Abstract: An echo canceler has its otherwise conventional tapped delay line basis functions augmented with at least one pair of basis functions realizing two quadrature responses of a complex pole. The frequency of this complex pole is selected as substantially equal to the frequency of the dominant pole of a highpass filter inserted into the echo path of the echo canceler, where the dominant pole is the pole closest to the unit circle. This echo canceler arrangement results in a substantial improvement in attainable echo-return-loss enhancement (ERLE).

Abstract: A network-based CODEC (coder-decoder) includes an echo canceler. This CODEC recognizes the presence of a data call by detecting predefined signaling portions of a modem handshaking process. For each detected data call, the CODEC uses a stored channel model for performing echo cancellation during the data call. The CODEC trains off-line during selected segments of the modem call and then stores the new channel model for use in a future data call.

Abstract: A network-based CODEC (coder-decoder) includes an echo canceler. This CODEC recognizes the presence of a data call by detecting predefined signaling portions of a modem handshaking process. For each detected data call, the CODEC uses a stored channel model for performing echo cancellation during the data call. The CODEC trains off-line during selected segments of the modem call and then stores the new channel model for use in a future data call.

Abstract: Quantization noise, introduced into data transmission when analog signals are translated into PCM code using the logarithmic conversion of &mgr;-law or A-law rules necessary to accommodate the transmission of voice signals, limits the maximum attainable speed of data transmission. However, when the PCM code signals represent data (rather than voice signals), linear conversion of analog data signals into the PCM code would avoid such logarithmic quantization noise. To signal the translating codec that a digital modem call is being made, the digital modem allocates one or more of the least significant bit positions of the code representing the 2100 Hz answer-tone-with-phase-reversal (specified in ITU-T Recommendation G. 165) to send a repetitive pattern “P1” to signal to the associated codec that a digital modem connection has been made.

Abstract: An adaptive filter and method that proportionately adjusts individual tap gain distributors such that the individually adjusted tap gains are not necessarily equal to one another and that the average of tap gains remains substantially constant. The filter employs a proportionate normalized least-means-squares (PNLMS) method of adaptation that imparts improved convergence speed over prior art adaptive filters that utilize a normalized least-means-squares (NLMS) method of adaptation, without affecting the adaptation quality of the filter.

Abstract: A system and method for enhancing interactive communication between video conferencing devices of the type in which a delay is inserted into the audio transmission path to provide lip synchronization of the image and speech of the respective users thereof. Each video conferencing device includes a display device for displaying images of at least one communicating party and a speech communicating system for communicating with the communicating party. In accordance with one embodiment of the invention, a speech detecting circuit detects an utterance by a first user of a first video conferencing apparatus. An audible or visual indication is provided to at least a second user of a second video conferencing apparatus before the utterance is reproduced. As a result, the potential for simultaneous speaking by two or more users is substantially reduced.

Abstract: The detection and cancellation of echo in a communications system is enhanced by taking advantage of the accumulated knowledge relating to the echo path delay that an adaptive filter acquires as a result of its adaptation process. Such knowledge improves significantly the ability of an associated Near-End-Speech (NES) detector to detect for the presence of near end speech.

Abstract: The detection of echo is enhanced using a coarse near-end speech detector at a Y-side of an echo canceler and using a more accurate near-end speech detector at the E-side of the echo canceler, such that the latter detector uses a representation of far-end speech and a threshold value to determine if the magnitude of the echo canceler output is larger than that of an echo signal only. If it is, and there is an absence of an indication of the presence of near-end speech from the coarse detector, then the more accurate speech detector adjusts the value of the threshold in a first direction, otherwise it adjusts the threshold in a second direction, to more "finely tone" such accuracy.

Abstract: In a digital communication system, synchronization of a plurality of asynchronous digital signals is usually achieved by stuffing one or more pulses in a pulse stream. Thereby, stuffing equalizes the input signal bit rate with the usually fixed bit rate of a signal on a transmission link. For example, in pulse stuffing, single pulses are added at certain allowed times to equalize the input bit rate with the transmission link bit rate. On the other hand, in block stuffing, a block of pulses is added. However, conventional block stuffing usually requires a narrowband phase-locked loop for removing jitter in the received signal. As a result, slower acquisition of synchronization occurs.

Abstract: An analog-to-digital counting encoder or a digital-to-analog counting decoder for a pulse code modulation signal typically requires a function generator using precision components to generate a piecewise linear comparison signal corresponding to a segment companding law. As a result, nonproportional component change introduces nonlinear distortion in the encoded or decoded signal. To mitigate nonlinear distortion and to provide a less expensive digital signal processing arrangement, the invention resides in an improved nonprecision component function generator. The function generator includes two serially connected integrators which are advantageously switched in a complementary fashion to provide the comparison signal.

Abstract: A digital processor apparatus is described which operates upon nonuniformly encoded digital words each one of which has .mu.-law encoded format of a sign bit to indicate polarity, segment bits to indicate range, and interval bits to indicate the location of the sample within the segment. A predetermined number of the digital words is stored in shift registers while the apparatus determines the maximum segment value represented in that block of predetermined number of words. The interval bits for one of the stored digital words are stored in a shift register along with a bit whose value is determined from the segment bits for that word. This shift register is then shifted by an amount which is dependent on the difference between the segment value for the digital word under consideration and the maximum segment value for that block of digital words.