The data acquired at different scalp EEG electrodes when human subjects are exposed to speech stimuli are highly redundant. The redundancy is partly due to volume conduction effects and partly due to localized regions of the brain synchronizing their activity in response to the stimuli. In a competing talker scenario, we use a recent measure of directed redundancy to assess the amount of redundant information that is causally conveyed from the attended stimuli to the left temporal region of the brain. We observe that for the attended stimuli, the transfer entropy as well as the directed redundancy is proportional to the correlation between the speech stimuli and the reconstructed signal from the EEG signals. This demonstrates that both the rate as well as the rate-redundancy are inversely proportional to the distortion in neural speech tracking. Thus, a greater rate indicates a greater redundancy between the electrode signals, and a greater correlation between the reconstructed signal and the attended stimuli. A similar relationship is not observed for the distracting stimuli.

Hearing aid (HA) users often experience increased listening effort, particularly in noisy environments. While noise reduction (NR) algorithms aim to alleviate this, traditional electroencephalography (EEG) methods based on power analysis have limited success in assessing the listening effort in this population. This study proposes a novel method using a whole-head synchronization map analysis that uses local connectivity, a measure of statistical dependencies within localized brain regions. We use EEG electrodes to define a region based on the surrounding electrodes in the first-order neighborhood. This approach was tested using EEG data from 22 HA users with active or inactive NR engaged in a continuous speech-in-noise (SiN) task at low (3dB) and high (8dB) signal-to-noise ratio (SNR) levels. Whole-head synchronization was quantified using circular omega complexity (COC), a multivariate phase synchrony measure. Results showed increased local connectivity in the alpha band (8–12 Hz) within frontal and occipital regions during SiN condition compared to the background noise-only (NO) condition. Furthermore, NR activation impacted the synchronization map differently at the two SNRs of the experiment, with greater effect observed at low SNR, primarily in the left parietal region and alpha band. This behavior is in line with that of existing measures for listening effort, and therefore suggests that EEG local connectivity analysis holds promise as a tool for objectively assessing listening effort in HA users, especially in challenging listening environments.

We present a joint design of sound zone control filters and robust audio coding for wireless low frequency sound zones. The audio signal is filtered using sound zone control filters and encoded using a multiple-description coder. The control filters and the multiple-description coder are combined in a nested loop. The inner loop performs filtering for sound zone control and generates multiple descriptions using oversampling and closed-loop prediction. The outer loop performs noise shaping and guarantees a trade-off between robustness and quality of the descriptions. A closed-form expression for the optimal sound-zone control filters are provided, and a simulation study demonstrates that even at moderate packet loss rates, a significant gain is possible compared to not using multiple descriptions.

It was recently shown that the combination of source prediction, two-times oversampling, and noise shaping, can be used to obtain a robust (multiple-description) audio coding framework for networks with packet loss probabilities less than 10%. Specifically, it was shown that audio signals could be encoded into two descriptions (packets), which were separately sent over a communication channel. Each description yields a desired performance by itself, and when they are combined, the performance is improved. This paper extends the previous work to an arbitrary number of descriptions (packets) by using fractional oversampling and a new decoding principle. We demonstrate that, due to source aliasing, existing MSE optimized reconstruction rules from noisy sampled data, performs poorly from a perceptual point of view. A simple reconstruction rule is proposed, that improves the PEAQ objective difference grades (ODG) by more than 2 points. The proposed audio coder enables low-delay high-quality audio streaming on networks with late packet arrivals or packet losses. With a coding delay of 2.5 ms, and a total bitrate of 300 kbps, it is demonstrated that mean PEAQ ODGs around -0.65 can be obtained for 48 kHz (mono) music (pop & rock), and packet loss probabilities of 20%.

In this paper, we consider the rate-distortion problem where a source X is encoded into k parallel descriptions Y1, . . ., Yk, such that the error signals X - Yi, i = 1, . . ., k, are mutually independent given X. We show that if X is one-sided exponentially distributed, the optimal decoder (estimator) under the one-sided absolute error criterion, is simply given by the maximum of the outputs Y1, . . ., Yk. We provide a closed-form expression for the rate and distortion for any k number of parallel descriptions and for any coding rate. We furthermore show that as the coding rate per description becomes asymptotically small, encoding into k parallel descriptions and using the maximum output as the source estimate, is rate-distortion optimal.

We consider a fixed-rate zero-delay source coding problem where a stationary vector-valued Gauss-Markov source is compressed subject to an average mean-squared error (MSE) distortion constraint. We address the problem by considering the Gaussian nonanticipative rate distortion function (NRDF) which is a lower bound to the zero-delay Gaussian RDF. Then, we use its corresponding optimal “test-channel” to characterize the stationary Gaussian NRDF and evaluate the corresponding information rates. We show that the Gaussian NRDF can be achieved by p-parallel fixed-rate scalar uniform quantizers of finite support with dithering signal up to a multiplicative distortion factor and a constant rate penalty. We demonstrate our framework with a numerical example.

Ozarow's test channel for the quadratic Gaussian (QG) multiple description (MD) problem consists of two correlated AWGN channels. It is known that simply replacing the AWGN channels by quantizers with equivalent statistical properties as the channels, will generally not lead to a rate-distortion optimal realization of the MD rate-distortion function. We have previously proposed a symmetric two-channel model for the QG MD problem for the case, where the two noise terms have equal variances. We show in this paper, that by replacing the AWGN channels of this model by quantizers that are statistical equivalent to the channels, will under high-resolution assumption be rate-distortion optimal. We furthermore extend this symmetric two-channel model to the asymmetric case, and provide a simple suboptimal implementation of the channel based on scalar quantizers. Simulations are provided to show the performance of the proposed implementation.

Due to the distributed structure of many modern audio transmission setups, it is likely to have an observation at the receiver which is correlated with the desired source at the transmitter. This observation could be used as side information to reduce the transmission rate using distributed source coding. How to integrate distributed source coding into the perceptual audio compression procedure is thus a fundamental question. In this paper, we take a completely analytical approach to this problem, in particular to the rate-distortion trade-off and the corresponding coding schemes. We then interpret the results from an audio coding perspective. The main result is that, to upgrade a regular perceptual audio coder to a distributed coder, one needs to revise the perceptual masking curve. The revised masking curve models the availability of the side information as an extra masking effect, yielding lower rates. Interestingly, this means that at least conceptually, the distributed coding scenario could be integrated into the audio coder with minor changes, and without destructing the original coder.

We formulate a new problem which bridges between source coding and enhancement in wireless acoustic sensor networks. We consider a network of wireless microphones, each of which encoding its own measurement under a covariance matrix distortion constraint and sending it to a fusion center. To process the data at the center, we use a recent spatio-temporal prediction filter. We assume that a weighted sum-rate for the network is specified. The problem is to allocate optimal distortion matrices to the nodes in order to achieve a maximum output SNR at the fusion center after processing the received data, while the weighted sum-rate for the network is no more than the specified value. We formulate this problem as an optimization problem for which we derive a set of equalities imposed on the solution by studying the KKT conditions. In particular, for the special case of scalar sources with two microphones and a sum-rate constraint, we derive the distortion allocation in closed form and will show that if the given sum-rate is higher than a critical value, the stationary points from the KKT conditions lead to distortion allocations which maximize the output SNR of the filter.

In this paper, we consider the problem of remote vector Gaussian source coding for a wireless acoustic sensor network. Each node receives messages from multiple nodes in the network and decodes these messages using its own measurement of the sound field as side information. The node's measurement and the estimates of the source resulting from decoding the received messages are then jointly encoded and transmitted to a neighbouring node in the network. We show that for this distributed source coding scenario, one can encode a so-called conditional sufficient statistic of the sources instead of jointly encoding multiple sources. We focus on the case where node measurements are in form of noisy linearly mixed combinations of the sources and the acoustic channel mixing matrices are invertible. For this problem, we derive the rate-distortion function for vector Gaussian sources and under covariance distortion constraints.

In this paper, we consider robust source coding in closed-loop systems. In particular, we consider a (possibly) unstable LTI system, which is to be stabilized via a network. The network has random delays and erasures on the data-rate limited (digital) forward channel between the encoder (controller) and the decoder (plant). The feedback channel from the decoder to the encoder is assumed noiseless. Since the forward channel is digital, we need to employ quantization. We combine two techniques to enhance the reliability of the system. First, in order to guarantee that the system remains stable during packet dropouts and delays, we transmit quantized control vectors containing current control values for the decoder as well as future predicted control values. Second, we utilize multiple description coding based on forward error correction codes to further aid in the robustness towards packet erasures. In particular, we transmit M redundant packets, which are constructed such that when receiving any J packets, the current control signal as well as J-1 future control signals can be reliably reconstructed at the decoder. We prove stability subject to quantization constraints, random dropouts, and delays by showing that the system can be cast as a Markov jump linear system.

In this paper we propose a novel spatial error concealment algorithm for video and images based on convex optimization. Block-based coding schemes in packet loss environment are considered. Missing macro blocks are sequentially reconstructed by filling them with a weighted set of templates extracted from the available neighbourhood. Moreover, a fast approximation of the optimization method is proposed. The technique produces high quality reconstructions that outperforms the state-of-the-art algorithms both in terms of PSNR and MS-SSIM.

We consider linear predictive coding and noise shaping for coding and transmission of auto-regressive (AR) sources over lossy networks. We generalize an existing framework to arbitrary filter orders and propose use of fixed-lag smoothing at the decoder, in order to further reduce the impact of transmission failures. We show that fixed-lag smoothing up to a certain delay can be obtained without additional computational complexity by exploiting the state-space structure. We prove that the proposed smoothing strategy strictly improves performance under quite general conditions. Finally, we provide simulations on AR sources, and channels with correlated losses, and show that substantial improvements are possible.

The problem of the rate region of the vector Gaussian multiple description with individual and central quadratic distortion constraints is studied. We have two main contributions. First, a lower bound on the rate region is derived. The bound is obtained by lower-bounding a weighted sum rate for each supporting hyperplane of the rate region. Second, the rate region for the scenario of the scalar Gaussian source is fully characterized by showing that the lower bound is tight. The optimal weighted sum rate for each supporting hyperplane is obtained by solving a single maximization problem. This is contrary to existing results, which require solving a min-max optimization problem.

We consider lossy compression of image sequences using l1-compression with overcomplete dictionaries. As a fidelity measure for the reconstruction quality, we incorporate the recently proposed structural similarity index measure, and we show that this leads to problem formulations that are very similar to conventional l1 compression algorithms. In addition, we develop efficient large-scale algorithms used for joint encoding of multiple image frames.

We characterize the rate-distortion function for zero-mean stationary Gaussian sources under the MSE fidelity criterion and subject to the additional constraint that the distortion is uncorrelated to the input. The solution is given by two equations coupled through a single scalar parameter. This has a structure similar to the well known water-filling solution obtained without the uncorrelated distortion restriction. Our results fully characterize the unique statistics of the optimal distortion. We also show that, for all positive distortions, the minimum achievable rate subject to the uncorrelation constraint is strictly larger than that given by the un-constrained rate-distortion function. This gap increases with the distortion and tends to infinity and zero, respectively, as the distortion tends to zero and infinity.

It was recently shown that the symmetric multiple-description (MD) quadratic rate-distortion function for memoryless Gaussian sources and two descriptions can be achieved by dithered Delta-Sigma quantization combined with memoryless entropy coding. In this paper,we generalize this result to stationary (colored) Gaussian sources by combining noise shaping and source prediction. We first propose a new representation for the test channel that realizes the MD rate-distortion function of a Gaussian source, both in the white and in the colored source case. We then show that this test channel canbe materialized by embedding two source prediction loops, one for each description, within a common noise shaping loop. While the noise shaping loop controls the tradeoff between the side and the central distortions, the role of prediction (like in differential pulse code modulation) is to extract the source innovations from the reconstruction at each of the side decoders, and thus reduce the coding rate. Finally, we show that this scheme achieves the MD rate-distortion function at all resolutions and all side-to-central distortion ratios, in the limit of high dimensional quantization.

In this paper we address the connection between the multiple-description (MD) problem and delta-sigma quantization. Specifically, we exploit the inherent redundancy due to oversampling in delta-sigma quantization, and the simple linear-additive noise model resulting from dithered lattice quantization, in order to construct a symmetric MD coding scheme. We show that the use of feedback by means of a noise shaping filter makes it possible to trade off central distortion for side distortion. Asymptotically as the dimension of the lattice vector quantizer and order of the noise shaping filter approach infinity, we show that the symmetric two-channel MD rate-distortion function for the memoryless Gaussian source and MSE fidelity criterion can be achieved at any resolution. This realization provides a new interesting interpretation for the information theoretic solution. The proposed design is symmetric in rate by construction and there is therefore no need for source splitting

We derive analytical expressions for the central and side quantizers in an n-channel symmetric multiple-description lattice vector quantizer which, under high-resolution assumptions, minimize the expected distortion subject to entropy constraints on the side descriptions for given packet-loss probabilities. The performance of the central quantizer is lattice dependent whereas the performance of the side quantizers is lattice independent. In fact the normalized second moments of the side quantizers are given by that of an L-dimensional sphere. Furthermore, our analytical results reveal a simple way to determine the optimum number of descriptions. We verify theoretical results with numerical experiments and show that with a packet-loss probability of 5%, a gain of 9.1 dB in MSE over state-of-the-art two-description systems can be achieved when quantizing a two-dimensional unit-variance Gaussian source using a total bit budget of 15 bits/dimension and using three descriptions. With 20% packet loss, a similar experiment reveals an MSE reduction of 10.6 dB when using four descriptions.