Abstract: Methods and systems for optimizing and audio communication system, including detecting a surrounding noise profile, determining one or more states of one or more noise-related conditions corresponding to the surrounding noise profile, and associating the one or more states of the one or more noise-related conditions to the surrounding noise profile.

Abstract: Methods and systems for optimizing and audio communication system, including detecting a surrounding noise profile, determining one or more states of one or more noise-related conditions corresponding to the surrounding noise profile, and associating the one or more states of the one or more noise-related conditions to the surrounding noise profile.

Abstract: Methods and systems for selectively switching between microphones in a plurality of microphones are disclosed, including providing a first state that corresponds to one or more microphones selected from the plurality of microphones. A subset of microphones may be selected from the plurality of microphones in response to determining an average power of an input signal for each of the plurality of microphones. A second state may be identified that includes at least one of the subset of microphones in response to evaluating the average powers of the input signals for the subset of microphones against a predetermined condition. A transition from the first state to the second state may be delayed in response to determining a transition delay time corresponding to the first state.

Abstract: Methods and systems for optimizing and audio communication system, including detecting a surrounding noise profile, determining one or more states of one or more noise-related conditions corresponding to the surrounding noise profile, and associating the one or more states of the one or more noise-related conditions to the surrounding noise profile.

Abstract: A method for reducing a computational complexity of an m-stage adaptive filter is provided by updating recursively forward and backward error prediction square terms for a first portion of a length of the adaptive filter, and keeping the updated forward and backward error prediction square terms constant for a second portion of the length of the adaptive filter.

Abstract: Methods and systems for selectively switching between microphones in a plurality of microphones are disclosed, including providing a first state that corresponds to one or more microphones selected from the plurality of microphones. A subset of microphones may be selected from the plurality of microphones in response to determining an average power of an input signal for each of the plurality of microphones. A second state may be identified that includes at least one of the subset of microphones in response to evaluating the average powers of the input signals for the subset of microphones against a predetermined condition. A transition from the first state to the second state may be delayed in response to determining a transition delay time corresponding to the first state.

Abstract: A method for reducing a computational complexity of an m-stage adaptive filter is provided by expanding a weighted sum of forward prediction error squares into a corresponding binomial expansion series, expanding a weighted sum of backward prediction error squares into a corresponding binomial expansion series, and determining coefficient updates of the adaptive filter with the weighted sums of forward and backward prediction error squares approximated by a select number of terms of their corresponding binomial expansion series.

Abstract: A comfort noise generator (104) suitable for use in a communication system includes a finite impulse response (FIR) filter (136), a random number generator (140), and a coefficient updater (138). The coefficient updater (138) determines an updated set of filter coefficients (142) based on the signal frame of the input signal (102). The updated set of filter coefficients (142) is output to the FIR filter (136). The FIR filter (136) shapes a white noise signal (146) supplied by the random number generator (140) to provide a simulated background noise signal, or comfort noise signal (122). The comfort noise signal (122) is selectively output from an echo suppression system or corresponding method to overwrite or suppress reflected residual echoes.

Abstract: Methods and apparatus are provided for an echo cancellation system. The echo cancellation system comprises an automatic gain control (AGC), a first scalar, a second scalar, a signal summing stage, and an adaptive filter. The AGC is responsive to a first signal in a reference path and a second signal in a near end path. The signal summing stage is between the first and second scalar in the near end path and the first and second scalars are responsive to the AGC. The adaptive filter is responsive to the first signal and provides a third signal to the signal summing stage that corresponds to an echo signal and is subtracted from a signal in the near end path. The second scalar has a scale rate that is the inverse of the scale rate of the first scalar such that unity scaling occurs on the near end path.

Abstract: An echo canceling system receives and transmits audio signals between a far end and a near end. During single talk, which is when only one end is originating audio, the path back to the originator is impeded by echo cancellation and attenuation. When there is double talk, which is when both ends are originating audio, the attenuation is removed, or at least significantly reduced. This is achieved by using ERLE, which itself is a known signal used for other purposes in an echo cancellation system, to provide information as to when double talking is occurring. This allows for stopping the attenuation for the double talk situation, which is the desired result.

Abstract: A speech recognition system (10) having a sampler block (12) and a feature extractor block (14) for extracting time domain and spectral domain parameters from a spoken input speech into a feature vector. A polynomial expansion block (16) generates polynomial coefficients from the feature vector. A correlator block (20), a sequence vector block (22), an HMM table (24) and a Viterbi block (26) perform the actual speech recognition based on the speech units stored in a speech unit table (18) and the HMM word models stored in the HMM table (24). The HMM word model that produces the highest probability is determined to be the word that was spoken.

Abstract: A comfort noise generator (104) suitable for use in a communication system includes a finite impulse response (FIR) filter (136), a random number generator (140), and a coefficient updater (138). The coefficient updater (138) determines an updated set of filter coefficients (142) based on the signal frame of the input signal (102). The updated set of filter coefficients (142) is output to the FIR filter (136). The FIR filter (136) shapes a white noise signal (146) supplied by the random number generator (140) to provide a simulated background noise signal, or comfort noise signal (122). The comfort noise signal (122) is selectively output from an echo suppression system or corresponding method to overwrite or suppress reflected residual echoes.

Abstract: A voice decoder detects channel errors and loss of cryptographic synchronization using the change in spectral energy between sequential frames. The change in energy between frames is determined between corresponding LSP's of said successive frames and summed together. A running average of the change in energy for a predetermined number of frames is maintained. Current voice frames are eliminated based on the difference between the change in energy associated with the current frame and the running average. Accordingly, offensive audio associated with such channel errors or cryptographic synchronization loss is eliminated.

Abstract: A speech recognition system (10) having a sampler block (12) and a feature extractor block (14) for extracting time domain and spectral domain parameters from a spoken input speech into a feature vector. A polynomial expansion block (16) generates polynomial coefficients from the feature vector. A correlator block (20), a sequence vector block (22), an HMM table (24) and a Veterbi block (26) perform the actual speech recognition based on the speech units stored in a speech unit table (18) and the HMM word models stored in the HMM table (24). The HMM word model that produces the highest probability is determined to be the word that was spoken.

Abstract: An apparatus comprises a first and a second input and a processor coupled to the inputs. The processor estimates peak and minimum levels of each of the first and second input signals. A first signal regulator is coupled to the first input and to a first output for delivering a first output signal from the first signals. A second signal regulator is coupled to the second input and to a second output for delivering a second output signal from the second audio signals. A gain adjustment device is coupled to the processor and the first and second signal regulators. The gain adjustment device provides control signals to the first and second signal regulators to adjust an output signal level of the first output signal in response to minimum and peak levels of the first and second input signals. The output signal level is continuously variable over a range.

Abstract: A new method for Code Excited Linear Predictive (CELP) coding of speech reduces the computational complexity by removing a convolution operation from a recursive loop used to poll the adaptive code book vectors. In a preferred embodiment, an impulse function of a short term perceptually weighted filter is first convolved with perceptual weighted target speech and the result cross-correlated with each vector in the codebook to produce an error function. The vector having the minimum error function is chosen to represent the particular speech frame being examined.

Abstract: A low-power dual voltage drive circuit and method which includes controlling a stepper motor overdriven with a high voltage. The stepper motor drive circuit includes a high voltage supply for overdriving the stepper motor, a low voltage supply, and an overdrive switch mechanism connected between the high voltage supply means and the stepper motor input. The method includes repetitively triggering the overdrive switch in response to a control signal, overdriving the stepper motor with the high voltage, producing an overdrive current, sensing the overdrive current, producing an additional control signal when the overdrive current reaches a preset value, switching off the overdrive current in response to the second control signal, resetting the overdrive switch to enable the overdrive switch to again receive the first control signal, and driving the stepper motor with a nominal low voltage.

Abstract: A new way of determining correlation coefficients for stochastic codebook vectors for CELP coding of speech takes advantage of the sparsely populated nature of stochastic codebook vectors. N valued input signals (e.g., convolution vectors) to be correlated with N valued codebook vectors are fed to an N by N multiplexer or other selection means and the signal values either passed to an accumulator or not according to the state of N select inputs or other identification means determined from a memory store (e.g., an EPROM) whose entries correspond to the non-zero values of the codebook vectors. The accumulator output is the correlation of the codebook vector with the input signal. A sequencer steps through the entire codebook to provide correlation values for each vectors. The results are used to determine the optimum stochastic codebook vector for replicating the particular speech frame being analyzed.

Abstract: A new way of determining autocorrelation coefficients for adaptive codebook vectors for CELP coding of speech simplifies and improves the accuracy of the autocorrelation coefficient determination for the situation where the codebook vector length being analyzed is less than a speech frame length. This is important in synthesizing short pitch period speech. Copy-up of the shortened codebook vector to equal the frame length is not needed and autocorrelation coefficient errors associated with copy-up are avoided. The improved system relies on calculating autocorrelation coefficients of the first (shortest) vector and then obtaining subsequent autocorrelation coefficients for successive vectors of increasing length by a simple end correction technique until the vector length equals the frame length. The autocorrelation coefficients are scaled by multiplying them by the ratio of the frame length to the vector length.

Abstract: A new way of CELP coding speech simplifies the recursive loop used to poll code adaptive book vectors by reducing the number of autocorrelation operations that must be performed with the K vectors of the codebook each having N entries. Autocorrelation is initially performed for only a small number P<<N autocorrelation coefficients in each codebook vector and the values found are used to scan through all the codebook vectors looking for those S vectors (S<K) which give the best match to the input speech. The autocorrelation function for the S vectors is then recalculated for R entries (P<R.ltoreq.N) in the codebook vectors to determine which of the S codebook vectors and associated gain gives the best match to the input speech.