Abstract: Devices, systems, and methods of enhanced automatic speech recognition. An acoustic microphone senses or captures acoustic signals that are uttered by a human speaker. An optical microphone or laser microphone transmits a laser beam aimed towards the face of the human speaker; receives a reflected optical feedback signal; and produces an optical self-mix signal by self-mixing interferometry. The self-mix signal is used by a training unit of a speech recognition processor. Optionally, the self-mix signal is used by an utterance recognition unit of the speech recognition processor. Optionally, the self-mix signal is utilized for enhancing the acoustic signals, or for constructing a digital filter that is applied to the acoustic signal; and the enhanced acoustic signal, or the filtered acoustic signal, is used by the training unit or by the a recognition unit of the speech recognition processor.

Abstract: System and method for producing enhanced speech data associated with at least one speaker. The process of producing the enhanced speech data comprises: receiving distant signal data from a distant acoustic sensor; receiving proximate signal data from a proximate acoustic sensor located closer to the speaker than the distant acoustic sensor; receiving optical data originating from an optical unit configured for optically detecting acoustic signals in an area of the speaker and outputting data associated with speech of the speaker; processing the distant and proximate signals data for producing a speech reference and a noise reference; operating an adaptive noise estimation module, which identifies stationary and/or transient noise signal components, using the noise reference; and operating a post filtering module, which uses the optical data, speech reference and identified noise signal components for creating an enhanced speech data.

Abstract: A method and adaptive processor for estimating a reference signal, comprising receiving or determining a number of groups and respective group sizes, each group associated with a range of delay values of the reference signal, determining a multiplicity of coefficients, each coefficient associated with a specific delay of the reference signal, determining a multiplicity of weights, each weight associated with one of the groups, multiplying each sample of the reference signal having a delay by a corresponding coefficient to obtain a first product, and summing a multiplicity of first products associated with a group into a group sum signal sample, and multiplying each group sum by a weight associated with the group to obtain a second product, and summing all second products to obtain an estimated signal value.

Abstract: A method and apparatus for filtering a reference signal by a series of taps having coefficients, comprising dividing the taps into at least one layer of a plurality of groups, each group comprising a plurality of taps, as a first layer, forming a plurality of coefficients arranged in at least one group, as at least one subsequent layer and forwarding multiply-accumulate results of each group in each layer sequentially to the at least one group in the at least one subsequent layer to generate an output signal, and adjusting the coefficients based on an error signal formed as a difference between the output signal and a received input signal related to the reference signal and an interim error signal as a difference between the input signal and a summation of the multiply-accumulate results of at least one layer.

Abstract: A method of determining if an echo canceller comprised in a communication channel that is used to transmit at least one tonal signal and characterized by a model echo path of the channel, is acceptably adapted to cancel echoes in the channel, the method comprising: determining a performance measure how well the canceller cancels echoes of at least one test signal characterized by at least one frequency different from a frequency of the at least one tonal signal; and determining if the canceller is acceptably adapted responsive to the measure.

Abstract: A method for Comfort Noise Generation (CNG) comprising the steps of recording information of Background Noise (BGN); generating white noise samples; and generating Comfort Noise (CN) by applying coefficients that are extracted from said information of BGN on White Noise (WN) samples.

Abstract: The method and system use a speaker available in an audio end device when not in use as a microphone. The speaker is utilized for enhancing the sound quality received by the person or device at the near end, i.e., the user of the device, or the speaker at the far end, i.e., the speaker with whom the near-end speaker is communicating. Exemplary applications are shown for such utilization, including single or dual microphone near-end automatic listening enhancement, acoustic echo cancellation or suppression with or without double talk detection, and microphone array noise reduction.

Abstract: Disclosed is a method and circuit for echo suppression or cancellation. According to some embodiments of the present invention, there is provided a circuit including a Noise-Enhanced-Receiving-Direction unit and Narrow-Band Remover unit or units.

Abstract: As part of a system and method for transmitting a signaling tone over a packet switched data (or other communication) network, an audio encoder may have a speech encoding unit and a tone distorter. The tone distorter may distort one or more components of a signaling tone within an audio signal and the speech encoding unit may encode speech and the distorted signaling tone within the audio signal. A control signal may be sent to a tone generator to reproduce the distorted tone.

Abstract: A method for using a waveform segment in place of a missing portion of an audio waveform generated in response to a packet stream encoding portions of the audio waveform, the method comprising: phase matching a trailing portion of the waveform segment with a trailing portion of the audio waveform that follows the missing portion; and adding the phase matched waveform segment to the audio waveform.

Abstract: Disclosed is an echo suppresser or cancellation circuit including a speech signal extrapolation unit. The speech extrapolation unit may provide extrapolated signal segment approximations to an echo segment replacement unit. The echo segment replacement unit may replace a segment of a first speech signal suspected of being corrupted with an echo component of a second signal using the extrapolated signal segment approximated to represent the corrupted signal segment.

Abstract: A method of maintaining signal convergence of a transmitter and a receiver of an adaptive differential pulse code modulated (ADPCM) communication system during discontinuous transmission, by generating synchronized comfort noise frame in the transmitter and in the receiver during quiet periods and mutually updating the receiver's decoder and encoder on their states.

Abstract: A method of error protection for wireless communication of packets of sampled digitized audio data according to a selected communication protocol, comprising: identifying the least significant bits of each sample of audio data encoded in the packets based on the communication protocol used; selectively replacing one or more of the least significant bits with values that enable performing error correction on the rest of the data encoded in the packet; transmitting the altered packet from a transmitting party to a receiving party; using the replaced bits to detect or correct errors in the transmitted packet; wherein said communication protocol uses an ADPCM encoding scheme.

Abstract: As part of a system and method for transmitting a signaling tone over a packet switched data (or other communication) network, an audio encoder may have a speech encoding unit and a tone distorter. The tone distorter may distort one or more components of a signaling tone within an audio signal and the speech encoding unit may encode speech and the distorted signaling tone within the audio signal. A control signal may be sent to a tone generator to reproduce the distorted tone.

Abstract: A method for using a waveform segment in place of a missing portion of an audio waveform generated in response to a packet stream encoding portions of the audio waveform, the method comprising: phase matching a trailing portion of the waveform segment with a trailing portion of the audio waveform that follows the missing portion; and adding the phase matched waveform segment to the audio waveform.

Abstract: A method of determining if an echo canceller comprised in a communication channel that is used to transmit at least one tonal signal and characterized by a model echo path of the channel, is acceptably adapted to cancel echoes in the channel, the method comprising: determining a performance measure how well the canceller cancels echoes of at least one test signal characterized by at least one frequency different from a frequency of the at least one tonal signal; and determining if the canceller is acceptably adapted responsive to the measure.

Abstract: Disclosed is a method and circuit for echo suppression or cancellation. According to some embodiments of the present invention, there is provided a circuit including a Noise-Enhanced-Receiving-Direction unit and Narrow-Band Remover unit or units.

Abstract: Disclosed is an echo suppresser or cancellation circuit including a speech signal extrapolation unit. The speech extrapolation unit may provide extrapolated signal segment approximations to an echo segment replacement unit. The echo segment replacement unit may replace a segment of a first speech signal suspected of being corrupted with an echo component of a second signal using the extrapolated signal segment approximated to represent the corrupted signal segment.

Abstract: A method apparatus and system are provided for detecting pulses in a received signal wherein the spacing between the pulses may alternate in an expected manner. Generally, the method exploits intrinsic characteristics in the ensemble pulses, by performing correlation between pulses in a data stream representative of the received signal. Pulses may be related to an ensemble of pulses if their correlation result complies with a predefined criteria and the spacing between the correlated pulses is substantially the expected spacing. The correlation may be performed between two successive pulses, or it may be performed between non-successive pulses. The pulses may be alternating “start” and “stop” pulses of telephonic dial (rotary) pulses, and the expected spacing between the pulses may alternate between two values, “break” period and “make” period. Pulses may be correlated if their energy exceeds a predefined threshold value.

Abstract: As part of a system and method for transmitting a signaling tone over a packet switched data (or other communication) network, an audio encoder may have a speech encoding unit and a tone distorter. The tone distorter may distort one or more components of a signaling tone within an audio signal and the speech encoding unit may encode speech and the distorted signaling tone within the audio signal. A control signal may be sent to a tone generator to reproduce the distorted tone.