROW (FDAF-PEM-AFROW) algorithm. We show that FDAF-PEM-AFROW is related to the best, i.e., minimum-variance, linear unbiased estimate (BLUE) of the echo path. We derive and define the instantaneous pseudo-correlation (IPC) measure between the near-end signal and the loudspeaker signal. The IPC measure serves as an indication of the effect of a DT situation occurring during adaptation due to the correlation between these two signals. Based on the good results obtained using FDAF-PEM-AFROW, we derive a practical and computationally efficient algorithm for DT-robust AEC and for AFC. The proposed algorithm features two modifications in the FDAF-PEM-AFROW: (a) the WIener variable Step sizE (WISE), and (b) the GRAdient Spectral variance Smoothing (GRASS), leading to the WISE-GRASS-FDAF-PEM-AFROW. The WISE modification is implemented as a single-channel noise reduction Wiener filter where the Wiener filter gain is used as a variable step size in the adaptive filter. On the other hand, the GRASS modification aims at reducing the variance in the noisy gradient estimates based on time-recursive averaging of instantaneous gradients.In the last part of this thesis, the nonlinear response of (active) loudspeakers forming part of the audio chain is studied. We consider the description of odd and even nonlinearities in (active) loudspeakers by means of periodic random-phase multisine signals. The fact that the odd nonlinear contributions are more predominant than the even ones implies that at least a 3rd-order nonlinear model of the loudspeaker should be used. Therefore, we consider the identification and validation of a model of the loudspeaker using several linear-in-the-parameters nonlinear adaptive filters, namely, Hammerstein and Legendre polynomial filters of various orders, and a simplified $3$rd-order Volterra filter of various lengths. In our measurement set-up, the obtained results imply however, that a $3$rd-order nonlinear filter fails to capture all the nonlinearities, meaning that odd and even nonlinear contributions are produced by higher-order nonlinearities. High-order Volterra filters are impractical in AEC and AFC due to their large computational complexity together with inherently slow convergence. On the other hand, the kernel affine projection algorithm (KAPA) has been successfully applied to many areas in signal processing but not yet to nonlinear AEC (NLAEC). In KAPA, and kernel methods in general, the kernel trick is applied to work implicitly in a high-dimensional (possibly infinite-dimensional) space without having to transform the input data into this space. This is one of the most appealing characteristics of kernel methods, as opposed to nonlinear adaptive filters requiring explicit nonlinear expansions of the input data as, for instance, Volterra filters. In fact, all computations can be done by evaluating the kernel function in the input space. This fact provides powerful modeling capabilities to kernel adaptive algorithms where the computational complexity will be determined by the input dimension, independent of the order of the nonlinearity. Our contributions in this context are to apply KAPA to the NLAEC problem, to develop a sliding-window leaky KAPA (SWL-KAPA) that is well suited for NLAEC applications, and to propose a suitable kernel function, consisting of a weighted sum of a linear and a Gaussian kernel. nrpages: 257 status: published