97 results about "Speech denoising" patented technology

The invention relates to the technical field of dual-mode interphones, and discloses an AI-based dual-mode interphone satellite communication noise reduction method and system, and the method comprises the steps: carrying out the orbit equation solving of satellite ephemeris data, and obtaining a satellite elevation angle and a channel fading parameter; receiving a first voice signal through a dual-mode interphone, and identifying a modulation mode according to the first voice signal; calculating a first noise reduction mask according to the channel fading parameter and the modulation mode; compressing the standard noise reduction neural network to form a lightweight noise reduction model; and starting an FM spectrum subtraction branch or a TETRA filtering branch in the lightweight noise reduction model according to the modulation mode, and performing frequency domain processing on the first noise reduction mask and the first voice signal to obtain a second voice signal, thereby providing stable and reliable real-time voice noise reduction capability for the portable dual-mode interphone in a dynamic satellite environment.

The invention relates to an array microphone noise reduction recording method based on cascade noise reduction and blind source separation, which belongs to the technical field of voice signal processing and recording, and comprises the following steps: configuring a multi-channel array microphone, ensuring that the amplitude and phase of a channel signal are consistent, and collecting an original multi-channel voice signal; a weighted kernel function blind source separation algorithm is adopted, signal-to-noise ratio distribution characteristics of signals are extracted, kernel function weights are given, and target voice and interference signal components are obtained through decoupling of an independent component analysis model; executing target-oriented adaptive cascade noise reduction, locking the voice of a keynote speaker through directional pickup, reducing noise, filtering out reverberation, and enhancing the voice of a far-field target by combining a voice mask neural network with a far-field pickup algorithm in sequence; and processing the target voice through voice feature perception lossless coding and storing the target voice. According to the invention, stable acquisition of multi-channel signals, accurate separation of mixed signals and layered suppression of noise reverberation are realized, the signal-to-noise ratio and definition of far-field voice are significantly improved, and the method is suitable for single-person speaking or multi-person dialogue scenes.

The invention particularly relates to a voice noise reduction method suitable for different noise environments, and relates to the technical field of voice signal processing. Analyzing a noise environment and characteristics; adaptive selection of a noise reduction algorithm; adaptive noise reduction processing is executed; and noise reduction effect evaluation and iterative optimization are carried out. According to the invention, refined noise classification is matched with the algorithm, so that the limitation of a traditional single algorithm in a complex environment is solved; the steady-state noise is dynamically counteracted by adopting an adaptive filter, the frequency domain gain of the unsteady-state noise is adjusted in real time through Wiener filtering, and the pulse noise is subjected to dynamic threshold processing by combining median filtering and wavelet transform; especially, the design of cooperation factors in wavelet transform can accurately distinguish signal details and noise: intensively suppress pulse noise, loosely retain details such as voices and consonants, and realize dynamic balance of noise reduction intensity and signal distortion.

The invention discloses a voice noise reduction method and device for a multi-person scene, electronic equipment and a storage medium, and relates to the technical field of voice signal processing. The method comprises the following steps: acquiring an audio signal and a video image of a target space scene; determining audio perception information matched with the audio signal and a visual voice activity detection result corresponding to the video image; fusing the audio perception information and the visual voice activity detection result, and determining a target voice sounding object when the voice activity fusion result indicates that the voice activity exists; and positioning a human face corresponding to the target voice sounding object to obtain position change information, updating a pickup direction, controlling beam forming processing, and outputting a denoised target voice signal. According to the scheme provided by the invention, the voice of the current effective speaker can be accurately separated and enhanced from the aliasing audio signals, meanwhile, the interference of other speakers and environmental noise is inhibited, and help is provided for realizing high-quality voice interaction and recognition.

The invention relates to the technical field of intelligent communication, in particular to an AI (Artificial Intelligence)-based communication voice noise reduction method, which comprises the following steps of: 1, collecting noise of multiple scenes, recording noisy voice through a microphone array, and covering complex environments such as traffic, restaurants, meetings and the like; 2, constructing a dialect voice database, adding a noise mixing layer (such as sudden whistling and keyboard tapping) and a scene evolution layer, and simulating a real environment; according to the method, multi-scene noise is collected, noisy voice is recorded through a microphone array, a dialect voice database is constructed, a noise mixing layer is added, voice features of different dialects in a complex environment are simulated, a deep learning model is deployed on edge equipment, and low-delay noise reduction is achieved through real-time processing. Mechanical noise, aerodynamic noise and external environment noise are distinguished through the dynamic noise recognition module, the noise reduction intensity is adjusted to adapt to scenes such as elevators, meeting rooms and vehicles, and therefore the method can adjust the noise reduction intensity in a self-adaptive mode.

The invention relates to a voice signal filtering and noise reduction method based on a Mama generative adversarial network, and the method comprises the following steps: designing a noise synthesis strategy for different application scenes, collecting real environment noise to construct a mixed noise library, and constructing a voice data set with noise and a voice data set without noise, and the voice data set and the voice data set do not need to be matched; constructing a Mama-based generative adversarial network, and realizing end-to-end conversion from the noisy voice to a time sequence filtering result; the construction of the model comprises the steps of designing a generator network, designing a discriminator network and optimizing a loss function. According to the invention, the robustness and generalization ability of voice signal noise reduction can be improved.

The present invention discloses a speech enhancement method that combines hearing loss compensation and speech noise reduction, comprising: extending and embedding a hearing loss map along the frequency axis to obtain a hearing loss spectrum, and superimposing the hearing loss spectrum with the complex spectrum features of the noisy training speech; constructing a metric generative adversarial network model based on frequency-time convolution recursion, the model's main structure comprising a compensation generator and a metric discriminator; alternately training the compensation generator and the metric discriminator to optimize the metric generative adversarial network model; superimposing the complex spectrum features of the test speech with the hearing loss spectrum and inputting them into the trained compensation generator, and reconstructing an enhanced speech waveform of the test speech based on the output of the compensation generator. The present invention uses a metric generative adversarial network to simultaneously achieve noise reduction and hearing loss compensation for a specific audiogram, capable of stably and effectively improving the effectiveness of hearing loss compensation in noisy environments. The method is ingenious and novel, with promising application prospects.

A noise reduction method and device, a terminal sound transmission method and device, and a server noise reduction processing method and device, the noise reduction device comprising: a sound signal acquisition module for generating a sound signal according to collected sound information, the sound information comprising a voice and a first environment sound; the memory is suitable for storing a voice noise reduction program, the voice noise reduction program is generated according to an environment sound signal, and the environment sound signal is generated according to a second environment sound; and the processor is suitable for executing the voice noise reduction program to carry out noise reduction processing on the sound signal. According to the invention, the voice noise reduction program suitable for the current scene can be obtained, and sound signal noise reduction requirements in different scenes can be met.

The invention discloses a noise suppression method and system for a microphone, and belongs to the technical field of audio signal processing. The method comprises the following steps: main audio processing equipment acquires an original audio stream to obtain complex frequency spectrum data of a current frame; synchronously acquiring real-time operation state data of the audio processing equipment; dynamically generating a task allocation strategy; generating a subtask processing result, and performing data exchange; constructing a complete full-band gain matrix for the current frame; and reconstructing a denoised time domain audio signal through inverse transformation. According to the invention, the noise reduction task is intelligently decomposed and dynamically allocated to the heterogeneous computing units of the main and auxiliary devices for cooperative processing through a cooperative scheduling mechanism, so that the technical effect of realizing high-quality and low-delay real-time voice noise reduction on the mobile terminal with limited resources is achieved; the problems of processing delay, asynchronous state updating, reduced noise reduction quality and incoherent voice caused by insufficient computing power of the terminal in the prior art are solved.

An embodiment of this application provides a speech processing method and apparatus, and an apparatus for speech processing, applied to a terminal device, where the terminal device is equipped with at least two microphones. The method includes: performing summation on signals received by the at least two microphones to obtain a first signal, and performing subtraction on the signals received by the at least two microphones to obtain a second signal; performing blind separation on the first signal and the second signal to obtain a speech signal and a noise signal; and performing adaptive noise cancellation on the speech signal based on the noise signal to obtain a target speech signal. The embodiments of this application can optimize a speech denoising effect, and further improve the speech recognition accuracy of the terminal device in a complex and changeable environment with large noise or strong interference.

The present disclosure provides a method for training a noise reduction model, a speech noise reduction method, an apparatus and an electronic device. The method comprises: obtaining source noisy speech data, and performing reverberation processing on the source noisy speech data to generate corresponding training samples, wherein the training samples comprise a plurality of samples, and each sample comprises multi-channel reverberation noisy speech data; training a to-be-trained speech noise reduction model based on the multi-channel reverberation noisy speech data included in the training samples to obtain a trained first speech noise reduction model. In the present disclosure, the human ear has a good listening experience for the noise reduction speech output by the trained speech noise reduction model, improves the propagation accuracy of the information carried in the speech data, reduces the influence of noise on the propagation of the information carried in the speech data, improves the robustness of the speech noise reduction model, strengthens the applicability and practicality of the speech noise reduction model, and optimizes the training method and training effect of the speech noise reduction model.

The present application relates to the field of speech processing technology and discloses a two-stage speech denoising method, apparatus, computer device, computer-readable storage medium and computer program product based on amplitude spectrum and complex spectrum. The method comprises performing a preprocessing operation on the acquired original speech signal and dividing it into a number of time frames according to a preset interval; determining the amplitude, phase information and complex spectrum of each different frequency component; using the amplitude of each different frequency component to perform noise estimation, performing preliminary noise suppression on the original noisy speech to obtain a preliminary denoised amplitude spectrum; based on the preliminary denoised amplitude spectrum, combined with the phase information, converting it into a preliminary denoised complex spectrum; using the complex spectrum of the original noisy speech and the preliminary denoised complex spectrum to perform noise estimation, and converting it into a complex spectrum of the target speech signal. The present application has the effect of improving the denoising accuracy when processing noisy speech with a low signal-to-noise ratio or a mixture of multiple types of noise.

The invention discloses a voice noise reduction method and system based on subspace filtering, a medium and equipment, and belongs to the technical field of software engineering and communication, and the method comprises the steps: carrying out the framing processing of noisy voice; extracting an autocorrelation coefficient and a Mel-frequency cepstral coefficient of each frame, splicing the autocorrelation coefficient and the Mel-frequency cepstral coefficient to form a feature vector, inputting the feature vector into a noise estimation network, and predicting to obtain a noise autocorrelation coefficient of each frame; according to the autocorrelation coefficient of each frame, the noise autocorrelation coefficient and the unit matrix, constructing a subspace separation matrix and carrying out eigenvalue decomposition to obtain eigenvalues and a eigenmatrix; converting the framing data into a subspace by using the feature matrix to obtain a subspace signal in each direction; calculating a subspace gain matrix in combination with the characteristic value, and filtering the subspace signal to obtain a noise reduction signal; and performing inverse transformation on the noise reduction signal according to the feature matrix to obtain noise reduction voice frames, and performing frame combination to obtain noise reduction voice. Therefore, by implementing the invention, the voice noise reduction effect can be improved, and the calculation amount and complexity of the algorithm can be reduced.

The application provides a speech noise reduction network training method and device, electronic equipment and computer readable storage medium, comprising: performing short-time Fourier transform on sample speech data in a sample data set to obtain sample time-frequency domain features; the sample speech data includes noise speech data and clean speech data, and the sample time-frequency domain features include noise time-frequency domain features and clean time-frequency domain features; calculating the noise time-frequency domain features through a neural network model to obtain predicted time-frequency domain features; evaluating the difference between the predicted time-frequency domain features and the clean time-frequency domain features through a loss function to obtain a function value, and determining whether the function value is less than a preset loss threshold; wherein the loss function is switched periodically during the training process; if yes, it is determined that the neural network model converges, and a speech noise reduction network is obtained. According to the application, the speech noise reduction network considering noise reduction and speech fidelity effect is trained without increasing model parameters.

The present application relates to a method for speech noise reduction with controllable time delay, comprising the following steps: framing the noisy speech, transforming the noisy speech into a complex spectrum of the noisy speech in the time-frequency domain; determining a gain function based on the complex spectrum of the noisy speech; the gain function being a real number or a complex number; determining a time domain filter based on the gain function, wherein the order of the time domain filter is set according to the time delay requirement; inputting the noisy speech into the time domain filter for noise reduction processing that meets the time delay requirement to obtain pure speech. The method proposed in the embodiment of the present application can achieve advanced speech noise reduction performance under low-latency conditions, reduce computational complexity, and improve robustness.

The embodiment of the invention relates to the technical field of vehicles, and provides a vehicle-mounted communication method and device and electronic equipment. The method comprises the following steps: acquiring a multi-modal signal of a first user, wherein the multi-modal signal comprises a first sound signal and an image signal of the first user; determining text information corresponding to the multi-modal signal based on the multi-modal signal of the first user; determining a second sound signal corresponding to the text information based on the text information; and outputting the second sound signal to the second user. Therefore, according to the technical scheme provided by the embodiment of the invention, the original voice is not output to the second user after noise reduction, but the new voice, namely the second sound signal, is synthesized to the second user, so that the interference of noise can be avoided, and the sound quality of the vehicle-mounted call voice is ensured.

The present invention discloses a speech noise reduction method, apparatus, device, and computer-readable storage medium. The method comprises: obtaining preliminary noise reduction data obtained by performing preliminary noise reduction on original speech data collected by a microphone; determining correction standard data corresponding to each first frequency point based on a speech model, wherein the speech model is spectral data determined based on a Gaussian mixture model; comparing the frequency point data of each first frequency point in the preliminary noise reduction data with the correction range defined by the correction standard data of the corresponding frequency point, correcting the frequency point data that exceeds the corresponding correction range to fall within the corresponding correction range, obtaining corrected noise reduction data, and using the corrected noise reduction data as the noise reduction result of the original speech data. The present invention provides a speech noise reduction scheme that further improves the speech noise reduction effect by further correcting the noise reduction speech data after preliminary noise reduction using a speech noise reduction algorithm.

The application provides an interactive method with voice noise reduction, a computer device, a storage medium and a system. The interactive method with voice noise reduction comprises the following steps: obtaining a user voice signal picked up by a voice collector; performing voice noise reduction on the user voice signal; uploading the user voice signal after voice noise reduction to a voice recognition cloud; obtaining semantic information obtained after processing by the voice recognition cloud; and generating a task code based on the semantic information, the task code being used to be transmitted to an executing mechanism of an indoor robot to solve the problem that the robot can accurately locate the position of a user in a complex indoor environment.

The application discloses a voice noise reduction method and device based on a memory-computing integrated chip, a medium and equipment. The method comprises the following steps: extracting weight data corresponding to a preset level of a voice noise reduction model to obtain a first weight matrix; performing positive and negative weight division processing on the first weight matrix to obtain a second weight matrix; performing weight quantization on the second weight matrix to obtain a third weight matrix; adding an additional weight to the third weight matrix to obtain a target weight matrix; and adjusting the conductance value of a memristor in a pre-constructed memory-computing integrated chip according to the elements of the target weight matrix to embed the voice noise reduction model in the memory-computing integrated chip to obtain a voice noise reduction chip. The voice noise reduction model is integrated on the memory-computing integrated chip to reduce noise, the difficulty of deploying the voice noise reduction model on the terminal side is reduced, and the efficiency of voice noise reduction is improved.

The invention discloses a voice noise reduction method and device, an earphone, a medium and a program product. The method comprises the following steps: acquiring voice sequence data; performing at least one self-attention operation on the voice sequence data; obtaining the output of the self-attention mechanism; noise reduction is carried out according to the output of the self-attention mechanism; the sg attention operation comprises the following steps: carrying out linear processing on the input sequence data by utilizing a preset query-key matrix and a preset value matrix to obtain a query-key vector and a value vector; the query-key matrix corresponds to a product between a preset query matrix and a preset key matrix; obtaining a target attention score according to the transpose of the voice sequence data and the query-key vector; carrying out Softmax operation on the target attention score to obtain an attention weight; and performing matrix multiplication on the attention weight and the value vector to obtain the output of the self-attention operation. In this way, the harsh requirement of the embedded device for high real-time performance in a voice interaction scene can be met.

A speech noise reduction method, an apparatus, a terminal device, a computer readable storage medium, and a computer program product, relating to the technical field of audio processing. The method comprises: collecting a first microphone signal by means of a microphone, the first microphone signal comprising a speech signal of a first speaker (901); on the basis of a speech feature of the first speaker, running a TSE algorithm to perform noise reduction on the first microphone signal (902); and when a running stop condition of the TSE algorithm is detected, stopping running the TSE algorithm, the condition comprising at least one of: the speaker moving away from and then approaching the microphone again and the speaker changing (903). In this way, if the speaker moves away from and then approaches the microphone again, it is highly probable that the speaker in a call has changed, and then running of the TSE algorithm is stopped, so that the problem that the call quality is seriously affected due to seriously damaging the speech of the speaker caused by changing the speaker but continuing to run the TSE algorithm can be avoided. Alternatively, the running of the TSE algorithm is stopped when it is determined that a speaker change has occurred, thereby also ensuring the quality of a voice call.

The invention relates to the technical field of sailor assessment management, and particularly discloses an online management method and system based on sailor assessment, and the method comprises the steps: collecting an original voice signal during sailor assessment, and a propeller rotating speed, a sea condition grade and a ship navigational speed at a corresponding moment; the original voice signals and the physical parameters are input into a physical constraint noise reduction network to serve as a regularization item embedding loss function, ocean background noise is separated, and pure sailor voice signals are output; inputting the pure sailor voice signals into a navigation term recognition network, calculating the similarity between voice features and IMO standard term prototype vectors, and outputting term accuracy scores and term confusion point information; based on the pure crew voice signal, the term accuracy score and the sea condition level, a voice signal-to-noise ratio, term accuracy and a sea condition influence factor are fused to calculate a communication capability index; generating a personalized performance evaluation report and pushing training courses; according to the invention, the problems of difficult voice noise reduction and disjunction between evaluation and training in a complex marine environment are solved.

The invention particularly relates to a voice noise reduction method based on deep spanning feature extraction and feature cross fusion, which comprises the following steps: a double-flow network structure comprising an amplitude feature extraction branch and a phase feature extraction branch is constructed, the amplitude feature extraction branch adopts a multi-scale coding-decoding structure and a time sequence convolutional network for joint modeling, and the phase feature extraction branch adopts a multi-scale coding-decoding structure; local amplitude details are extracted to be dependent on long-range time; the phase feature extraction branch uses a hierarchical decoding structure for recovering multi-level phase structure features. Moreover, a cross-flow feature fusion mechanism is designed between double flows, so that the amplitude features and the phase features are dynamically complementary in the deep network. According to the method, acoustic physical priori and data driving characteristics are combined, noise reduction can be accurately carried out under complex background noise, and the performance of a voice noise reduction task in a non-stationary noise environment is effectively improved; therefore, problems of insufficient time-frequency correlation modeling, insufficient amplitude and phase feature utilization and the like of the existing speech enhancement technology in a complex noise scene can be solved.

The application discloses a cockpit voice noise reduction function automatic test method, device and equipment, and relates to the technical field of noise reduction test. The method comprises the following steps: collecting mixed audio data flow of a target vehicle cockpit, analyzing the mixed audio data flow to obtain noise audio data; performing noise reduction processing on the noise audio data to obtain noise reduction audio data, and converting the noise reduction audio data into test audio data in a target playback format; determining a frequency band noise reduction comprehensive score and a defect distribution atlas of the test audio data based on the difference values of the test audio data and reference audio data on different target frequency bands; and generating a test result of the voice noise reduction function based on the frequency band noise reduction comprehensive score and the defect distribution atlas of the test audio data. In the foregoing manner, discrete test steps are integrated into a continuous automatic flow, differences caused by human operation are eliminated as much as possible, distortion in the transmission or conversion process is basically avoided, and the accuracy of the test result is improved.

Disclosed is speech noise reduction method including: acquiring first speech data, which is collected by means of a microphone, and acquiring second speech data, which is collected by means of a bone conduction sensor; and inputting speech data in a first frequency band of the first speech data and speech data in a second frequency band of the second speech data into a speech fusion noise reduction network and performing prediction, so as to obtain target noise reduced speech data, wherein the first frequency band is higher than the second frequency band, and the speech fusion noise reduction network is obtained by means of performing training in advance by performing training using noisy microphone speech data and noisy bone conduction speech data as input data, and using clean microphone speech data corresponding to the noisy microphone speech data as a training label.

The application provides a multi-modal fusion voice noise reduction method and device, electronic equipment and storage medium, comprising: collecting multi-modal signals and aligning the multi-modal signals to obtain synchronous data streams, wherein the multi-modal signals include AC microphone signals, BC microphone signals and IMU signals; performing air conduction difference noise reduction on two AC microphone signals to obtain AC difference synchronous signals; determining whether it is valid speech through a preset voice activity detection mechanism; in response to valid speech, performing fusion processing on the AC difference synchronous signals, BC synchronous signals and IMU synchronous signals based on a constructed neural network to generate a pure speech signal. Through the collection and fusion processing of multi-modal data, the application has stronger robustness than the traditional single-modal or AC+BC noise reduction method, can reduce non-speech interference, and is suitable for extreme noise environments with low signal-to-noise ratio.

The invention provides a multi-modal fusion voice noise reduction method and device, electronic equipment and a storage medium, and the method comprises the steps: collecting multi-modal signals, and carrying out the data alignment of the multi-modal signals, so as to obtain a synchronous data stream, the multi-modal signals comprising an AC microphone signal, a BC microphone signal and an IMU signal; the method comprises the following steps: performing air conduction differential noise reduction on two paths of AC microphone signals to obtain an AC differential synchronization signal; judging whether the voice is effective or not through a preset voice activity detection mechanism; and in response to the effective voice, performing fusion processing on the AC differential synchronization signal, the BC synchronization signal and the IMU synchronization signal based on the constructed neural network to generate a pure voice signal. Compared with a traditional single-mode or AC + BC noise reduction method, the method has the advantages that through collection and fusion processing of the multi-mode data, the robustness is higher, non-voice interference can be reduced, and the method is suitable for an extreme noise environment with a low signal-to-noise ratio.

The invention belongs to the technical field of artificial intelligence, and relates to a voice noise reduction and voice recognition joint optimization method under a multi-task learning framework, which comprises the following steps: acquiring an original voice signal flow containing environmental noise; integrating the analysis results of the two channels to generate joint context intelligence; generating a multi-dimensional processing strategy instruction set comprising a plurality of control dimensions for each time-frequency unit according to the joint context intelligence; driving the parameterization noise reduction processor to process the initial time-frequency domain characteristics so as to output customized noise-reduced characteristics; constructing an initial recognition hypothesis path containing a plurality of candidate recognition results and corresponding confidence coefficients based on the joint context information; and selecting the candidate recognition result with the highest confidence coefficient after re-scoring, and decoding the candidate recognition result into a final recognition text. The method solves the problems that the overall performance of the system is limited due to lack of information interaction and cooperation, a globally optimal solution cannot be formed, and variable and complex real noise scenes are difficult to deal with.

The invention discloses a real-time voice noise reduction method and system based on a deep neural network, and is applied to the technical field of voice noise reduction. The method comprises the following steps: obtaining a pure voice sample, and adding random noise to the pure voice sample to obtain a noisy voice sample; preprocessing the voice sample to obtain a training data set; constructing a generative adversarial neural network for real-time voice noise reduction, wherein the generative adversarial neural network comprises a generator and a discriminator; iteratively training a generator and a discriminator based on the training data set and evaluating network performance; a generator is deployed, a noisy voice signal is obtained in real time and preprocessed, the preprocessed noisy voice is input into the generator, and a noise-reduced voice signal is obtained. According to the invention, the real-time processing requirement is met while the voice noise reduction performance is ensured.

The application relates to a voice noise reduction method, device and storage medium, wherein the voice noise reduction method comprises the following steps: acquiring a voice signal, extracting a frequency domain component of the voice signal, and dividing the frequency domain component into a high-frequency component and a low-frequency component according to the frequency size, wherein the frequency of the high-frequency component is greater than that of the low-frequency component; acquiring a sound source distance of the voice signal; and performing noise reduction on the voice signal, and in the case that the sound source distance is not lower than a preset threshold, the noise reduction intensity of the high-frequency component is not higher than that of the low-frequency component, thereby solving the problem of voice distortion when the sound source distance is far, and improving the quality of the voice signal.