Abstract: Methods, devices, and systems for voice activity detection. An audio signal is received by receiver circuitry. A pitch analysis is performed on the received audio signal by pitch analysis circuitry. A higher-order statistics analysis is performed on the audio signal by statistics analysis circuitry. Logic circuitry determines, based on the pitch analysis and the higher-order statistics analysis, whether the audio signal includes a voice region. The logic circuitry outputs a signal indicating that the audio signal includes voice if the audio signal was determined to include a voice region or indicating that the audio signal does not include voice if the audio signal was determined not to include a voice region.

Abstract: An apparatus including circuitry configured for: defining at least one ambience audio representation, the ambience audio representation includes at least one respective diffuse background audio signal and at least one parameter, the at least one parameter associated with the at least one respective diffuse background audio signal and further associated with at least one frequency range or at least one part of the frequency range, at least one time period or at least one part of the time period and a directional range for a defined position within an audio field, wherein the at least one ambience component representation is configured to be used in rendering an ambiance audio signal by a 6-degrees-of-freedom or enhanced 3-degrees-of-freedom Tenderer by processing, based on the at least one ambience audio representation and a listener position and/or direction, the respective diffuse background audio signal.

Abstract: Controlling a concealment method for a lost audio frame associated with a received audio signal is provided. At least one bin vector of a spectral representation for at least one tone is obtained, wherein the at least one bin vector includes three consecutive bin values for the at least one tone. Whether each of the three consecutive bin values has a complex value or a real value is determined. Responsive to the determination, the three consecutive bin values are processed to estimate a frequency of the at least one tone based on whether each bin value has a complex value or a real value.

Abstract: Communication terminal includes a first microphone system, a second microphone system, and a noise reduction processing unit (NRPU). The NRPU receives a primary signal from the first microphone system and a secondary signal from the second microphone system. The NRPU dynamically identify an optimal transfer function of a correction filter which can be applied to the secondary signal provided by the second microphone system to obtain a correction signal. The correction signal is subtracted from the primary signal to obtain a remainder signal which approximates a signal of interest contained within the primary signal.

Abstract: Techniques disclosed herein are directed towards streaming keyphrase detection which can be customized to detect one or more particular keyphrases, without requiring retraining of any model(s) for those particular keyphrase(s). Many implementations include processing audio data using a speaker separation model to generate separated audio data which isolates an utterance spoken by a human speaker from one or more additional sounds not spoken by the human speaker, and processing the separated audio data using a text independent speaker identification model to determine whether a verified and/or registered user spoke a spoken utterance captured in the audio data. Various implementations include processing the audio data and/or the separated audio data using an automatic speech recognition model to generate a text representation of the utterance.

Abstract: An apparatus for decoding an encoded audio signal to obtain a reconstructed audio signal is provided, having: a receiving interface for receiving one or more frames, a coefficient generator, and a signal reconstructor. The coefficient generator is configured to determine one or more first audio signal coefficients, and one or more noise coefficients. Moreover, the coefficient generator is configured to generate one or more second audio signal coefficients, depending on the one or more first audio signal coefficients and depending on the one or more noise coefficients.

Abstract: A method and system for providing end-to-end multi-task denoising for joint signal distortion ratio (SDR) and perceptual evaluation of speech quality (PESQ) optimization is herein disclosed. According to one embodiment, an method includes receiving a noisy signal, generating a denoised output signal, determining a signal distortion ratio (SDR) loss function based on the denoised output signal, determining a perceptual evaluation of speech quality (PESQ) loss function based on the denoised output signal, and optimizing an overall loss function based on the PESQ loss function and the SDR loss function.

Abstract: A method and an apparatus for packet loss concealment, and a decoding method and an apparatus employing same are provided. A method for time domain packet loss concealment includes checking whether a current frame is either an erased frame or a good frame after the erased frame, when the current frame is either the erased frame or the good frame after the erased frame, obtaining signal characteristics, selecting one of a phase matching tool and a smoothing tool based on a plurality of parameters including the signal characteristics, and performing a packet loss concealment processing on the current frame based on the selected tool.

Abstract: An approach is described that obtains spectrum coefficients for a replacement frame of an audio signal. A tonal component of a spectrum of an audio signal is detected based on a peak that exists in the spectra of frames preceding a replacement frame. For the tonal component of the spectrum a spectrum coefficients for the peak and its surrounding in the spectrum of the replacement frame is predicted, and for the non-tonal component of the spectrum a non-predicted spectrum coefficient for the replacement frame or a corresponding spectrum coefficient of a frame preceding the replacement frame is used.

Abstract: A method, computer program product, and computing system for receiving audio encounter information from a microphone array. Speech activity within one or more portions of the audio encounter information may be identified based upon, at least in part, a correlation among the audio encounter information received from the microphone array. Location information for the one or more portions of the audio encounter information may be determined based upon, at least in part, the correlation among the signals received by each microphone of the microphone array. The one or more portions of the audio encounter information may be labeled with the speech activity and the location information.

Abstract: A method of communication includes determining, at a mobile device, a speech quality metric for an incoming speech signal associated with a voice call. The speech quality metric is based on an environment of the mobile device. The method also includes converting incoming speech associated with the incoming speech signal to text in response to a determination that the speech quality metric fails to satisfy a speech quality metric threshold. The method further includes displaying the text at a display screen of the mobile device during the voice call.

Abstract: A voice signal processing method according to an embodiment of the present disclosure for overcoming the problem includes: acquiring a real-time near-end noise signal; acquiring a far-end voice signal according to an incoming call; measuring subjective speech quality and perceptual-objective speech quality of test signals generated based on a reference signal and the real-time near-end noise signal; selecting at least one speech quality enhancement method based on the subjective speech quality and the perceptual-objective speech quality, and determining parameters that are to be applied to the selected at least one speech quality enhancement method; and enhancing speech quality of the far-end voice signal by using the selected at least one speech quality enhancement method, based on the determined parameters, wherein the test signals are generated by mixing the acquired real-time near-end noise signal with the reference signal whose speech quality is enhanced by applying a combination of parameter values to s

Abstract: A condition monitoring apparatus which monitors a condition of equipment including a rotor includes a storage and a calculator. The storage stores a plurality of results obtained from a plurality of divided data strings, respectively, resulting from division of a data string obtained by sampling at a regular time interval of signals from sensors provided in the equipment. The calculator estimates a plurality of rotation speeds corresponding to the plurality of divided data strings from the plurality of results accumulated in the storage, respectively, corrects the plurality of divided data strings based on the plurality of rotation speeds, respectively, and generates a corrected data string by combining the plurality of corrected divided data strings. A condition monitoring apparatus capable of accurately analyzing data obtained from the sensors while increase in cost is suppressed and influence by variation in rotation is lessened is thus provided.

Abstract: Various implementations include a wearable audio device that includes an array of microphones and an audio processing system for processing audio signals from the array of microphones. The audio processing system includes a system for analyzing the audio signals and generating a first set of sub-array signals that preserve signals originating from a mouth of a user using a fixed coefficient beamformer, and generating a second set of sub-array signals associated with noise sources originating from directions other than the mouth of the user, wherein each sub-array signal includes a sub-combination of signals from the array of microphones; a primary generator that generates a primary signal by combining the first set of sub-array signals; and an adaptive filter for processing the primary signal, wherein the adaptive filter utilizes the second set of sub-array signals as a set of noise reference signals to adaptively steer a null towards the noise sources to generate an enhanced primary signal.

Abstract: An apparatus, system, and method for selectively enhancing a desired sound while suppressing noise is provided. An apparatus comprises a plurality of microphones situated at spatially diverse locations to provide microphone signals, a spatial filter coupled to the microphones, the spatial filter configured to spatially filter the microphone signals, a noise suppressor coupled to the spatial filter, the noise suppressor for suppressing noise, and an audio amplifier, the audio amplifier coupled to the noise suppressor, the audio amplifier for amplifying an audible output signal. In accordance with at least one embodiment, the noise suppressor comprises a voice activity detector coupled to the spatial filter, the voice activity detector for detecting voice activity and for selecting an updated spatial parameter value for the spatial filter to use for performing further spatial filtering.

Abstract: A method and an apparatus for packet loss concealment, and a decoding method and an apparatus employing same are provided. A method for time domain packet loss concealment includes checking whether a current frame is either an erased frame or a good frame after the erased frame, when the current frame is either the erased frame or the good frame after the erased frame, obtaining signal characteristics, selecting one of a phase matching tool and a smoothing tool based on a plurality of parameters including the signal characteristics, and performing a packet loss concealment processing on the current frame based on the selected tool.

Abstract: A method and a device for improving audio processing performance are provided. In some embodiments, the method includes: acquiring audio data in an audio call in a terminal; processing the audio data by utilizing an audio processing algorithm, to obtain a characteristic parameter of the audio data; analyzing the processing quality of the audio data according to the characteristic parameter of the audio data; and if the processing quality of the audio data does not reach a preset quality standard, optimizing the audio processing algorithm by utilizing optimization data matching the terminal.

Abstract: A wireless device is described. The wireless device includes at least two microphones on the wireless device. The microphones are configured to capture sound from a target user. The wireless device also includes processing circuitry. The processing circuitry is coupled to the microphones. The processing circuitry is configured to locate the target user. The wireless device further includes a communication interface. The communication interface is coupled to the processing circuitry. The communication interface is configured to receive external device microphone audio from at least one external device microphone to assist the processing circuitry in the wireless device to locate the target user.

Abstract: A method and device are provided for determining an optimized scale factor to be applied to an excitation signal or a filter during a process for frequency band extension of an audio frequency signal. The band extension process includes decoding or extracting, in a first frequency band, an excitation signal and parameters of the first frequency band including coefficients of a linear prediction filter, generating an excitation signal extending over at least one second frequency band, filtering using a linear prediction filter for the second frequency band. The determination method includes determining an additional linear prediction filter, of a lower order than that of the linear prediction filter of the first frequency band, the coefficients of the additional filter being obtained from the parameters decoded or extracted from the first frequency band and calculating the optimized scale factor as a function of at least the coefficients of the additional filter.

Abstract: Background noise estimators and methods are disclosed for estimating background noise in an audio signal. Some methods include obtaining at least one parameter associated with an audio signal segment, such as a frame or part of a frame, based on a first linear prediction gain, calculated as a quotient between a residual signal from a 0th-order linear prediction and a residual signal from a 2nd-order linear prediction for the audio signal segment. A second linear prediction gain is calculated as a quotient between a residual signal from a 2nd-order linear prediction and a residual signal from a 16th-order linear prediction for the audio signal segment. Whether the audio signal segment comprises a pause is determined based at least on the obtained at least one parameter; and a background noise estimate is updated based on the audio signal segment when the audio signal segment comprises a pause.

Abstract: The present disclosure generally relates to methods of using near infrared radiation, such as from a near infrared spectroscopy device, to monitor physical activity in a person. In one aspect, a method of measuring physiological parameters is provided. The method further includes determining a NIRS-derived measure of a tissue of a person using near infrared spectroscopy over a time period, associating the NIRS-derived measure to the time period to determine a function-related change, and associating the function-related change to a biomechanical function.

Abstract: Various embodiments provide a communication system that monitors environmental noise at a computing device. The communication system additionally monitors call audio received at the computing device. Based on the environmental noise and the received call audio, the communication system determines that a user of the computing device is unlikely to hear the received call audio when played back by the computing device. In response to determining that the user of the computing device is unlikely to hear the received call audio, the communication system visually displays a transcription of at least a portion of the received call audio.

Abstract: Techniques are provided in which an audio signal for transmission to a receiving device is acquired at a network device. The audio signal is analyzed for an audio feature to be suppressed or enhanced during playback of the audio signal at the receiving device. The audio feature is detected based on the analysis. The audio signal is encoded for transmission over a network to the receiving device. The encoded audio signal is transmitted to the receiving device. A packet is generated comprising an audio feature descriptor indicating where in the audio signal the audio feature is located to enable the receiving device to suppress or enhance the audio feature during playback of the audio signal. The packet comprising the audio feature descriptor is transmitted to the receiving device.

Abstract: A voice converting apparatus and a voice converting method are provided. The method of converting a voice using a voice converting apparatus including receiving a voice from a counterpart, analyzing the voice and determining whether the voice abnormal, converting the voice into a normal voice by adjusting a harmonic signal of the voice in response to determining that the voice is abnormal, and transmitting the normal voice.

Abstract: A method of removing noise and an electronic device thereof are provided. The electronic device includes a speaker, a first sensor configured to output a first signal by obtaining noise around the electronic device, a second sensor configured to output a second signal by detecting vibrations around the electronic device, and a controller configured to determine whether the second signal satisfies a condition of not exceeding a certain threshold value, and output an acoustic signal to be output through the speaker based on the first signal according to a result of the determination.

Abstract: Online source separation may include receiving a sound mixture that includes first audio data from a first source and second audio data from a second source. Online source separation may further include receiving pre-computed reference data corresponding to the first source. Online source separation may also include performing online separation of the second audio data from the first audio data based on the pre-computed reference data.

Abstract: A method and apparatus for suppressing vocoder noise are provided. In the method, first information and second information are received from a channel decoder, the first information indicating whether a decoded data frame has an error and the second information being a channel quality metric, error concealment voice decoding is performed on the decoded data frame if the first information indicates that no channel decoding error has been generated and the second information is smaller than a predetermined first threshold, and normal voice decoding is performed on the decoded data frame if the first information indicates that no channel decoding error has been generated and the second information is equal to or larger than the first threshold.

Abstract: A method for controlling an electronic apparatus which uses voice recognition, and an electronic apparatus applying the same are provided. The method includes, in response to a voice start command being input through a first microphone included in the electronic apparatus, converting a mode of the electronic apparatus to a first voice task mode and displaying a first icon corresponding to the first voice task mode; in response to the voice start command being input through a second microphone included in a remote control device, converting the mode of the electronic apparatus to a second voice task mode and displaying a second icon corresponding to the second voice task mode; displaying voice guide information; and in response to a user voice being input while the voice guide information is displayed, displaying a list including a plurality of items related to the user voice.

Abstract: Reconstructed picture quality for a video codec system may be improved by categorizing reconstructed pixels into different histogram bins with histogram segmentation and then applying different filters on different bins. Histogram segmentation may be performed by averagely dividing the histogram into M bins or adaptively dividing the histogram into N bins based on the histogram characteristics. Here M and N may be a predefined, fixed, non-negative integer value or an adaptively generated value at encoder side and may be sent to decoder through the coded bitstream.

Abstract: A method and system are described herein that employ a lost frame concealment technique for processing data frames received during transmission over a communications channel. The lost frame concealment technique involves determining whether a current data frame is a bad frame, performing source decoding on the current data frame with one or more parameters that are limited by a first set of one or more values if the current data frame is a bad frame, and performing source decoding on the current data frame with one or more parameters that are not limited if the current data frame is a good frame.

Abstract: A microphone array processing system and method carried out in the system. In one embodiment, the system includes: (1) a beamformer configured to perform adaptive beamforming on gain-compensated signals received from a plurality of microphones, the adaptive beamforming including dynamic range compression and diagonal loading of a sample correlation matrix based on order statistics and (2) a postfilter configured to receive an output of the beamformer and reduce noise components remaining from the beamforming.

Abstract: In a speech processing apparatus, an acquisition unit is configured to acquire a speech. A separation unit is configured to separate the speech into a plurality of sections in accordance with a prescribed rule. A calculation unit is configured to calculate a degree of similarity in each combination of the sections. An estimation unit is configured to estimate, with respect to the each section, a direction of arrival of the speech. A correction unit is configured to group the sections whose directions of arrival are mutually similar into a same group and correct the degree of similarity with respect to the combination of the sections in the same group. A clustering unit is configured to cluster the sections by using the corrected degree of similarity.

Abstract: A match score provides a semantically-meaningful quantification of the aural similarity of two chromae from two corresponding audio sequences. The match score can be applied to the chroma pairs of two corresponding audio sequences, and is independent of the lengths of the sequences, thereby permitting comparisons of matches across subsequences of different length. Accordingly, a single cutoff match score to identify “good” audio subsequence matches can be determined and has both good precision and good recall metrics. A function for determining the match score is determined by establishing a function PM indicating probabilities that chroma correspondence scores indicate semantic correspondences, and a function PR indicating probabilities that chroma correspondence scores indicate random correspondences, repeatedly updating PM and the match function based on existing values of PM and the match function as applied to audio subsequences with known semantic correspondences.

Abstract: An apparatus for generating spectral replacement values for an audio signal has a buffer unit for storing previous spectral values relating to a previously received error-free audio frame. Moreover, the apparatus includes a concealment frame generator for generating the spectral replacement values, when a current audio frame has not been received or is erroneous. The previously received error-free audio frame has filter information, the filter information having associated a filter stability value indicating a stability of a prediction filter. The concealment frame generator is adapted to generate the spectral replacement values based on the previous spectral values and based on the filter stability value.

Abstract: A method for speech recognition, the method includes: extracting time-frequency speech features from a series of reference speech elements in a first series of sampling windows; aligning reference speech elements that are not of equal time span duration; constructing a common subspace for the aligned speech features; determining a first set of coefficient vectors; extracting a time-frequency feature image from a test speech stream spanned by a second sampling window; approximating the extracted image in the common subspace for the aligned extracted time-frequency speech features with a second coefficient vector; computing a similarity measure between the first and the second coefficient vector; determining if the similarity measure is below a predefined threshold; and wherein a match between the reference speech elements and a portion of the test speech stream is made in response to a similarity measure below a predefined threshold.

Abstract: A signal processing apparatus includes an adder that acquires a plurality of input signals from a plurality of microphones and calculates an added value obtained by adding the input signals together, a subtracter that acquires a plurality of input signals from the plurality of microphones and calculates a subtracted value obtained by subtracting one input signal from the other input signal, and a determination unit that determines whether noise is included in the input signals based on the added value and the subtracted value.

Abstract: A coding apparatus is provided that improves the quality of a decoded signal in a hierarchical coding (scalable coding) scheme in which a coding target band is selected in each hierarchy (layer). The coding apparatus includes a first layer coding section that selects a first quantization target band of inputted spectrum and generates first layer coded information including first band information of the selected band. An adder generates a first layer difference spectrum using a first decoded signal generated using the first layer coded information and the inputted spectrum. A second layer coding section generates second layer coded information including second band information of the selected band, wherein first layer coding section determines a method of quantizing the gain of the inputted spectrum from a plurality of candidates based on the first band information and second band information.

Abstract: A method, computer program product, and computer system for inferring quality of unified telephony calls based on real-time bi-directional exploitation of a full reference algorithm. A client copy of an audio file may be received from a client device associated with a unified telephony call. The client copy of the audio file may be compared to a local copy of the audio file using a reference algorithm. A quality of service associated with the client device may be determined based upon, at least in part, a difference between the client copy of the audio file and the local copy of the audio file.

Abstract: A method for mixing of audio signals that allows maintaining of a consistent perceived sound level for the mixed signal by holding the sound level of the dominant signal in the mix constant by adjusting the sound level of the non-dominant signal(s) in relation to the dominant signal. It further includes receiving of a mixing balance input, which denotes the adjustable balance between the main and associated signals. It further includes identification of the dominant signal from the mixing balance input and mixing metadata, from which an appropriate scale factor for the non-dominant signal may also be determined directly from the scaling information, without the need for any analysis or measurement of the audio signals to be mixed. It further includes scaling the non-dominant signal in relation to the dominant signal and combining the scaled non-dominant signal with the dominant signal into a mixed signal.

Abstract: Provided is a multi-point connection device including: a first signal receiving unit which receives a first signal containing a plurality of constituent elements and first analysis information expressing the relationship between the constituent elements contained in the first signal; a second signal receiving unit which receives a second signal containing a plurality of constituent elements and second analysis information expressing the relationship between the constituent elements contained in the second signal; a signal mixing unit which mixes the first signal and the second signal; and an analysis information mixing unit which mixes the first analysis information and the second analysis information.

Abstract: A signal processing device includes a noise analysis unit for analyzing a frequency component of a noise signal obtained by converting a collected sound into an electrical signal, a plurality of filtering units for carrying out predetermined filtering operations on the noise signal on the basis of an analysis result, and an output control unit for temporally varying a synthesis rate of outputs of the plurality of filtering units according to a change in the analysis result of the noise analysis unit. When the analysis result of the noise analysis unit changes, one filtering unit starts a predetermined filtering operation by characteristics different from those of other filtering units that carry out predetermined filtering operations on the noise signal according to the change in the analysis result of the noise analysis unit.

Abstract: An encoder for providing an audio stream on the basis of a transform-domain representation of an input audio signal includes a quantization error calculator configured to determine a multi-band quantization error over a plurality of frequency bands of the input audio signal for which separate band gain information is available. The encoder also includes an audio stream provider for providing the audio stream such that the audio stream includes information describing an audio content of the frequency bands and information describing the multi-band quantization error. A decoder for providing a decoded representation of an audio signal on the basis of an encoded audio stream representing spectral components of frequency bands of the audio signal includes a noise filler for introducing noise into spectral components of a plurality of frequency bands to which separate frequency band gain information is associated on the basis of a common multi-band noise intensity value.

Abstract: Techniques for a computing device operating in limited-access states are provided. One example method includes determining, by a computing device, that a notification is scheduled for output by the computing device during a first time period and that a pattern of audio detected during the first time period is indicative of human speech. The method further includes delaying output of the notification during the first time period and determining that a pattern of audio detected during a second time period is not indicative of human speech. The method also includes outputting at least a portion of the notification at an earlier in time of an end of the second time period or an expiration of a third time period.

Abstract: In prediction of a speech quality evaluation score such as a phone speech, even when a background noise exists, a subjective opinion score is predicted with high precision. A speech quality evaluation system that outputs a predicted value of the subjective opinion score for an evaluation speech such as a far-end speech of a phone, includes a speech distortion calculation unit that conducts, after calculating frequency characteristics of the evaluation speech, a process of subtracting given frequency characteristics from frequency characteristics of the evaluation speech, and calculates the speech distortion on the basis of the frequency characteristics after the subtracting process has been conducted, and a subjective evaluation prediction unit that calculates the predicted value of the subjective opinion score on the basis of the speech distortion.

Abstract: Methods and systems for using pitch predictors in speech/audio coders are provided. Techniques for optimal pre- and post-filtering are presented, and a general result that post-filtering is more effective than pre-filtering is derived. A practical paired-zero filter design for the low-rate regime is proposed, and this design is extended to handle frequency-dependent periodicity levels. Further, the methods described provide a general performance measure for a post-filter that only uses information available at the decoder, thereby allowing for the optimization or selection of a post-filter without increasing the rate.

Abstract: A voice quality measurement device that measures voice quality of a decoded voice signal outputted from a voice decoder unit. The voice quality measurement device includes a packet buffer unit and a voice information monitoring unit. The packet buffer unit accumulates voice packets that arrive non-periodically as voice information, and outputs the voice information to the voice decoder unit periodically. The voice information monitoring unit monitors continuity of the voice information inputted to the voice decoder unit, and calculates an index of voice quality of the decoded voice signal that reflects acceptability of this continuity.

Abstract: MDCT or FFT-based audio coding algorithms often have the problem named here spectral pre-echoes when coding an energy attack signal. This invention presents several possibilities to avoid the spectral pre-echoes existing in decoded signal segment before the energy attack point. The spectral envelope before the attack point can be improved by performing spectrum smoothing, replacing the segment of having spectral pre-echoes or filtering the segment with a combined filter obtained by doing LPC analysis.

Abstract: An audio apparatus is provided. The audio apparatus includes an input configured to receive an audio signal containing noise; a period estimation unit configured to estimate a period of a noise pattern in the audio signal; a noise reducer configured to subtract and remove the noise pattern from the audio signal in a frequency domain by using the estimated period of the noise pattern; a noise updater configured to update the noise pattern according to a change in amplitude of the noise; and an output configured to output the audio signal obtained by removing the noise pattern.

Abstract: A howling canceller which suppresses occurrence of howling even when an open loop gain exceeds “1” in the whole reproduction band. In the howling canceller, an adaptive filter (107) operates a digital received voice signal with a tap coefficient to generate a pseudo echo; a subtractor (108) subtracts the pseudo echo from a digital transmitted voice signal to generate a residual signal; and an amplitude limiting circuit (110) limits the absolute value of the amplitude of the digital received voice signal to be equal to or smaller than a predetermined threshold which ensures that all of a D/A converter (101), a power amplifier (102), a speaker (103), a microphone (104), a microphone amplifier (105), and an A/D converter (106) operate in a linear operation area, and outputs the amplitude-limited digital received voice signal to the D/A converter (101) and the adaptive filter (107).

Abstract: Described herein are multi-channel noise suppression systems and methods that are configured to detect and suppress wind and background noise using at least two spatially separated microphones: at least one primary speech microphone and at least one noise reference microphone. The multi-channel noise suppression systems and methods are configured, in at least one example, to first detect and suppress wind noise in the input speech signal picked up by the primary speech microphone and, potentially, the input speech signal picked up by the noise reference microphone. Following wind noise detection and suppression, the multi-channel noise suppression systems and methods are configured to perform further noise suppression in two stages: a first linear processing stage that includes a blocking matrix and an adaptive noise canceler, followed by a second non-linear processing stage.