Abstract: The convergence time of a blind, adaptive equalizer is shortened by using a tracking generator. The tracking generator comprises a smoothing filter which receives and smoothes a tap coefficient error estimate derived from an output data stream. Thereafter, a fraction of the smoothed estimate is generated. It is the use of this function of the smoothes estimate which allows the convergence time to be shortened.

Abstract: The convergence time of a blind, adaptive equalizer is shortened by using a tracking generator. The tracking generator comprises a smoothing filter which receives and smoothes a tap coefficient error estimate derived from an output data stream. Thereafter, a fraction of the smoothed estimate is generated. It is the use of this function of the smoothes estimate which allows the convergence time to be shortened.

Abstract: A method of converting between a sampling rate associated with a first audio format and a second audio format includes up-sampling an input signal sampled at the sample rate associated with the first audio format. Then, the up-sampled signal is filtered as a function of a fractional delay to generate an output signal sampled at the sample rate associated with the second audio format. The fractional delay is computed from the sample rates associated with the first and second audio formats. In one embodiment, the sample rates that are converted between are associated with a compact disc format having a sample rate of about 44.1 kHz and a digital audio tape format having a sample rate of about 48 kHz. In such case, the input samples are preferably up-sampled by a factor of two and the samples are then preferably filtered in accordance with a third order six taps coefficient finite impulse response filtering technique.

Abstract: The convergence time of a blind, adaptive equalizer is shortened by using a tracking generator. The tracking generator comprises a smoothing filter which receives and smoothes a tap coefficient error estimate derived from an output data stream. Thereafter, a fraction of the smoothed estimate is generated. It is the use of this function of the smoothes estimate which allows the convergence time to be shortened.

Abstract: A method and apparatus are disclosed for numerically controlling a ring oscillator. The disclosed programmable period ring oscillator selectively switches pairs of inverters into or out of the ring oscillator to provide a desired frequency. In one implementation, a programmable period ring oscillator provides a range of five to nine inverters that may selectively be included in the ring oscillator in increments of two inverters. A frequency synthesizer is disclosed that aligns the phase of the programmable ring oscillator with a reference signal. The frequency synthesizer generates a phase difference signal that is that is representative of the phase difference between the reference signal and the ring oscillator output. The phase difference signal is utilized to correct the frequency of the ring oscillator, so that the mean phase of the ring oscillator corresponds to the mean phase of the reference signal.

Abstract: A QAM data signal timing recovery loop feedback element provides a fixed sampling time offset adjustment to two continuously variable digital rate interpolators/decimators to produce a quadrature output stream at a programmed rational rate multiple of the actual baud rate of the received data signal. The continuously variable digital rate interpolators/decimators are configured at startup so as to produce output streams at the same programmed rational rate multiple of the nominal baud rate of the anticipated received data signal, assuming the fs sample timing offset adjustment stream provided by the timing recovery feedback element to be identically 0. The “nominal” fixed sampling rate fs of the received analog input signal need not be rationally related to the nominal baud rate of the anticipated received data signal.

Abstract: A method of efficiently demodulating and isolating a signal within a fixed possibly wider band spectral region, having any of a wide range of baud rates. The signal is sampled at a fixed, first frequency which remains fixed no matter what the baud rate. Thus, the sample rate does not necessarily correspond to the desired baud rate. The sampled signal is then demodulated and low pass filtered to create baseband samples at the first frequency, which are then subject to user-specified arbitrary rate change in a continuously variable interpolator/decimator (continuously variable digital delay (CVDD) device), and decimated by a programmable power of 2, to produce samples at a second frequency. The second frequency is preferably determined to be a whole number multiple of the desired baud rate, e.g., twice the desired baud rate. The samples are equalized to produce output symbols at the target baud rate.

Abstract: A digital carrier recovery system includes at least two modes of operation, namely, an acquisition mode and a tracking mode. The bandwidth of the carrier recovery loop filter is different for the acquisition mode and the tracking mode. In the acquisition mode, the digital phase-locked loop seeks and locks to the long term frequency offset of the received carrier signal. In the tracking mode, the digital phase-locked loop tracks the instantaneous variations in the carrier phase. Switching between the acquisition mode and the tracking mode is realized digitally, and includes programmable hysteresis, resulting in optimal performance in the presence of signals having high levels of phase noise (jitter). More specifically, the carrier recovery loop filter “locks” to the pilot signal of an incoming signal, e.g., a vestigial side band (VSB) video signal, by employing a so-called digital vector tracking phase-locked loop that demodulates the VSB signal.

Abstract: A digital timing recovery system advantageously employs both demodulated I-phase and Q-phase components to more accurately locate the synchronization signal of an incoming VSB signal. The Q-phase component is advantageously employed to detect the phase error. The use of the Q-phase component provides a more accurate measure of the phase error and results in a larger (wider) acquisition range for timing frequency offset. More specifically, the timing recovery system of this invention performs symbol clock recovery based on the VSB signal segment synchronization (sync) signal and generates a pulse density modulated (PDM) phase difference signal that controls a voltage controlled crystal oscillator (VCXO) in the phase-locked loop. This is realized, in one embodiment of the invention, by correlating received sync segment data with the known sync signal pattern and searching for "peaks" in the correlation values that are periodic at the known sync segment data rate.

Abstract: A new filter that may be employed in timing recovery circuits, and automatic gain control circuits, employs a so-called vector tracking filter (VTF). The VTF includes a complex filter with a time constant so long that it might be considered a leaky integrator. In operation, the VTF builds up an average vector (timing (gain) estimate vector) having a direction that is the average of the estimated timing (gain) error, which is stored in the VTF. When the VTF is employed in an automatic gain control arrangement, the vector becomes a scalor and only the amplitude is tracked. When an arbitrary timing (gain) correction is made, to the incoming signal, this causes a rotation of the timing (gain) estimate vector. In order to track this rotation, a comparable rotation is made to the stored timing (gain) estimate vector.

Abstract: Timing recovery circuits, automatic gain control circuits and the like, employ a so-called vector tracking filter (VTF) in conjunction with other timing recovery techniques. The VTF includes a complex filter with a time constant so long that it might be considered a leaky integrator. In operation, the VTF builds up an average vector (timing (gain estimate vector) having a direction that is the average of the estimated timing (gain) error, which is stored in the VTF. When an arbitrary timing (gain) correction is made, to the incoming signal, this causes a rotation of the timing (gain) estimate vector. In order to track this rotation, a comparable rotation is made to the stored timing (gain) estimate vector. This allows the stored timing (gain) estimate vector to build up, and at any time the stored timing (gain) estimate vector will be the same as if the current incoming phase had been constant for all previous time.