Abstract: An audio device is disclosed, comprising multiple microphones and an audio module. The multiple microphones generate multiple audio signals. The audio module coupled to the multiple microphones comprises a processor, a storage media and a post-processing circuit. The storage media includes instructions operable to be executed by the processor to perform operations comprising: producing multiple instantaneous relative transfer functions (IRTFs) using a known adaptive algorithm according to multiple spectral representations for multiple first sample values in current frames of the multiple audio signals; and, performing distractor suppression over the multiple spectral representations and the multiple IRTFs using an end-to-end neural network to generate a compensation mask. The post-processing circuit generates an audio output signal according to the compensation mask.

Abstract: A lens assembly is provided, comprising a first lens, a second lens, a limiting component, a positioning component, and a connecting component. The first lens is stacked with the second lens. The limiting component comprises a first limiting piece and a second limiting piece, a receiving groove is formed between the first limiting piece and the second limiting piece. The positioning component comprises a positioning bulge, the positioning bulge is arranged on a side of the second lens near the first lens. The connecting component comprises a first connecting structure and a second connecting structure, the first connecting structure is arranged on the first lens, the second connecting structure is arranged on the second lens, the second connecting structure is bonded with the first connecting structure.

Abstract: A lens assembly comprises a first lens, a second lens, a third lens and a baffle ring. The first lens comprises a first surface. The second lens comprises a second surface, in an assembly direction, the first surface is spaced from the second surface. The baffle ring is positioned between the first lens and the second lens, the baffle ring comprises a first end and a second end, the first end is connected to the first surface, the second end is connected to the second surface, a pouring space is formed between the first surface, the second surface and an inner surface of the baffle ring. And the third lens is formed by solidifying a pouring material poured into the pouring space. The application can improve the production yield of the lens assembly.

Abstract: An audio signal enhancement system is disclosed. The audio signal enhancement system includes a neural network-like human voice detection module and a frequency shifting module. The neural network-like human voice detection module is used to detect a human voice of an input signal it receives. The frequency shifting module is coupled to the neural network-like human voice detection module and used to perform a frequency shifting process on the human voice to generate an enhanced human voice.

Abstract: A fan control system applied to N fans inside a computer system is disclosed, comprising: a main microphone, a control circuitry, a wave generation circuitry and a number N of fan controllers. The control circuitry calculates a basic frequency value according to a temperature inside the computer system, and continuously updates a parameter by any known optimization algorithm according to a main audio signal from the main microphone. The wave generation circuitry generates N square waves according to the basic frequency value and the parameter. The N fan controllers respectively form and transmit N modulation signals to the N fans according to the N square waves and N tachometric signals from the N fans. The parameter is one of a frequency variation and a set of phase differences, and the N square waves have the same frequency.

Abstract: A hearing device is disclosed, comprising a main microphone, M auxiliary microphones, a transform circuit, a processor, a memory and a post-processing circuit. The transform circuit transforms first sample values in current frames of a main audio signal and M auxiliary audio signals from the microphones into a main and M auxiliary spectral representations. The memory includes instructions to be executed by the processor to perform operations comprising: performing ANC over the first sample values using an end-to-end neural network to generate second sample values; and, performing audio signal processing over the main and the M auxiliary spectral representations using the end-to-end neural network to generate a compensation mask. The post-processing circuit modifies the main spectral representation with the compensation mask to generate a compensated spectral representation, and generates an output audio signal according to the second sample values and the compensated spectral representation.

Abstract: A fan control system applied to N fans inside a computer system is disclosed, comprising: a main microphone, a control circuitry, a wave generation circuitry and a number N of fan controllers. The control circuitry calculates a basic frequency value according to a temperature inside the computer system, and continuously updates a parameter by any known optimization algorithm according to a main audio signal from the main microphone. The wave generation circuitry generates N square waves according to the basic frequency value and the parameter. The N fan controllers respectively form and transmit N modulation signals to the N fans according to the N square waves and N tachometric signals from the N fans. The parameter is one of a frequency variation and a set of phase differences, and the N square waves have the same frequency.

Abstract: There is provided a smart engine including a profile collector and a main processing module. The profile collector is configured to store a plurality of profiles, one or more suitable profiles being dynamically selected according to an instruction from a user or an automatic selector. The main processing module is connected to the profile collector and directly or indirectly connected to a sensor, and configured to perform a detailed analysis to determine detailed properties of features, objects, or scenes based on suitable sensor data from the sensor.

Abstract: A hearing device is disclosed, comprising a main microphone, M auxiliary microphones, a transform circuit, a processor, a memory and a post-processing circuit. The transform circuit transforms first sample values in current frames of a main audio signal and M auxiliary audio signals from the microphones into a main and M auxiliary spectral representations. The memory includes instructions to be executed by the processor to perform operations comprising: performing ANC over the first sample values using an end-to-end neural network to generate second sample values; and, performing audio signal processing over the main and the M auxiliary spectral representations using the end-to-end neural network to generate a compensation mask. The post-processing circuit modifies the main spectral representation with the compensation mask to generate a compensated spectral representation, and generates an output audio signal according to the second sample values and the compensated spectral representation.

Abstract: A speech enhancement apparatus is disclosed and comprises an adaptive noise cancellation circuit, a blending circuit, a noise suppressor and a control module. The ANC circuit filters a reference signal to generate a noise estimate and subtracts a noise estimate from a primary signal to generate a signal estimate based on a control signal. The blending circuit blends the primary signal and the signal estimate to produce a blended signal. The noise suppressor suppresses noise over the blended signal using a first trained model to generate an enhanced signal and a main spectral representation from a main microphone and M auxiliary spectral representations from M auxiliary microphones using (M+1) second trained models to generate a main score and M auxiliary scores. The ANC circuit, the noise suppressor and the trained models are well combined to maximize the performance of the speech enhancement apparatus.

Abstract: Provided is a speech recognition system including an acoustic model, a decoding graph module, a history buffer, and a decoder. The acoustic model is configured to receive an acoustic input from an input module, divide the acoustic input into audio clips, and return scores evaluated for the audio clips. The decoding graph module is configured to store a decoding graph having at least one possible path of the keyword. The history buffer is configured to store history information corresponding to the possible path in the decoding graph module. The decoder is connected to the acoustic model, the decoding graph module, and the history buffer, and configured to receive the scores from the acoustic model, loop up the possible path in the decoding graph module, and predict an output keyword.

Abstract: A speech enhancement apparatus is disclosed and comprises an adaptive noise cancellation circuit, a blending circuit, a noise suppressor and a control module. The ANC circuit filters a reference signal to generate a noise estimate and subtracts a noise estimate from a primary signal to generate a signal estimate based on a control signal. The blending circuit blends the primary signal and the signal estimate to produce a blended signal. The noise suppressor suppresses noise over the blended signal using a first trained model to generate an enhanced signal and a main spectral representation from a main microphone and M auxiliary spectral representations from M auxiliary microphones using (M+1) second trained models to generate a main score and M auxiliary scores. The ANC circuit, the noise suppressor and the trained models are well combined to maximize the performance of the speech enhancement apparatus.

Abstract: A spread spectrum based audio frequency communication system at least includes a transmitting apparatus. The transmitting apparatus includes a first dot-product module, a summation module, a transmitting modulation module, a mixture module, a digital-to-analog converter, and a transmitter. The first dot-product module is configured to perform a dot-product of a first data and a first pseudo-noise code, and derive a first spreading data. The summation module is configured to sum up the first spreading data and a second spreading data to form a summed data. The transmitting modulation module is configured to vary a carrier signal with the summed data to form a modulated signal. The mixture module is configured to mix the modulated signal and an acoustic signal up to form a mixed signal. The digital-to-analog converter is configured to convert the mixed signal into acoustic waves. The transmitter transmits the acoustic waves.

Abstract: A method for matrix by vector multiplication, applied in an artificial neural network system, is disclosed. The method comprises: compressing a plurality of weight values in a weight matrix and indices of an input vector into a compressed main stream; storing M sets of synapse values in M memory devices; and, performing reading and MAC operations according to the M sets of synapse values and the compressed main stream to obtain a number M of output vectors. The step of compressing comprises: dividing the weight matrix into a plurality of N×L blocks; converting entries of a target block and corresponding indices of the input vector into a working block and an index matrix; removing zero entries in the working block; shifting non-zero entries row-by-row to one of their left and right sides in the working block; and, respectively shifting corresponding entries in the index matrix.

Abstract: There is provided a smart engine including a profile collector and a main processing module. The profile collector is configured to store a plurality of profiles, one or more suitable profiles being dynamically selected according to an instruction from a user or an automatic selector. The main processing module is connected to the profile collector and directly or indirectly connected to a sensor, and configured to perform a detailed analysis to determine detailed properties of features, objects, or scenes based on suitable sensor data from the sensor.

Abstract: A method of manufacturing a piezoelectric element includes: forming a patterned mask layer over a substrate, in which the patterned mask layer has an opening exposing a portion of the substrate; forming a piezoelectric element in the opening; and removing the patterned mask layer to obtain the piezoelectric element, in which the piezoelectric element has a central portion and a peripheral portion adjacent to the central portion, and the peripheral portion has a maximum height greater than a height of the central portion.

Abstract: A local learning system in a local artificial intelligence (AI) device includes at least one data source, a data collector, a training data generator, and a local leaning engine. The data collector is connected to the at least one data source, and used to collect training data. The training data generator is connected to the data collector, and used to analyze the training data to produce paired examples for supervised learning, or unlabeled data for unsupervised learning. The local leaning engine is connected to the training data generator, and includes a local neural network. The local neural network is trained by the paired examples or the unlabeled data in a training phase, and makes inference in an inference phase.

Abstract: A spread spectrum based audio frequency communication system at least includes a transmitting apparatus. The transmitting apparatus includes a first dot-product module, a summation module, a transmitting modulation module, a mixture module, a digital-to-analog converter, and a transmitter. The first dot-product module is configured to perform a dot-product of a first data and a first pseudo-noise code, and derive a first spreading data. The summation module is configured to sum up the first spreading data and a second spreading data to form a summed data. The transmitting modulation module is configured to vary a carrier signal with the summed data to form a modulated signal. The mixture module is configured to mix the modulated signal and an acoustic signal up to form a mixed signal. The digital-to-analog converter is configured to convert the mixed signal into acoustic waves. The transmitter transmits the acoustic waves.

Abstract: A method for matrix by vector multiplication, applied in an artificial neural network system, is disclosed. The method comprises: compressing a plurality of weight values in a weight matrix and indices of an input vector into a compressed main stream; storing M sets of synapse values in M memory devices; and, performing reading and MAC operations according to the M sets of synapse values and the compressed main stream to obtain a number M of output vectors. The step of compressing comprises: dividing the weight matrix into a plurality of N×L blocks; converting entries of a target block and corresponding indices of the input vector into a working block and an index matrix; removing zero entries in the working block; shifting non-zero entries row-by-row to one of their left and right sides in the working block; and, respectively shifting corresponding entries in the index matrix.

Abstract: A memory device is disclosed. The memory device comprises N flash memories and a flash manager. The flash manager comprises an interleave/de-interleave buffer and an addressing circuit. The interleave/de-interleave buffer operates according to a mode signal. The addressing circuit sequentially converts N input address signals to transmit N converted address signals. For write operations, the interleave/de-interleave buffer interleaves a write parameter stream into N interleaved streams according to the mode signal indicative of interleave mode and the N interleaved streams in conjunction with the N converted address signals are written into the N flash memories in parallel. For read operations, N read streams are read from the N flash memories in parallel in response to the N converted address signals and the interleave/de-interleave buffer de-interleaves the N read streams into a de-interleaved parameter stream according to the mode signal indicative of de-interleave mode.