Abstract: A signal processing system (50) performs real-time pitch shifting for applications such as karaoke, tapeless answering machines, and the like while minimizing distortion. A digital input signal is sampled and stored at successive locations in a variable-size buffer (62) at an input sample rate. Data from the variable-size buffer (62) is interpolated according to a pitch-shifting ratio. An adaptive pitch estimator (61) continually estimates the fundamental frequency of the digital input signal, and the signal processing system (50) adjusts the buffer size of the variable-size buffer (62) in response thereto. The signal processing system (50) changes the buffer size to store the digital input signal for an integral number of periods of the estimated fundamental frequency.

Abstract: The characteristics of a room in which a speakerphone (20) is located are measured by determining a time between a test signal and its first attack, and a number of sample periods between the first attack and a time when the average power in the echo falls below a threshold. The first-attack time determines a pre-filter delay and the number of sample periods determines a tap length for an adaptive echo-canceling filter (62). In a teleconferencing environment, an annoying initialization sequence is avoided by initializing filter coefficients for each microphone (140), and saving the initial filter coefficients generated thereby in a corresponding nonvolatile memory (104). In response to an off-hook signal, the coefficients are retrieved from the nonvolatile memory (104). During operation, the coefficients are constantly updated. If another microphone (141) is enabled, the stored coefficients corresponding to that microphone (141) are dynamically substituted for the present coefficients.

Abstract: A method of dual-tone multifrequency (DTMF) detection which decimates and adaptively filters an input signal is provided to efficiently detect a presence of a DTMF signal. The input signal is provided to a half-band filter (14) to be decimated in frequency in accordance with Nyquist's theory. A decimated input signal is subsequently processed to form a low frequency component signal and a high frequency component signal. The low frequency component signal is again decimated by a decimator (24). The decimated low frequency component signal and the high frequency component signal are each filtered by an adaptive fir filter (22, 26) to provide a first and a second frequency parameter and a first and a second gain factor, respectively. The first and second frequency parameters and the first and second gain factors are then tested by a tone identifier (28) to determine if the input signal includes a valid DTMF signal.

Abstract: In an asynchronous communication system such as a V.32 modern (80), an input signal is sampled at a near-end clock rate. Each sample is then interpolated in an interpolation filter (92) to provide corresponding interpolated values. The interpolation filter (92) uses a selected one of a predetermined number of sets of windowed sinc function coefficients, each set having a successively greater phase offset. A time drift between near-end and far-end clocks is measured by tracking the coefficient shift in a passband phase-splitting, fractionally-spaced equalizer (95). When the time drift exceeds a threshold, a subsequent set of windowed sinc function coefficients is selected. When the time drift exceeds the threshold after the last set of coefficients is used, the first set is again selected and an interpolated value is either dropped or repeated in forming the far-end data samples.

Abstract: A digital receiver (10) has a symmetrical base band processor (28, 30, 32, 34, and 36) which concurrently provides a left and a right channel of audio information. Because of the symmetrical design of the base band processor, each of the left and right channels of information has perfectly balanced phase and amplitude with respect to each other. The base band processor includes an adaptive gain compensator (28), multipliers (30, 32), and high pass filters (34, 36). Compensator (28) iteratively derives a gain factor which is multiplied by both an in-phase component and a quadrature component of a quadrature modulated input signal to respectively provide a composite of the left and right channels and a difference between the left and right channels. Subsequently, adders (44, 48) arithmetically manipulate the composite and difference of the channels to separate the left and the right audio information.

Abstract: A digital receiver, such as "C-QUAM" receiver (10), has phase error correction. In another form, a software program may be executed by a conventional digital signal processor to also implement phase error correction. A digital input signal is demodulated to form an in-phase and a quadrature component. The in-phase and quadrature components are processed by a digital envelope detector (24) to form a composite signal containing left and right audio channel information. The in-phase component and composite signal are both processed by a reciprocal cosine estimator (28) and a quadrature channel circuit (38) to provide a difference signal also containing left and right audio channel information. The difference signal is input to phase error correction circuitry (16, 22, 26) to estimate a phase error of the digital input signal. The estimated phase error is then used to correct an actual phase error of the digital input signal during demodulation.