Soft-output detector for partial response channels using vector quantization

The Internet community is very heterogeneous in terms of access bandwidth and terminal capabilities, hence, there is much interest for low-computation, software-only, scalable video coders that guarantee universal access to video communication. Scalability allows users to achieve a fair quality of service in relation to their resources. Low complexity, on the other hand, is necessary in order to ensure that also users with low computing power can be served.In this work, we propose a multiplication-free video codec, whose complexity is much reduced with respect to standard coders at the price of a limited increase in memory requirements. To this end we resort to very simple coding tools such as table lookup vector quantization (VQ) and conditional replenishment. We start from the simple coder proposed in [1], which already guarantees high scalability and limited computational burden, and improve upon it by further reducing complexity, as well as the encoding rate, with no effect on the encoding quality. The main innovation is the use of ordered VQ codebooks, which allows the encoder to generate correlated indexes, unlike in conventional VQ. Index correlation, in turn, allows us to carry out conditional replenishment (the most time-consuming operation in the original coder) by working on indexes rather than on block of pixels, and to reduce drastically its complexity. In addition, we also take advantage of the correlation among indexes to compress them by means of a predictive scheme, which leads to a 15-20% rate reduction in the base layer, without significant increase in complexity. Thanks to these and other minor optimizations we have obtained improved performance and, more important, a 60-70% reduction of the encoding time (on a general purpose machine) with respect to [1].

This paper is concerned with the applications of majority-logic decodable nonbinary low-density parity-check (LDPC) codes to partial response (PR) channels. We propose a joint detection/decoding algorithm that works in an iterative manner by exchanging messages between the Viterbi detector and a generalized majority logic decoder (GMLGD). The Viterbi detector is implemented over a sectionalized trellis. The hard-decisions made by the Viterbi detector are then passed to the decoder. The decoder delivers as output estimates of each coded symbols, which will be utilized in next iteration to update the branch metrics of the trellis. Since the proposed algorithm requires only integer operations and finite field operations, it can be implemented with simple combinational logic circuits. Simulation results and complexity analysis show that, when compared with the conventional turbo equalizer implemented with the BCJR algorithm and the Q-ary sum-product algorithm (BCJR-QSPA), the proposed algorithm has a much lower complexity but suffers from a little performance degradation. So the proposed algorithm provides a good candidate for trade-offs between performance and complexity.

The adaptive multi-rate wideband (AMR-WB) speech codec is widely used in modern mobile communication systems for high speech quality in handheld devices. Nonetheless, a major disadvantage is that vector quantization (VQ) of immittance spectral frequency (ISF) coefficients takes a considerable computational load in the AMR-WB coding. Accordingly, a binary search space-structured VQ (BSS-VQ) algorithm is adopted to efficiently reduce the complexity of ISF quantization in AMR-WB. This search algorithm is done through a fast locating technique combined with lookup tables, such that an input vector is efficiently assigned to a subspace where relatively few codeword searches are required to be executed. In terms of overall search performance, this work is experimentally validated as a superior search algorithm relative to a multiple triangular inequality elimination (MTIE), a TIE with dynamic and intersection mechanisms (DI-TIE), and an equal-average equal-variance equal-norm nearest neighbor search (EEENNS) approach. With a full search algorithm as a benchmark for overall search load comparison, this work provides an 87% search load reduction at a threshold of quantization accuracy of 0.96, a figure far beyond 55% in the MTIE, 76% in the EEENNS approach, and 83% in the DI-TIE approach.

This paper will describe the vector quantization (VQ) of speech spectrum parameters as used in state of the art real time speech coding algorithms. The complexity of these speech coding algorithms, combined with the constraints of real time operation, limit the practical size of the vector quantization tables. This conflicts with the requirement that large VQ tables are needed to provide high quality coded speech. The goal of this paper is to demonstrate the feasibility of using tree searched vector quantizers to overcome this conflict.

On the transformed entropy-constrained vector quantizers employing Mandala block for image coding

Adaptive predistortion is one of the most promising methods to linearize RF power amplifiers. The size of gain look-up table is a performance limiting factor and so are various error sources existing in the system. In this paper, we study the table size effects in combination with system imperfections. It is shown that errors introduced in the adaptation path can cause an inverse relationship between the table size and the output adjacent channel power ratio. Thus, this implies that there exists an optimal table size in practice and a further increase of the table size beyond that value can even result in linearity degradation instead of improvement. The reason behind this unexpected result is discussed. This study indicates that the table size shall be properly chosen based on a good understanding of various error sources located in the system.

We study the performance of turbo codes for magnetic recording channels using realistic channel models. The main objective is to provide a benchmark for the future studies on the subject. We examine a variety of schemes including the use of parallel concatenated turbo codes with and without turbo equalization and convolutional codes with an interleaver (can be regarded as a serial concatenation of convolutional codes with the partial response channel). We investigate the effect of the (turbo) code block length, the use of different code structures, and the effects of precoding and the media noise on the system performance. We assume that the partial response target is EPR4.

A major task in processing scanned images is converting the color values from the non-colorimetric color space used by the scanner to a device-independent color space or the color space of an output device. For example, typical color scanners produce 24-bit RGB values for each color pixel, while most printers use color spaces such as CMY or CMYK. While this conversion can be accomplished using a lookup table, the 16 million entries required makes the direct lookup table prohibitively large. This suggests use of a sparse lookup table, where only certain values are directly mapped. For all color values that are not directly mapped, the unknown values may be calculated from the nearest table entries via trilinear interpolation, a computationally intensive process. This paper presents a method of efficiently converting from a set of noisy color values to a set of device colorants. Using a deterministic process, 24-bit scanned color values are reduced to dithered 12-bit RGB table indices. After the reduction, a small but complete lookup table with 4096 entries converts the RGB values directly to the output color space. This stochastic interpolation process, while minimizing banding and abrupt color transitions, eliminates the need for trilinear interpolation of the data and significantly reduces the size of the lookup table.

We propose a video coding and delivery scheme which is geared towards low bit-rate and real-time performance requirements. We use a finite state wavelet-based hierarchical lookup vector quantization (FSWHVQ) scheme, which embeds the optical flow calculations in table-lookups. This video coding scheme is both fast (table-lookups) and accurate (dense motion field), and avoids the blocking artifacts and poor prediction which plagues block coding schemes at low bit rates. For restricted image compression/transmission scenarios like teleconferencing, for which a good training set may be available, the FSWHVQ scheme may be viewed as storing an internal representation in its lookup tables, a valid and complete model of the problem domain.

We propose novel techniques to reduce the complexity of the well-known Bahl, Cocke, Jelinek, and Raviv (BCJR) algorithm when it is employed as a detection algorithm in turbo equalization schemes. In particular, by also considering an alternative formulation of the BCJR algorithm, which is more suitable than the original one for deriving reduced-complexity techniques, we describe three reduced-complexity algorithms, each of them particularly effective over one of the three different classes of channels affected by intersymbol interference (minimum-phase, maximum-phase, and mixed-phase channels). The proposed algorithms do not explore all paths on the trellis describing the channel memory, but they work only on the most promising ones, chosen according to the maximum a posteriori criterion. Moreover, some optimization techniques improving the effectiveness of the proposed solutions are described. Finally, we report the results of computer simulations showing the impressive performance of the proposed algorithms, and we compare them with other solutions in the literature.

The performance of underwater acoustic communication systems is sensitive to inter-symbol interference caused by delay spread developed from multipath signal propagation. The multipath nature of underwater channels causes signal distortion and error floor. In order to improve the performance, it is necessary to employ an iterative coding scheme. Of the various iterative coding schemes, turbo code and convolutional code based on the BCJR algorithm have recently dominated this application. In this study, the performance of iterative codes based on turbo equalizers with equivalent coding rates and similar code word lengths were analyzed. Underwater acoustic communication system experiments using these two coding techniques were conducted on Kyeong-chun Lake in Munkyeong City. The distance between the transmitter and receiver was 400 m, and the data transfer rate was 1 Kbps. The experimental results revealed that the performance of turbo codes is better for channeling than that of convolutional codes that use a BCJR decoding algorithm.

We propose a real-time image coding system, which is capable of adapting instantaneously to the available channel bandwidth. The range of operational bandwidths for this system has a finer calibration than ordinary hierarchical vector quantization (HVQ) or wavelet based hierarchical vector quantization (WHVQ) methods suggested in the literature. These properties make it very attractive for networks with fluctuating available bandwidth, like the Internet. All encoder and decoder operations are strictly constant time per pixel, proceeding through table lookups, and are intrinsically suitable for parallel and hardware implementation.

In the C-band high-speed long-reach optical intensity modulation direct detection (IM/DD) communication system, severe signal distortion, including chromatic dispersion and device bandwidth limitations, making the size of lookup tables (LUTs) required by equalizers too large to be operated. In this paper, we propose a cluster-assisting lookup table (CLUT) approach to reduce the table size significantly. In addition, we also propose multiplication-free CLUT-based decision-feedback equalizer (CLUT-DFE) and CLUT-based log-maximum a posteriori estimation with a fixed number of surviving states (CLUT-FS-MAP) to effectively compensate for channel distortions. Using the k-means clustering algorithm, the proposed CLUT significantly shortens the table size by reducing the number of independent variable values input. The table size of traditional LUTs increases exponentially with the filter tap number, making it impossible for LUTs to replace filters with tap numbers greater than 20. The CLUT breaks this limitation. The proposed schemes are experimentally performed in a C-band 61-Gb/s IM/DD optical on-off keying (OOK) system featuring a 3.65-GHz 3-dB bandwidth over a 100-km standard single mode fiber (SSMF). Compared to traditional LUT, CLUT reduces table size by about 10 orders of magnitude with the same performance. To the best of our knowledge, this study represents the first deployment of ultrasmall-size CLUT in equalization.

Turbo codes are applied to magnetic recoding channels by treating the channel as a rate-one convolutional code that requires a soft a posteriori probability (APP) detector for channel inputs. The complexity of conventional APP detectors, such as the BCJR algorithm or the soft-output Viterbi algorithm (SOVA), grows exponentially with the channel memory length. This paper derives a new APP module for binary intersymbol interference (ISI) channels based on minimum mean squared error (MMSE) decision-aided equalization (DAE), whose complexity grows linearly with the channel memory length, and it shows that the MMSE DAE is also optimal by the maximum a posteriori probability (MAP) criterion. The performance of the DAE is analyzed, and an implementable turbo-DAE structure is proposed. The reduction of channel APP detection complexity reaches 95% for a five-tap ISI channel when the DAE is applied. Simulations performed on partial response channels show close to optimum performance for this turbo-DAE structure. Error propagation of the DAE is also studied, and two fixed-delay solutions are proposed based on combining the DAE with the BCJR algorithm.

We propose a new iterative timing recovery scheme based on per-survivor processing that jointly performs timing recovery and turbo equalization on partial response channels with error-correction codes. The scheme embeds the timing recovery process inside the Bahl, Cocke, Jelinek, and Raviv (BCJR) equalizer, using per-survivor processing. This per-survivor BCJR equalizer then iteratively exchanges soft information with an error-correction decoder. Results indicate that the proposed scheme yields a better performance than a conventional receiver that performs timing recovery and turbo equalization separately, especially when the channel encounters severe timing jitter noise. We also present evidence that suggests that the proposed scheme can correct a cycle slip much more efficiently than the others.