An efficient soft-input soft-output iterative decoding algorithm for block turbo codes (BTCs) is proposed. The proposed algorithm utilizes Kaneko's (1994) decoding algorithm for soft-input hard-output decoding. These hard outputs are converted to soft-decisions using reliability calculations. Three different schemes for reliability calculations incorporating different levels of approximation are suggested. The algorithm proposed here presents a major advantage over existing decoding algorithms for BTCs by providing ample flexibility in terms of performance-complexity tradeoff. This makes the algorithm well suited for wireless multimedia applications. The algorithm can be used for optimal as well as suboptimal decoding. The suboptimal versions of the algorithm can be developed by changing a single parameter (the number of error patterns to be generated). For any performance, the computational complexity of the proposed algorithm is less than the computational complexity of similar existing algorithms. Simulation results for the decoding algorithm for different two-dimensional BTCs over an additive white Gaussian noise channel are shown. A performance comparison of the proposed algorithm with similar existing algorithms is also presented.

The concept of a general nonlinear parity-check equation was first proposed by Pietrobon et al. who used them to construct rotationally invariant trellis codes for two dimensional signal sets. This paper applies this technique to trellis-coded modulation (TCM) codes with multidimensional MPSK and MQAM constellations to construct nonlinear phase rotationally invariant multi-D codes. Based on this construction technique, a series of the best fully phase rotationally invariant multi-D trellis codes for a variety of signal constellations is obtained through a systematic code search program. The code search results show that most codes achieve the same asymptotic coding gain as the best linear, but nonphase rotationally invariant codes found by Pietrobon et al.

A high speed 800 channel digital interpolator network has been designed. The interpolator network accepts data from the demultiplexer onboard a satellite which separates 800 telephone channels from ground stations in an FDMA format into individual channels in a TDM format at 45 ksamples/sec/ch. The interpolator then changes the sampling rate of the data input to 64 ksamples/sec/ch so that the demodulator onboard the satellite can correctly demodulate all of the 800 channels with a requirement of two samples/symbol data rate. An interpolator network has been implemented using 0.8 micron CMOS technology and is suitable for real time processing.

An 8PSK scheme with the least significant bit (LSB) Golay coded is proposed and its performance is analyzed. With hard decision decoding, the proposed scheme can achieve 1 dB asymptotic coding gain and also save 26% bandwidth, as compared with uncoded QPSK scheme.>

The quality of visual communications has greatly improved with HDTV. Due to the advantages of digital communications, and the trend towards it, it is desirable to develop a cost effective digital codec for HDTV signals. One of the criteria in the transmission of digital signals via satellites is the transmission bandwidth which dictates the transmission costs. Apart from development of good compression schemes, due consideration should also be given to the ease of hardware implementation, which is crucial in deciding factors like processing delay and the level of encoder/decoder complexity. With the above mentioned factors in mind a cost-effective HDTV codec is proposed. Few US organizations have addressed the problem of HDTV distribution via communication satellites. The majority of efforts in the development of an HDTV codec have been aimed at the 100 to 140 Mbits/sec data rate. To facilitate cost effective satellite transmission, the DS3 rate of 45 Mbits/sec, is desirable. The purpose of the present research is to compress the HDTV signals to around 20 Mbits/sec so that 2 HDTV channels can be transmitted through the DS3 rate channel. To achieve this degree of compression a new scheme is proposed. First subband coding is used to decompose the image into different frequency bands. Next motion compensation is applied to the low band, and the resulting interframe difference is discrete cosine transform (DCT) coded. The high bands are also DCT coded. The DCT coefficients of the difference signal of the low band and the DCT coefficients of the high band are then vector quantized.>

A generalized covariance matrix computation processor has been designed. The covariance matrix is the first processing step for the computation of the direction of arrival (DOA) algorithms. This covariance processor is useful for computing the covariance matrix for various DOA algorithms. In this work three DOA algorithms namely MUSIC algorithm for narrowband signal and BASS-ALE & bilinear transformation algorithms for wideband signals have been considered. The processor has been implemented using 0.8 micron CMOS technology and is suitable for real time processing.>

The transmission of high-definition television (HDTV) signals on available digital networks and satellites requires the adoption of sophisticated compression techniques to limit the bit rate requirements and to provide high-quality and reliable service to customers. For processing and transmission of image signals, a low-complexity codec without visible degradation is desired. A low-complexity intraframe subband image coding algorithm is developed. The low band is DPCM encoded and the high bands are PCM encoded. An efficient entropy coder is designed which reduces the overall bit rate significantly. It is shown that high-quality HDTV images can be obtained at as low a bit rate as 45 Mb/s or less with a very low-complexity encoder. For dividing the image into subbands, a new class of quadrature mirror filters (QMFs) called generalized quadrature mirror filters (GQMFs) is used for filtering. Performance is also evaluated by using short kernel filters (SKFs), which are easy to implement and require very few computations.>

A reconfigurable transmultiplexer that is capable of on-board demultiplexing of a varying number of single channel per carrier frequency division multiple access (FDMA) channels with varying bit rates is presented. The multiplexing algorithm selected for demultiplexing the FDMA channels is the polyphase FFT (fast Fourier transform) method, which requires a bank of filters followed by an FFT operation. A reconfigurable shared filter bank and reconfigurable pipelined FFT architecture are designed to implement the bank of filters and FFT operations for two different cases. The architecture is suitable for satellite on-board processing as it is reconfigurable and modular and can perform its processing in real time without large buffers. The architecture is illustrated specifically for demultiplexing 800 channels, at 64 kbps or a mix of 400 channels at 64 kbps and 12 channels at 2.048 Mbps or 24 channels at 2.048 Mbps.>

A real-time vector quantizer (VQ) architecture for broadcast quality encoding of color TV images is presented. The architecture maps the mean/quantized residual vector quantizer (MQRVQ), an extension of mean/residual VQ, onto a VLSI/LSI chip set. The MQRVQ contributes to the feasibility of the VLSI architecture through the use of a simple multiplication-free distortion measure and reduction of the required memory per codevector. In other words, the complexity-reduced MQRVQ allows each subcodebook of 128 codevectors (10 kb) and its associated encoding elements to be fitted onto a VLSI chip. This architecture reduces encoding search complexity through partitioning of a large codebook into on-chip memories of a concurrent VLSI chip set. There are 64 of these VLSI chips to accommodate a codebook with size up to 2/sup 13/=8192 codevectors. The architecture is capable of real-time processing of 480*768 pixels per frame with refreshing rate of 30 frames/second.>

The authors investigate image sequence coding by means of a classified vector quantization technique in which 3-D cubes of the image sequence are coded. A four-image-color sequence (512*768*4) is divided into 32*32*4 or smaller cubes, which are classified as to horizontal, vertical, and temporal activity. Once the final form of the cube has been determined, a 3-D discrete cosine transform is applied, and the resulting coefficients are zone-sampled. These coefficients are then vector-quantized, and the resultant codebooks are optimized by both the LBG (Linde-Buzo-Gray) method and a simulated annealing technique. The average bit rate is in the range of 0.34-0.45 bits/pixel for active sequences with signal/noise ratio values (peak signal to root-mean-square noise) in the range of 36 to 41 dB overall.>

This paper presents an adaptive algorithm for motion compensated color image coding. The algorithm can be used for video teleconferencing or broadcast signals. Activity segmentation is used to reduce the bit rate and a variable stage search is conducted to save computations. The adaptive algorithm is compared with the nonadaptive algorithm and it is shown that with approximately 60 percent savings in computing the motion vector and 33 percent additional compression, the performance of the adaptive algorithm is similar to the nonadaptive algorithm. The adaptive algorithm results also show improvement of up to 1 bit/pel over interframe DPCM coding with nonuniform quantization. The test pictures used for this study were recorded directly from broadcast video in color.

In motion compensated coding of the NTSC composite video signal using block matching method, the motion vector can be determined either by using the composite signal itself or by using only the luminance component of the composite signal. However the computational requirement and the compression obtained may be quite different in the two cases. This paper provides this comparison using an adaptive variable stage motion search motion compensated coding scheme. It is found that for approximately the same signal to noise power ratio, composite coding requires more computations but also results in higher data compression as compared to coding the composite signal based on luminance component only.

A new technique for separating the composite NTSC television signal into its components is presented in this paper. An allpass filter is used to separate the luminance and chrominance components. The chrominance is then separated into its I and Q components using an algebraic technique rather than the conventional demodulation and filtering technique.

In orthogonal transform coding of NTSC video signal, the data are processed in blocks. Processing a block of data that is maximally correlated leads to larger compression. This correspondence describes a new architecture for arranging the pixels in a block that yields higher correlation than the adjacent pixel blocks. The data are processed using adaptive, intrafield, discrete cosine transform, variable bit allocation coding. In subjective tests good quality pictures were obtained for an average of about 2 bits/pel for intrafield processing.

The ill-posed nature of direct deconvolution has prevented its use in applications such as channel equalization or inversion of transducer distortion. In this paper we discuss the use of a recursive nonlinear filter. The design of the proposed filter is accomplished in two stages. First, a constrained least-squares technique is employed to obtain, off-line, a filtered signal from a typical distorted signal. Then this pair of signals is used in a model identification scheme, to arrive at the nonlinear recursive filter. Since the final filter is recursive, and synthesized in one of the two alternative structures, both the memory and computation requirements are low enough to permit efficient on-line implementation.