A low-bit-rate image codec for videophone and B/2B-ISDN (integrated services digital network) applications is presented. The encoder skips two frames over three images and encodes one image over three, using DCT (discrete cosine transform)-Huffman coding, and sends this information together with the motion vectors computed on overlapped blocks. At the receiver the missing frames are reconstructed by linear interpolation using coded frames and corresponding motion vectors. The resulting video sequences indicate that for a videophone in ISDN network applications, an acceptable quality image is obtained.>

A stability condition analysis is performed for weighted Least Squares Algorithms. It is shown that the introduction of a constant in the prediction error energy updating recursion can make the algorithm stable. Coefficient wordlength estimations can then be derived from those available for Least Mean Squares Algorithms. The results given are useful in designing adaptive filters based on least squares techniques in their simplest forms.

A simple exponential technique is presented for the real time power estimation of non stationary signals. It is used to adjust the coefficient adaptation step size of a LMS filter. Its efficiency is demonstrated with an adaptive linear predictor for speech. Improved performance with respect to conventional techniques appears. However the main advantage of that exponential estimation is its simplicity, since it can be implemented with just one, or even no full multiplications or divisions.

Leakage terms are introduced in certain adaptive filters to improve stability and robustness to transmission errors. The effects of that modification to the conventional procedure is analyzed for the gradient technique. Estimations are given for the resulting bias of the coefficients and error signal power increase. As an illustration, a prediction filter in an ADPCM codec is considered.

The performance results of the adaptive DPCM speech codecs which have been used in conjunction with an Adaptive Frequency Mapping system (AFMAP) are presented. The AFMAP preprocessor operates on a wideband speech (0.3 - 7.6 KHz) and compresses the speech signal into 0.3 - 3.3 KHz telephone channel. This signal is subsequently digitized by DPCM codecs employing simple, but efficient prediction algorithms in order to reproduce wideband speech from AFMAP postprocessor output. The informal listening tests, SNRSEG measures indicate that AFMAP system in tandem with DPCM codecs significantly enhances the recovered speech compared to other bandwidth compression systems. The reproduced AFMAP speech is also preferable to 0.3 - 3.3 KHz telephone speech, digitized by the some codecs at the same transmission bit rates.

The field of applications of adaptive digital filters is growing and, as for fixed coefficient filters, system designers and project engineers need simple means for computational complexity assessment. An attempt is made in the present paper to derive simplified expressions for coefficient and internal data wordlengths as a function of specified parameters. Only the gradient technique in direct and cascade forms is considered.

A comparative study of DPCM-AQF speech digitizers employing various second order prediction algorithms is presented. A new prediction scheme called Correlation Switched Prediction is also discussed. Computer simulation results for the range of 16 to 32 kbits/sec, indicate an improved segmented signal-to-noise ratio (SNRSEG) performance when using Forward Block Adaptive (FBA) instead of Backward Sequentially Adaptive Prediction. The introduction, in DPCM-AQF, of the relatively simple Correlation Switched Predictor produces SNRSEG values comparable to those obtained from DPCM-AQF-FBA. A more complex prediction scheme which combines Correlation Switched and Backward Sequential Adaptive Prediction (CSP-SGEP) is also considered. DPCM-AQF using the latter prediction is shown to provide the best overall SNRSEG performance.

The sequential gradient estimation predictor is compared in detail to the stochastic approximation predictor, and both are evaluated in an ADPCM codec. A switched predictor having two coefficients is then described for use in a DPCM-AQF codec. This predictor divides the range of the correlation coefficient of the speech signal into zones, and as the correlation coefficient changes zones, the predictor coefficients undergo a substantial modification. By this method the adaptation rate of the predictor is improved, particularly during transitions between unvoiced and voiced sounds.

A sequential gradient estimation predictor [SGEP] for speech signals is presented. In a given sampling interval, each prediction coefficient in turn is increased and decreased in value by a prescribed amount, while the other coefficients are kept constant. Two predictions are then made and the better enables the coefficient to be modified in the correct direction, but by an amount determined by a number of factors. The superiority of SGEP over the stochastic approximation predictor is illustrated by waveforms and snr (signal to noise ratio) performance curves.