30 results about "Distortion free" patented technology

This disclosure provides methods, devices, and systems for signal processing. The present implementations relate more specifically to a spatio-temporal beamformer. In some aspects, a beamforming system may receive an audio signal via a plurality of microphones, the audio signal including a number (B) of frames for each of the plurality of microphones, each of the B frames for each of the plurality of microphones including a number (N) of time-domain samples. For a first microphone, the beamforming system may transform the B*N time-domain samples into B*N / 2 first frequency-domain samples; transform the B*N / 2 first frequency-domain samples into B*N / 2 second frequency-domain samples; and determine a probability of speech associated with the B*N / 2 second frequency-domain samples based on a neural network model. The beamformer system may determine a minimum variance distortionless response (MVDR) beamforming filter based at least in part on the probability of speech for the first microphone.

A grating waveguide device that reduces rainbow artifacts, and a waveguide system. The grating waveguide device comprises a waveguide substrate (2) and a grating structure. The grating structure comprises a first in-coupling grating (3-1) and a first out-coupling grating (4-1), a grating region of the first out-coupling grating (4-1) having a grating line overlapping structure of multiple dimensions; alternatively, the grating structure comprises a second in-coupling grating (3-2), a fold grating (5) and a second out-coupling grating (4-2), a grating line region of the fold grating (5) and a grating line region of the second out-coupling grating (4-2) having an overlapping region. Grating vectors of the grating structure form one and only one closed path in a k-space. When a projected image of an imaging device enters an eye box via the grating structure of the waveguide substrate (2) as a projected light beam, an image displayed by the grating waveguide device is distortion-free and aberration-free compared with the projected image of the imaging device, and diffracted light generated by the diffraction of ambient light via the grating structure does not enter the eye box, such that human eyes cannot observe the rainbow artifact phenomenon in the eye box.

The invention discloses a windowing compensation method and device for frequency domain signal processing, electronic equipment and a storage medium. The windowing compensation method comprises the following steps: A, generating a window function sequence; b, generating an overlapped and added window function sequence; c, obtaining a difference value sequence; d, performing amplitude scaling and repeated expansion on the difference value sequence to generate a compensation sequence; e, obtaining a window function sequence for compensation and correction; f, obtaining a square root sequence of the compensation correction window function; and G, performing windowing processing before the FFT and after the IFFT by using the square root sequence of the compensation correction window function. According to the windowing compensation method provided by the invention, the complete distortion-free reduction of the input signal can be ensured, the problem of spectrum leakage in the FFT and IFFT conversion processes is eliminated, the overhead of computing resources is saved, and the windowing compensation method is very suitable for engineering application in a real-time scene.

A method and device for performing a perception task are disclosed. The method and device incorporate a dense regression model. The dense regression model advantageously incorporates a distortion-free convolution technique that is designed to accommodate and appropriately handle the varying levels of distortion in omnidirectional images across different regions. In addition to distortion-free convolution, the dense regression model further utilizes a transformer that incorporates an spherical self-attention that use distortion-free image embedding to compute an appearance attention and uses spherical distance to compute a positional attention.

Techniques, including devices and systems implementing the techniques, for using speech enhancement to provide optimal denoised output. One example system generally includes a device of a user, a first sensor coupled to the device, a second sensor coupled to the device, and one or more processors coupled to the device. The one or more processors are generally, individually or collectively, configured to receive, at the first sensor, a first audio signal, receive, at the second sensor, a second audio signal, determine a minimum variance distortionless response (MVDR) using at least the second audio signal, and determine a mixed audio signal using a condition of an environment of the device and at least one of the first audio signal, the second audio signal, or the MVDR.

The invention discloses a system and a method for lossless dynamic measurement of microstructure characteristics of a material surface, and relates to the technical field of optical measurement, and the method comprises the following steps: S1, placing an object to be measured on a two-dimensional stepping motor platform, and adjusting the platform to be horizontal; and starting the solid laser. According to the invention, an optical system composed of a solid laser, a beam expander, an aperture diaphragm and a spectroscope generates an accurate detection light beam, a spatial filter designed according to a confocal diffraction limit is matched to effectively suppress defocus stray light interference, and a multi-parameter fusion judgment mechanism is combined to synchronously extract a light intensity peak value and a light spot centroid offset, so that the detection accuracy is improved. The focus is judged only when the light intensity reaches the maximum value and the mass center offset is smaller than the threshold value, the positioning precision is greatly improved, Z-axis coordinate errors are dynamically corrected through real-time vibration compensation and thermal drift compensation, the problems of sample damage and optical method noise interference in traditional contact measurement are solved, and undistorted data support is provided for microstructure three-dimensional reconstruction.

A feedforward amplifier circuit, an audio amplifier, and an audio playback device relate to the field of feedforward amplifier circuits. The amplifier circuit comprises: a main amplifier circuit, an adder circuit, an auxiliary amplifier circuit, and a subtractor circuit. The main amplifier circuit adds distortion to an input signal to output a distorted signal, and inputs the distorted signal into the adder circuit and the subtractor circuit for feedforward correction. The auxiliary amplifier circuit amplifies the input signal and uses it as the positive input of the subtractor, so that the signal gain output by the subtractor equals the signal gain of the main amplifier. The subtractor circuit uses the distorted signal as a negative input, reversely amplifies the distorted signal, and then performs feedforward. The adder circuit superimposes the distorted signal output by the main amplifier circuit with the reversely amplified distorted signal output by the subtractor, and outputs a distortion-free signal. By eliminating the reliance on the open-loop frequency-amplitude characteristics and feedback loop of the main amplifier, the circuit structure and component parameter selection are simplified, and costs are saved.

The present invention discloses a method for drawing an arc without distortion for a head-up display, comprising: S1, constructing a rectangular plane coordinate system with the center of the arc as the origin, determining the range of a rectangular scanning frame by the center, radius and line width of the arc, wherein the rectangular scanning frame encompasses the arc and three mirrored arcs respectively referenced by the x-axis, y-axis and center; S2, traversing all sampling points in the rectangular scanning frame based on a multi-sampling algorithm, obtaining the distance value between the sampling point and the center of the arc, calculating the difference between the distance value and the radius, and judging whether the sampling point is within the arc angle; combining the judgment result and the calculated difference to determine the weight value to achieve an anti-aliasing effect, and completing the drawing of a distortion-free arc. The present invention can solve the problem that the anti-aliasing effect of the arc section is poor and the extension line of the arc section does not pass through the center of the circle, is applicable to any line width, radius and span, and has low hardware requirements.

The invention relates to a method for creating undistorted projections on the surface of real-world objects in the case of artistic lighting of buildings, structures and similar objects, and in the case of light markings at industrial facilities and on civil objects. The method comprises performing geodetic measurements of the objects and their spatial surroundings to create a vertical control network, performing photogrammetric measurements or laser scans of the real-world objects to be projected thereon and objects in their spatial surroundings, creating a 3D model of the real-world objects, during the rendering process, real-world objects are referenced to a given coordinate system, a projection zone is determined and calculated, the position of the projector is calculated, taking into account the position of an object in the space surrounding that blocks the path of the projection beam from the projector to the object on which it is to be projected. The method includes creating a model of a "source (projector)-screen" system using coordinates obtained in a 3D modeling process, transmitting a projected image onto a gobo picture, installing a projector according to coordinates of a virtual projector in the 3D model of the real-world object, placing the picture with the image into the projector, and projecting the image onto a surface of the real-world object.

The application relates to a design method of a robust vector array differential beamformer, belonging to the field of sound signal processing, which comprises the following steps: setting system working parameters; transmitting a detection signal; judging whether a target signal is detected; receiving and preprocessing array signals; reconstructing full-array steering vectors by using Kronecker product operation; selecting an ideal differential beam pattern form, setting target direction distortionless response constraints and non-target direction response constraints; decoupling the system total weight vector into the Kronecker product form of spatial weight and channel weight by using the Kronecker product property, taking the white noise gain maximization as an objective criterion, and solving the optimal channel weight; constructing a weight approximation model based on the Maclaurin series expansion, solving the optimal spatial weight by using the least square method, synthesizing the system total weight vector, and realizing the design of a differential beamformer of any order. The application can significantly improve the robustness of the differential beamformer and effectively suppress spatial white noise.

This invention claims protection for a mixed-frequency time-difference amplification flow measurement method based on an improved ultrasonic time-difference method, belonging to the field of ultrasonic fluid velocity measurement technology. Addressing the problem of insufficient time-difference resolution in low-frequency ultrasonic flow measurement, this invention constructs a flow measurement architecture that deeply coordinates fiber optic sensing and heterodyne mixing: 1) An interferometric fiber optic sensor is used to acquire upstream and downstream ultrasonic signals, whose high signal-to-noise ratio and phase fidelity provide a distortion-free signal basis for subsequent time-difference amplification; 2) The two signals are digitally mixed with a reference signal and low-pass filtered. Utilizing the complete phase information retained by the fiber optic sensor, the microsecond-level original time difference is losslessly converted into a significantly amplified apparent time difference on the difference frequency signal; 3) The apparent time difference is extracted by cross-correlation and then the amplification factor is used to inversely deduce the actual propagation time difference; 4) The flow velocity and flow rate are calculated. In this invention, fiber optic sensing ensures the measurability and phase accuracy of weak low-frequency signals, while heterodyne mixing breaks through the traditional time-domain resolution limit. The two complement each other and work synergistically to significantly improve the accuracy and stability of flow measurement under complex river conditions.

The present disclosure provides a speech enhancement and separation method and device based on directivity convolution beamforming, which comprises the following steps: firstly, constructing a history observation signal matrix from a microphone array through a multi-channel observation signal, and constructing a Kalman gain linear prediction error model according to the history observation signal matrix; secondly, constructing a minimum variance distortionless response beamforming model based on a directivity beamformer and a maximum null beamformer, and estimating a speech plus noise covariance matrix; thirdly, using a time-varying variance based on a separated speech source estimation, combining the Kalman gain linear prediction error model and the minimum variance distortionless response beamforming model to establish a directivity convolution beamforming model, and completing the enhancement and separation of the speech signal through an alternating iteration method. The present disclosure can better suppress early reverberation and reduce late reverberation residues in a noisy reverberation environment; and through the construction of directivity gain and the real-time estimation of a noise covariance matrix, more robust speech separation performance is achieved.

The application discloses a kind of professional noise exposure individualized noise reduction earplug self-adapting filtering method and system, its method includes: outward microphone picks up environmental sound stream;DSP carries out real-time acoustic scene decomposition, decomposes sound using time-frequency analysis, and identifies "human voice", "alarm" and other safety key acoustic events in real time by lightweight CNN;DSP combines personal audiogram, and executes frequency band adaptive noise reduction strategy: attenuates reverse wave to background noise frequency band;Safety key acoustic event frequency band is removed from noise reduction, and is played after individualized gain compensation according to personal audiogram.The application realizes the perfect fusion of hearing protection and context awareness, while protecting hearing, distortionlessly retains key safety information, greatly improves the wearing compliance and work safety of noise reduction earplug.

Magnetic resonance imaging (“MRI”) using a spin- and gradient-echo (“SAGE”) pulse sequence with blip up-down acquisition (“BUDA”) encoding enables distortion-free, high-resolution diffusion-weighted imaging and / or quantitative parameter mapping. Phase-encoding polarities are alternated across shots during a multi-shot acquisition. In each shot, multi-contrast data are acquired at echo times associated with a gradient echo, a mixed gradient-and-spin echo, and a spin echo. High in-plane resolution and distortion-free quantitative parameter maps can be generated, such as T2 maps. T2* maps, paramagnetic susceptibility maps, and diamagnetic susceptibility maps. Diffusion-weighted data can be acquired using diffusion encoding gradients and BUDA encoding, where multi-contrast data are acquired in the b=0 acquisition. Diffusion parameter maps can be generated from the b=0 and diffusion-weighted data.

This invention discloses an online intelligent detection and early warning system for metal processing quality based on AI vision, specifically relating to the field of metal processing quality inspection. The system employs a distributed industrial camera and intelligent light source in its image acquisition module, combined with FPGA synchronous control, to ensure clear and distortion-free images on high-speed production lines. The data preprocessing module enhances image quality through adaptive median filtering, distortion correction, and GPU acceleration. The AI ​​vision detection module accurately outputs defect locations, bounding boxes, and confidence scores based on an improved YOLOv3 model. The defect classification module uses a CNN-Transformer dual-branch network to determine five types of defects: cracks, scratches, dents, dimensional deviations, and surface deposits, and integrates dual-reset confidence verification. The early warning response module triggers graded audible and visual alarms and notifications based on defect levels, and coordinates with the PLC to execute shutdowns, marking and diverting processes, or continuous monitoring, achieving closed-loop management of "detection-classification-disposal."

The invention provides a phase-sensitive optical time domain reflectometer, a measurement method and a readable storage medium, and a signal generation circuit of the phase-sensitive optical time domain reflectometer is configured to generate a radio frequency signal to drive a light intensity modulator to modulate detection light. Therefore, synchronous detection can be carried out on the sensing optical fiber to be detected by utilizing two groups of frequency division multiplexing sequences with difference, the phase variation in adjacent sampling intervals can be accurately solved by combining a differential de-folding algorithm, then demodulation results corresponding to different frequency parts are respectively de-folded to obtain de-folded phase results, and the detection accuracy of the sensing optical fiber to be detected is improved. Therefore, the problem of large-strain phase folding is solved, the large-strain range is recovered, further time sequence interleaving processing is carried out on the unfolded phase results, the unfolded phase results are rearranged and combined according to the time sequence, and therefore a high-sampling-rate and undistorted reconstruction result is reconstructed. According to the invention, accurate measurement of high-frequency response and large-strain-range dynamic strain can be realized.

This application provides a sound perception and localization method for humanoid robots, relating to the field of robotics. It constructs a basic three-dimensional sampling array by scientifically arranging four microphones, employing high-precision omnidirectional condenser microphones to achieve distortion-free acquisition of environmental sound signals. Simultaneously, it outputs raw sound signal data, precise sound pressure amplitude data, and related identification information. The raw sound signal is preprocessed using A-weighted sound levels that conform to human hearing characteristics, while filtering out DC components and high-frequency interference noise to output a human-like sound signal. Then, based on the theory of pulsating spherical source radiation sound field and a point source approximation model, combined with the human-like sound pressure data acquired by the microphone array and preset coordinates, the three-dimensional spatial coordinates of the sound source are accurately calculated. Multiple sound source coordinate results are obtained through multi-microphone array combination calculations, thereby achieving accurate judgment and high-precision localization of the environmental sound source and its location.

Method for membrane excursion control of an electronic apparatus is disclosed herein. The method includes receiving an audio signal input from one or more audio sources and extracting audio parameters, which include speaker characteristics and signal features. These parameters are processed through a distortion classifier trained on a synthetic audio dataset generated by a diffusion model with a custom loss function. The distortion classifier identifies audio frames within the input signal that cause audible distortions. The identified frames are then processed by a plurality of artificial intelligence (AI)-based speaker models trained using the same synthetic dataset. The output of this process is a distortion-free set of audio frames, ensuring optimal playback quality.

The invention provides a speech enhancement and separation method and device based on directional convolution beam forming, and the method comprises the steps: constructing a historical observation signal matrix through a multi-channel observation signal according to a microphone array, and constructing a Kalman gain linear prediction error model according to the historical observation signal matrix; constructing a minimum variance undistorted response beam forming model based on a directional beam former and a maximum null beam former, and estimating a voice and noise covariance matrix; a directional convolution beam forming model is established by using the time-varying variance based on separated voice source estimation, a simultaneous Kalman gain linear prediction error model and a minimum variance undistorted response beam forming model, and enhancement and separation of voice signals are completed in an alternate iteration mode. According to the invention, the early reverberation can be better inhibited and the late reverberation residue can be reduced in the noise-containing reverberation-containing environment; and through construction of directional gain and real-time estimation of a noise covariance matrix, more robust voice separation performance is realized.

The present application discloses a calibration method and a calibration device. The calibration device is used to calibrate the non-linear distortion components of the IQ signals in a non-linear system. The non-linear system includes a front-end device and a digital down-converter. The calibration device includes a training signal source, a switch, a processor, and a calibration module. The training signal source and the receiving link of the external signal are selectively connected to the front-end device of the non-linear system through the switch. The digital down-converter is connected to the calibration module, and the processor is respectively connected to the training signal source, the switch, the digital down-converter, and the calibration module. The processor controls the switch to make the non-linear system in a training mode or a calibration mode. When the non-linear system is in the training mode, the processor calculates the calibration parameters and loads them into the calibration module. When the non-linear system is in the calibration mode, the calibration module performs non-linear calibration on the output signal of the digital down-converter to obtain distortion-free IQ signals. The present application can effectively remove the signal distortion generated by the non-linear system.

The present invention discloses a beamforming method, system, storage medium and electronic device having a constant beam deflection ability. The method includes obtaining a steering vector of a microphone array according to the directivity of each microphone in the microphone array and the steering vector of an omnidirectional microphone; obtaining a system of linear equations based on the steering vector, the distortionless constraint condition and the null constraint condition; constructing an optimization problem for minimizing the error between an actual beam and an ideal beam based on the system of linear equations and the white noise gain constraint condition, and obtaining a weight vector in the actual beam from the optimization problem. The beamformer designed by the present invention has a constant beam deflection ability, greatly expands its application fields, and also improves performance indexes such as its white noise gain and directivity factor.

Techniques, including devices and systems implementing the techniques, for using speech enhancement to provide optimal denoised output. One example system generally includes a device of a user, a first sensor coupled to the device, a second sensor coupled to the device, and one or more processors coupled to the device. The one or more processors are generally, individually or collectively, configured to receive, at the first sensor, a first audio signal, receive, at the second sensor, a second audio signal, determine a minimum variance distortionless response (MVDR) using at least the second audio signal, and determine a mixed audio signal using a condition of an environment of the device and at least one of the first audio signal, the second audio signal, or the MVDR.

The invention discloses a signal correction method for a pulse type terahertz system, and relates to the technical field of terahertz waves. After frequency spectrum data of terahertz signals are obtained through a terahertz waveform standard device, a distortion coefficient is calculated, frequency domain data, obtained in a terahertz system, of the terahertz signals are corrected through the distortion coefficient, and then undistorted terahertz time domain waveforms are obtained through inverse Fourier transform. The problems that in a current pulse type terahertz system, the inherent influence elimination cost of the system is high, and signals are prone to distortion are solved.

The invention provides a multi-channel FIADC design method based on second-order cone programming, belongs to the technical field of filter optimization, and aims to solve the problem that distortion-free recovery of signals cannot be realized due to errors in a reconstruction process of a digital reconstruction filter in an FIADC model. Comprising the following steps: S1, constructing initial perfect reconstruction conditions of a sampling filter bank and a digital reconstruction filter; s2, a sampling filter module is designed by adopting a standard filter design method, output signal frequency response of the sampling filter module is subjected to instantiation, and preliminary perfect reconstruction conditions are rewritten; s3, designing a digital reconstruction filter module in combination with an SOCP algorithm and the rewritten perfect reconstruction condition; and S4, completing the design of a multi-channel FIADC model, and verifying the error between the frequency spectrum of the reconstructed signal and the frequency spectrum obtained by directly sampling the broadband signal. Compared with an existing algorithm, the reconstruction error of the whole FIADC system is greatly reduced and is only 0.03, and distortion-free recovery of signals is better facilitated.

Systems and methods for improved audio signal processing include a plurality of audio input components configured to generate a plurality of audio input signals, and a logic device configured to receive the plurality of audio input signals, determine whether the plurality of audio signals includes a target audio associated with an audio source, estimate a relative position of the audio source with respect to the plurality of audio input components based on the plurality of audio signals and the determination of whether the plurality of audio signals includes the target audio, and process the plurality of audio signals to generate an audio output signal by enhancing the target audio based on the estimated relative position. The logic device is further configured to construct a cross-band aligned directional covariance matrix using a relative transfer based covariance and find a direction that minimizes a beam power under a distortionless criterion.

The present invention relates to the field of audio amplification technology, and specifically discloses a power amplifier distortion compensation system based on deep learning, comprising: an input signal acquisition module, a deep learning distortion recognition module, a vacuum tube simulation module, a distortion compensation module, a PWM generation and optimization module, a power amplification module, and a feedback learning module; the system of the present invention combines a deep learning algorithm with a multi-band dynamic feature extraction method to achieve high-precision recognition of complex nonlinear distortion types such as harmonic distortion and dynamic distortion. Compared with traditional spectrum analysis or static filtering compensation, the system can dynamically adjust the compensation strategy and significantly reduce the impact of distortion on sound quality. Through a nonlinear even harmonic enhancement algorithm, the natural timbre and auditory texture of the audio signal are highly restored. Through a GAN framework, intelligent repair of signal distortion is achieved. The generator dynamically generates a high-fidelity compensation signal, and at the same time, a discriminator learns the characteristics of different distortion types to ensure that the repaired signal is close to the distortion-free target.

This disclosure provides methods, devices, and systems for signal processing. The present implementations relate more specifically to a spatio-temporal beamformer. In some aspects, a beamforming system may receive an audio signal via a plurality of microphones, the audio signal including a number (B) of frames for each of the plurality of microphones, each of the B frames for each of the plurality of microphones including a number (N) of time-domain samples. For a first microphone, the beamforming system may transform the B*N time-domain samples into B*N / 2 first frequency-domain samples; transform the B*N / 2 first frequency-domain samples into B*N / 2 second frequency-domain samples; and determine a probability of speech associated with the B*N / 2 second frequency-domain samples based on a neural network model. The beamformer system may determine a minimum variance distortionless response (MVDR) beamforming filter based at least in part on the probability of speech for the first microphone.

A vision system for a vehicle includes a camera configured to generate one of a distorted image and an undistorted image. An alignment system is configured to generate a region of interest (ROI) of a camera, generate a bird's-eye view (BEV) homography to convert undistorted points in the ROI into BEV points, generate a BEV image based on one of a distorted image and the undistorted image and one of the BEV homography and an updated BEV homography, and generate an alignment image based on the BEV homography and one of the updated BEV homography and the one of the distorted image and the undistorted image and the one of the updated BEV homography and the one of the updated BEV homography and the one of the updated BEV homography and the one of the updated BEV homography and the one of the updated BEV homography and the one of the distorted image and the undistorted image. The method includes adaptively updating a BEV homography to generate an updated BEV homography, detecting a current position of a feature in a BEV image, predicting a future position of the feature for a next BEV image, comparing the current position of the feature in the BEV image to a previous predicted position, and generating an updated BEV homography. And converting BEV points in the BEV image into undistorted points in the perspective undistorted image using an inverse of the updated BEV homography.

A method for processing images to remove distortions of various types without requiring retraining for per-type processing models, applied in an electronic device, establishes a distortion removal model, and inputs distorted images in the distortion removal model. General distortions are removed by the distortion removal model and the restored images are input into a deep learning model for final image processing. The deep learning model of the present disclosure finally corrects distortion in the images so that distortion-free images are obtained.

The invention discloses a combined estimation method and system for array shape correction and noise source positioning during towed line array rotation, and a calculation device. The combined estimation method comprises the following steps: preliminarily estimating the positions of array elements of a towed line array; preliminarily positioning a sound source position; calculating the delay time of each array element relative to the first array element in the towed line array, and calculating a steering vector and an array manifold matrix; establishing a target cost function and a constraint condition to optimize and solve the correction of each array element in the towed line array; calculating the corrected array element position of the towed line array according to the array element position under the preliminarily estimated array shape and the correction amount of each array element of the towed line array; on the basis of a second-order polynomial fitting method, curve fitting is carried out on the positions of the array elements, and the positions of the array elements of the towed line array after curve fitting are obtained; and correcting the array flow pattern matrix based on the new array element position, and positioning the sound source position by using the minimum variance undistorted response algorithm again under the new array flow pattern.