Abstract: A method for improving the robustness of a speech processing system having at least one speech processing module comprises: receiving an input sound signal comprising audio and non-audio frequencies; separating the input sound signal into an audio band component and a non-audio band component; and identifying possible interference within the audio band from the non-audio band component. Based on such an identification, the operation of a downstream speech processing module is adjusted.

Abstract: A method for use in a biometric process, comprising: for a first function and a second function, applying an acoustic stimulus to a user's ear; and for the second function: receiving a response signal of a user's ear to the acoustic stimulus; and extracting, from the response signal, one or more features for use in a biometric process, wherein the first function is a function other than to induce the response signal for use in the biometric process.

Abstract: A method of audio processing comprises receiving an audio signal. A plurality of framed versions of the received audio signal are formed, each of the framed versions having a respective frame start position. One of the plurality of framed versions of the received audio signal is selected. The selected one of the plurality of framed versions of the received audio signal is used in a subsequent process.

Abstract: Audio circuitry, comprising: a speaker driver operable to drive a speaker based on a speaker signal; a current monitoring unit operable to monitor a speaker current flowing through the speaker and generate a monitor signal indicative of that current; and a microphone signal generator operable, when external sound is incident on the speaker, to generate a microphone signal representative of the external sound based on the monitor signal and the speaker signal.

Abstract: This application relates to time-encoding modulators (TEMs). A TEM receives an input signal (SIN) and outputs a time-encoded output signal (SOUT). A filter arrangement receives the input signal and also a feedback signal (SFB) from the TEM output, and generates a filtered signal (SFIL) based, at least in part, on the feedback signal. A comparator receives the filtered signal and outputs a time-encoded signal (SPWM) based at least in part on the filtered signal. The time encoding modulator is operable in a first mode with the filter arrangement configured as an active filter and in a second mode with the filter arrangement configured as a passive filter. The filter arrangement may include an op-amp, capacitance and switch network. In the first mode the op-amp is enabled, and coupled with the capacitance to provide the active filter. In the second mode the op-amp is disabled and the capacitance coupled to a signal path for the feedback signal to provide a passive filter.

Abstract: Embodiments of the disclosure include methods, apparatus and computer programs for detecting the proximity of an ear to an audio device. In one embodiment, the disclosure provides a system for detecting the presence of an ear in proximity to an audio device. The system comprises: an input for obtaining a data signal from the environment of the audio device; and an ear biometric authentication module configured to: compare one or more ear biometric features, extracted from the data signal, to an ear biometric template; and generate a first output indicative of the presence or absence of any ear in proximity to the audio device based on the comparison of the one or more extracted ear biometric features to the ear biometric template.

Abstract: The present disclosure relates to a neuron for an artificial neural network. The neuron comprises a dot product engine operative to: receive a set of weights; receive a set of data inputs based on a set of input data signals; and calculate the dot product of the set of data inputs and the set of weights to generate a dot product engine output. The neuron further comprises an activation function module arranged to apply an activation function to a signal indicative of the dot product engine output to generate a neuron output; and gain control circuitry. The gain control circuitry is operative to control: an input gain applied to the input data signals to generate the set of data inputs; and an output gain applied to the dot product engine output or by the activation function module. The output gain is selected to compensate for the applied input gain.

Abstract: The present disclosure relates to a neuron for an artificial neural network. The neuron includes: a first dot product engine operative to: receive a first set of weights; receive a set of inputs; and calculate the dot product of the set of inputs and the first set of weights to generate a first dot product engine output. The neuron further includes a second dot product engine operative to: receive a second set of weights; receive an input based on the first dot product engine output; and generate a second dot product engine output based on the product of the first dot product engine output and a weight of the second set of weights. The neuron further includes an activation function module arranged to generate a neuron output based on the second dot product engine output. The first dot product engine and the second dot product engine are structurally or functionally different.

Abstract: A method of speaker recognition comprises receiving an audio signal representing speech. A speaker change detection process is performed on the received audio signal. A trigger phrase detection process is also performed on the received audio signal. On detecting the trigger phrase in the received audio signal, a speaker recognition process is performed on the detected trigger phrase and on any speech preceding the detected trigger phrase and following an immediately preceding speaker change.

Abstract: One aspect of the disclosure provides a device, comprising: an allocation module, for determining one or more metrics of each of a plurality of data streams; a compression module, for compressing each of the plurality of data streams and generating a plurality of compressed data streams, the compression module applying a compression ratio that varies as a function of the metrics determined by the allocation module; and a buffer memory, for storing the plurality of compressed data streams.

Abstract: The present disclosure relates to a neuron for an artificial neural network, the neuron comprising: a first dot product engine and a second dot product engine. The first dot product engine is operative to: receive a first set of weights; receive a set of inputs; and calculate the dot product of the set of inputs and the first set of weights to generate a first dot product engine output. The second dot product engine is operative to: receive a second set of weights; receive the set of inputs; and calculate the dot product of the set of inputs and the second set of weights to generate a second dot product engine output; The neuron further comprises a combiner operative to combine the first dot product engine output and the second dot product engine output to generate a combined output, and an activation function module arranged to apply an activation function to the combined output to generate a neuron output.

Abstract: A biometric processor comprises: one or more inputs configured to receive first ear biometric data acquired in respect of a first ear of a user and second ear biometric data acquired in respect of a second ear of the user; a processing module configured to perform a biometric algorithm on the first ear biometric data and the second ear biometric data, based on a comparison of the first ear biometric data to a first stored ear biometric template for an authorised user and a comparison of the second ear biometric data to a second stored ear biometric template for the authorised user, to obtain respective first and second biometric scores; a fusion module configured to apply first and second weights to the respective first and second biometric scores to obtain first and second weighted biometric scores, and to combine at least the first and second weighted biometric scores to generate an overall biometric score, wherein the first and second weights are different to each other; and wherein a biometric result is b

Abstract: This application relates to microphone authentication apparatus for verifying whether or not an audio signal originated at a microphone. The microphone authentication apparatus has a comparison block configured to receive a first signal indicative of one or more spectral parameters of at least part of an audio signal to be verified, and compare the one or more spectral parameters to one or more predetermined characteristic microphone parameters relating to a characteristic resonance associated with an acoustic port of a microphone. The first signal may be an audio signal and the microphone authentication apparatus may have a feature extract module for determining the spectral parameters. Based on the comparison determination block may whether the audio signal originated from a microphone and may send a verification signal to a voice biometric module.

Abstract: This application relates to time-encoding modulators (TEMs). A TEM (100) receives an input signal (SIN) and outputs a time encoded signal (SPWM). A comparator (101) is located within a forward signal path of a feedback loop of the TEM. Also in the feedback loop are a filter (104) and a delay element (106) for applying a controlled delay. In some embodiments a latching element (101, 302; 106, 402) is located within the forward signal path to synchronise any signal transitions output from the latching element to a received first clock signal. Any signal transitions in the output (SOUT) from the modulator are thus synchronised to the first clock signal. In some embodiments the delay element (106) is a digital delay element which is synchronised to the first clock signal.

Abstract: A method of detecting a replay attack on a voice biometrics system comprises: receiving an audio signal representing speech; detecting a magnetic field; determining if there is a correlation between the audio signal and the magnetic field; and if there is a correlation between the audio signal and the magnetic field, determining that the audio signal may result from a replay attack.

Abstract: A method of detecting a replay attack comprises: receiving an audio signal representing speech; identifying speech content present in at least a portion of the audio signal; obtaining information about a frequency spectrum of each portion of the audio signal for which speech content is identified; and, for each portion of the audio signal for which speech content is identified: retrieving information about an expected frequency spectrum of the audio signal; comparing the frequency spectrum of portions of the audio signal for which speech content is identified with the respective expected frequency spectrum; and determining that the audio signal may result from a replay attack if a measure of a difference between the frequency spectrum of the portions of the audio signal for which speech content is identified and the respective expected frequency spectrum exceeds a threshold level.

Abstract: There is described a two-terminal multi-level memristor element synthesised from binary memristors, which is configured to implement a variable resistance based on unary or binary code words. There is further described a circuit such as a synapse circuit implemented using a multi-level memristor element.

Abstract: This application describes methods and apparatus for speaker recognition. An apparatus according to an embodiment has an analyzer (202) for analyzing each frame of a sequence of frames of audio data (AIN) which correspond to speech sounds uttered by a user to determine at least one characteristic of the speech sound of that frame. An assessment module (203) determines, for each frame of audio data, a contribution indicator of the extent to which the frame of audio data should be used for speaker recognition processing based on the determined characteristic of the speech sound. In this way frames which correspond to speech sounds that are of most use for speaker discrimination may be emphasized and/or frames which correspond to speech sounds that are of least use for speaker discrimination may be de-emphasized.

Abstract: A method of speaker recognition comprises receiving an audio signal representing speech. A speaker change detection process is performed on the received audio signal. A trigger phrase detection process is also performed on the received audio signal. On detecting the trigger phrase in the received audio signal, a speaker recognition process is performed on the detected trigger phrase and on any speech preceding the detected trigger phrase and following an immediately preceding speaker change.

Abstract: This application relates to method and apparatus for driving acoustic transducers, such as speakers or haptic transducers. A transducer driver circuit (200) has a hysteretic comparator (201) configured to compare, with hysteresis, an input signal (SIN) received at a first comparator input to a feedback signal (SFB) received at a second comparator input. Based on the comparison the hysteretic comparator (201) generates a pulse-width modulation (PWM) signal (SPWM) at a comparator output (206). An inductor (203) is coupled between the comparator output and an output node (204). In use a resistive component (208), which may comprise the transducer (301) is coupled to output node (204). The inductor (203) and resistive component (208) provide filtering to the PWM signal (SPWM). A feedback path extends between the output node (204) and the second comparator input to provide the feedback signal (SFB).