33 results about "Noise masking" patented technology

This disclosure provides methods, devices, and systems for audio signal processing. The present implementations more specifically relate to speech enhancement techniques for separating microphone signals into speech, echo, and noise signals. In some aspects, a speech enhancement system may include a delay estimator and an acoustic echo and noise (AEN) decoupling filter. The delay estimator receives a microphone signal via a microphone and a far-end audio signal for output via a speaker and estimates a reference audio signal based on a delay between the microphone signal and the far-end audio signal. In some aspects, the AEN decoupling filter may determine a speech mask, an echo mask, and a noise mask based on the microphone signal and the reference audio signal and may suppress an echo component and a noise component of the microphone signal based on the determined set of masks.

The embodiment of the invention provides an underwater acoustic signal noise reduction method and device, computer equipment and a readable storage medium. The method comprises the steps of obtaining a to-be-denoised time domain signal of a to-be-denoised underwater sound, and performing feature extraction on the to-be-denoised time domain signal through a target model to obtain a to-be-denoised feature tensor; calculating local position dependence features and local time dependence features between each local feature block and the adjacent feature block in parallel through a local department control circulation unit attention mechanism of the target model to obtain local noise features; calculating global noise features of the feature tensor to be subjected to noise reduction through a global gating circulation unit attention mechanism of the target model; reconstructing a noise time domain signal through a decoder of the target model based on a noise mask matrix generated by fusing the local noise features and the global noise features; and based on the noise time domain signal, performing noise filtering on the to-be-denoised time domain signal to obtain a target waveform signal. Therefore, the accuracy of the waveform signal obtained by reconstruction can be improved.

The invention discloses a mobile phone personalized sound effect adjusting method based on auricle shape recognition, which comprises the following steps: acquiring three-dimensional shape data of a user auricle through a mobile phone camera, and extracting key parameters such as contour feature points, concave depth and reflection angle of the auricle; calculating a propagation path, a reflection characteristic and a resonance characteristic of sound waves in auricles by using an acoustic modeling technology, and generating a personalized sound effect optimization model; calculating auricle sound insulation characteristics based on an acoustic modeling result, wherein the auricle sound insulation characteristics comprise external noise shielding capability and specific frequency band sound wave absorption capability; when the user answers the call, audio optimization parameters are dynamically generated in combination with the spectral characteristics of the surrounding noise, the frequency band characteristics of the call audio and the personalized sound effect optimization model; dynamically adjusting the audio output setting of the mobile phone based on the audio optimization parameters; according to the method, deep combination of the auricle form and audio optimization is realized, and the personalized auditory experience of the user is improved.

The application discloses a kind of underwater glider acoustic data processing method and storage medium based on Conformer, belong to underwater acoustic signal processing technical field.The method includes: using MCR-AAD model to the original acoustic data collected by glider Effective signal detection, output the probability of each frame as effective signal, and using double threshold post-processing method to determine effective signal and noise signal;The detected effective signal is input into the causal BGformer noise reduction model, generates and reverses noise mask by audio feature fusion, DCC encoder, Conformer decoder and label embedding module, to obtain the audio result after noise reduction.MCR-AAD model is composed of mel feature extractor, multi-scale depth separable convolution, dynamic position coding, Conformer encoder and multi-task classification module.The application constructs efficient, real-time underwater acoustic data processing scheme, effectively improves the application performance of glider in marine observation and target identification.

A sound noise-shielding device and a masking earphone are provided in this disclosure. Earphones, hearing aids and noise-masking earplugs are integrated. The noise-masking earplugs are used to directly block real sounds, so that for patients with misophonia, stress response caused by sounds received by the earphones can be reduced, and these sounds can be converted into other acceptable sounds by the bone conduction earphones. Sometimes there may be some mixed sounds at the same time, and the computer can't distinguish them because of their fusion. A separate network is used to extract fused audio separately, and a spectrogram of an individual sound is obtained correspondingly and detected. By training the network, a sound recognition model that can accurately detect sound information can be obtained.

The invention provides a method, a device and a system for audio signal processing. The present implementations more particularly relate to speech enhancement techniques for separating microphone signals into speech, echo, and noise signals. In some aspects, a speech enhancement system may include a delay estimator and an acoustic echo and noise (AEN) decoupling filter. The delay estimator receives a microphone signal via the microphone and receives a far-end audio signal for output via the speaker, and estimates a reference audio signal based on a delay between the microphone signal and the far-end audio signal. In some aspects, an AEN decoupling filter may determine a speech mask, an echo mask, and a noise mask based on a microphone signal and a reference audio signal, and may suppress echo and noise components of the microphone signal based on the determined set of masks.

This invention belongs to the field of communication noise detection and processing technology, and relates to a method and system for testing low-noise power supplies using waveforms. The method connects the power supply under test (DUT) and a voice gateway, and powers them on in real time. The voice gateway is connected to an external noise-shielded environment via its voice interface. Within this environment, the voice gateway is modulated to a periodic state of comfortable noise or a dial tone, and the sound is played. The playing sound is acquired and processed, and based on the processed sound and the sound signal waveform, the noise level of the DUT is detected in real time. This transforms uncontrollable human-induced noise into an electrical signal measurable by an oscilloscope, accurately determining whether the power supply meets requirements. In other words, by utilizing a clean environment, the unmeasurable power frequency noise on the telephone line is converted into a differential-mode signal measurable by an oscilloscope through a physical sound isolation process, providing testers with an objective experimental result.

The invention provides an acoustic characteristic determination method and device, electronic equipment, a system and a storage medium, and relates to the technical field of signal processing. The method comprises the following steps: controlling to continuously input an excitation signal to a detected sounding body, wherein the excitation signal is obtained by splicing an identifier signal different from a main signal in front of the main signal; acquiring a sound response signal which is output by the detected sounding body and aims at the excitation signal; performing cross-correlation analysis on the sound response signal based on the identifier signal to obtain an effective signal corresponding to the main signal contained in the sound response signal; and performing frequency domain analysis on the effective signal to obtain the acoustic characteristics of the detected sounding body. The correlation between the acoustic response signal and the excitation signal is enhanced through the identifier signal, cross-correlation errors caused by time migration, high noise covering and the like are reduced, the accuracy of effective signals is improved, and then the accuracy of acoustic characteristics is improved. In addition, the positions of the identifier signal and the effective signal in the sound response signal are quickly positioned, the algorithm complexity is low, and the signal processing efficiency is high.

The application discloses a kind of automobile air conditioner control method, electronic equipment and storage medium.The automobile air conditioner control method includes: in response to the start request of air conditioner, control condenser cooling fan earlier than or equal to electric compressor start;In response to the stop request of air conditioner, control condenser cooling fan later than or equal to electric compressor stop.The present application is based on the noise masking effect of human hearing, using condenser cooling fan noise as the incremental source of air conditioning system working background noise, by staggering the start-stop time interval of cooling fan and electric compressor, let the cooling fan start in advance, delay closing, increase the background noise in the process of electric compressor start-stop, solve the abrupt feeling generated by noise mutation in the process of electric compressor start-stop.

A device for recognizing a multi-channel input voice includes a time-frequency transformer that receives channel audio signals extracted from voice data recorded through microphones having unspecified microphone array forms and transforms the channel audio signals into time-frequency domain signals, a speaker and noise mask estimator that receives the time-frequency domain signals and estimates a time-frequency domain mask for voices and noise for speakers, a beamformer estimator that estimates a time-frequency domain signal for voice signals of the speakers from which the noise has been removed from the time-frequency domain signals by using the time-frequency domain mask, a time-frequency inverse transformer that inversely transforms the time-frequency domain signal from which the noise has been removed into a time domain signal, and a learning machine that trains the speaker and noise mask estimator based on a loss function obtained by comparing the inversely-transformed time domain signal and a pre-defined answer signal.

Apparatuses, systems, and techniques to generate blue noise masks for real-time image rendering and enhancement. In at least one embodiment, a noise mask is generated and applied to one or more images to generate one or more enhanced images for image processing (e.g., real-time image rendering). In at least one embodiment, the noise mask is able to handle the temporal domain (e.g., add time to the spatial domain) to improve image quality when rendering images over multiple frames.

The invention relates to a low-speed warning tone generation method and device of an electric vehicle, the vehicle and a storage medium, and the method comprises the steps that the sound style of the low-speed warning tone of the electric vehicle is determined; the whole vehicle frequency response of the electric vehicle and the pedestrian warning indicator frequency response are tested and analyzed, and the main frequency band of the low-speed warning tone is determined; on the basis of the main frequency band, respectively recording in-vehicle noise of the electric vehicle in a low-speed pedestrian warning tone road test and a low-speed pedestrian warning tone-free road test under the condition that the electric vehicle runs at a preset low speed to obtain a frequency range and noise masking data of the in-vehicle road noise, and outputting the in-vehicle road noise after the frequency range and the noise masking data meet preset conditions. And generating a low-speed prompt tone. Therefore, the problems that in the related technology, warning tone generated by an external loudspeaker is transmitted into a cabin through structural sound transmission and air sound transmission ways, so that the sound quality in a vehicle is poor, and the sound pressure level in the vehicle is increased once a self-adaptive control system is used for improving the warning loudness to compensate the external environment noise, and the sound quality is poor are solved. The driving comfort is influenced, and the like.

Techniques for enhancing an audio signal are provided herein. In some embodiments, a method for enhancing an audio segment may include receiving an input audio segment to be enhanced. The method may also include generating a feature indicative of an audio content type of the input audio segment. The method may also include generating a feature indicative of the presence of noise in the input audio segment by providing at least the feature indicative of the audio content type to the trained de-noising feature extraction model. The method may also include generating a de-noising mask based on a feature indicating the presence of noise in the input audio segment. The method may also include generating an enhanced audio segment at least by applying a de-noising mask to the input audio segment.

A vehicle noise masking device includes a database configured to store sound sources, a sound source classification unit configured to classify the sound sources into a first sound source group and a second sound source group, a communication unit, a first processing unit configured to analyze components of the noise measurement signals based on the traveling information and classify the noise measurement signals into vehicle-related noise and vehicle-unrelated noise, a second processing unit configured to extract a first masking sound source similar to the vehicle-related noise from the first sound source group, a third processing unit configured to extract a second masking sound source similar to the vehicle-unrelated noise from the second sound source group, and a fourth processing unit configured to generate a third masking sound source using at least one of the first masking sound source and the second masking sound source.

The present invention discloses a method and system for masking noise. The method for masking noise includes the following steps: training an intelligent neural network model with a variety of "ideal noises" that can soothe people's body and mind; using the trained intelligent neural network model to identify the actual noise to be masked and form a parameter matching table, the parameter matching table includes noise types, time scaling parameters, and slope transformation parameters, and at the same time obtaining the amplitude ratio of the actual noise; according to the parameter matching table, combined with the user's personalized settings, obtaining the ideal noise type, time scaling parameter, and slope transformation parameter with the best masking effect; generating the best personalized noise masking sound according to the best-matched ideal noise type, time scaling parameter, slope transformation parameter, and the amplitude ratio of the actual noise.

A vehicle noise masking apparatus and method are disclosed. The vehicle noise masking apparatus includes: a database configured to store a sound source; a sound source classification unit configured to classify sound sources into a first sound source group and a second sound source group; a communication unit; a first processing unit configured to analyze a composition of the noise measurement signal based on the driving information and classify the noise measurement signal into vehicle-related noise and vehicle-independent noise, a second processing unit configured to extract a first masking sound source similar to the vehicle-related noise from the first sound source group, a third processing unit configured to extract a second masking sound source similar to the vehicle-independent noise from the second sound source group, and a fourth processing unit configured to extract a second masked sound source similar to the vehicle-independent noise from the second sound source group, and a fourth processing unit configured to generate a third masked sound source using at least one of the first masked sound source and the second masked sound source.

This invention discloses a method and system for denoising magnetotelluric signals based on a self-supervised diffusion model, belonging to the field of magnetotelluric technology. The invention designs the TimeDART diffusion model as a two-stage architecture for constructing a noise classifier model and a diffusion denoising model. The initial input magnetotelluric data first passes through a noise classifier to generate a noise mask to identify target noise regions. Subsequently, the diffusion denoising model based on the diffusion model processes only these noise regions, generating denoised signal segments. Finally, the denoised noise regions are merged with the clean regions of the initial magnetotelluric data. By combining time-frequency analysis and VMD techniques to extract low-frequency trends from the magnetotelluric data, precise noise region localization, differentiated denoising, and smooth fusion can be achieved. While efficiently suppressing noise, it retains the effective signal characteristics to the greatest extent, improving the automation level of noise suppression, signal-to-noise identification accuracy, and denoising accuracy.

The present invention relates to the technical field of speaker recognition, and discloses a speaker recognition method and system for real scenarios, including: acquiring external sound information, processing it through multi-resolution convolutional kernels, extracting features in the time domain and performing feature fusion only in the time domain; after feature fusion in the time domain, using grouped convolution to extract features in the frequency domain, performing processing and feature fusion only in the frequency domain; after feature fusion in the frequency domain, performing masking processing, and finally outputting the recognition result. Through the noise masking module and multi-resolution feature extraction, the extraction of target speaker features is enhanced, and the influence of noise on the recognition result is reduced. At the same time, in a complex noise environment, the system can effectively focus on the target speaker, improving the stability and reliability of recognition. Through multi-scale feature extraction in the time domain and frequency domain, more comprehensive feature information is provided, which helps to capture the subtle differences of speakers.

The described aspects and implementations enable privacy-respecting detection, separation, and localization of sounds in vehicle environments. The techniques include obtaining, using audio detector(s) of a vehicle, a sound recording that includes a plurality of elemental sounds (ESs) in a driving environment of the vehicle, and processing, using a sound separation model, the sound recording to separate individual ESs of the plurality of ESs. The techniques further include identifying a content of individual ESs and causing a driving path of the vehicle to be modified in view of the identified content of the individual ESs. Further techniques include rendering speech imperceptibly by redacting temporal portions of the speech, using sound recognition models to identify and discard recordings of speech, and driving at speeds that exceed threshold speeds at which speech becomes imperceptible from noise masking.

The invention relates to a method and device for improving the dynamic range of analog-to-digital conversion input signals. The method for improving the dynamic range of the analog-to-digital conversion input signal comprises the following steps: dividing an analog audio signal to respectively obtain an amplified analog audio signal and a reduced analog audio signal, and respectively carrying out analog-to-digital conversion on the amplified analog audio signal and the reduced analog audio signal to respectively obtain an amplified digital signal and a reduced digital signal; judging whether the instantaneous amplitude of the amplified digital signal meets a full-scale threshold value or not; if not, adopting the amplified digital signal; if yes, gain compensation is carried out on the amplitude-reduced digital signal, and the digital signal after gain compensation is adopted. According to the method for improving the dynamic range of the analog-to-digital conversion input signal, the input dynamic range of the analog-to-digital converter is effectively expanded, the problems of large signal topping distortion and small signal noise masking are solved, and therefore the overall performance and the tone quality performance in the audio collection process are remarkably improved.

The invention relates to the technical field of music data processing, in particular to a violin playing auxiliary system based on artificial intelligence, and the system comprises an audio collection module which employs a microphone array to capture a violin playing sound field, and generates an original audio stream. A coding adaptation module generates a transmission code stream while extracting side information metadata including a noise masking threshold and a transient signal flag. The feature extraction module outputs a high-dimensional feature vector. The reconstruction module generates reconstructed audio. And the evaluation feedback module performs quantitative evaluation on the reconstructed audio, generates scores of intonation deviation degree, string rubbing depth and overtone purity, feeds the scores back to the coding adaptation module, and dynamically adjusts frequency weighting parameters of the psychological acoustic model. The music interface module outputs reconstructed audio and structured evaluation data. The system drives the coding adaptation module to adaptively optimize a compression strategy through score feedback of the evaluation feedback module, and the reconstruction module optimizes a generation process by using historical evaluation data to form a closed-loop data stream.

The application discloses a speech enhancement method, device and equipment. The method estimates the noise mask of the complex spectrum of the noisy speech by a noise mask prediction model based on complex value calculation. The model uses complex operation to better model the correlation between the amplitude spectrum and the phase spectrum by using prior knowledge, avoids treating the real part and the imaginary part as two unrelated parts for operation respectively like a real value network, generates enhanced speech data of the noisy speech data according to the masking value of the complex spectrum and the complex spectrum of the noisy speech, so that the amplitude spectrum and the phase spectrum can be enhanced at the same time. Therefore, the hearing quality of the enhanced speech can be effectively improved.

According to the control method of the vehicle electronic expansion valve, the expansion valve working information of each startup sound effect mode is determined according to the sound effect sound information of each startup sound effect mode, and then the startup sound effect duration information is determined according to the expansion valve working information of each startup sound effect mode; the method comprises the following steps: responding to a vehicle power-on signal and acquiring a current startup sound effect mode, then controlling an in-vehicle loudspeaker according to startup sound effect duration information and sound effect sound information of the current startup sound effect mode, finally controlling an electronic expansion valve according to expansion valve working information of the current startup sound effect mode, and controlling the electronic expansion valve based on a noise masking effect. And noise caused by power-on reset of the electronic expansion valve is covered by using a power-on sound effect generated by power-on of the vehicle.

The invention discloses a hearing aid fitting method for matching noise reduction and personalized auditory filtering characteristics. The method comprises the following steps of: firstly, measuring an individual hearing filtering characteristic (such as frequency resolution) through a noise masking experiment method and the like so as to refine and simulate an individual hearing contour; and then, combining the audio state subjected to noise reduction processing with the simulated individual auditory contour to predict the actual auditory experience of the user so as to evaluate the hearing aid effect. And finally, based on a predicted auditory experience result, utilizing a reinforcement learning algorithm to drive a strategy network, and cooperatively optimizing noise reduction parameters and hearing aid gain configuration. The process aims to finally improve the auditory experience of the user by continuously adjusting the two types of parameters. Through denoising-gain collaborative optimization, the speech recognition rate and hearing comfort can be improved in a complex noise scene, and the method has higher adaptation precision, higher environmental robustness and wide clinical application value.

A system configured to improve audio processing by performing dereverberation, noise reduction, and residual echo suppression during a communication session. The system may include a deep neural network (DNN) configured to jointly mitigate additive noise, reverberation, and residual echo. The DNN may be a convolutional recurrent network with dense connectivity (CRN-DC) and may be configured to process complex-valued spectrograms corresponding to the isolated audio data and / or estimated echo data generated by during echo cancellation. The DNN may generate a speech mask and / or an ambient noise mask, enabling the device to generate output audio data representing target speech and a variable amount of ambient noise. For example, the device may separately reconstruct the target speech using the speech mask and the background noise using the ambient noise mask, which enables the device to control the amount of ambient noise represented in the output audio data.

The invention provides a feasible method for generating a dance video based on music and lyrics, which adopts a time synchronization mechanism fusing lyric semantic analysis and music beat analysis to ensure that dance actions are aligned with lyric keywords and music rhythms; advanced network structures such as U-Net and Reference Net are introduced, so that the quality and expressive force of the generated dance motion sequence are further improved. The U-Net network structure is used for finely adjusting and optimizing the generated dance actions, and the coherence and naturalness of the actions are ensured by utilizing the strong feature fusion and context sensing capabilities of the U-Net network structure. And on the other hand, the Reference Net is used for introducing reference information, and the reference information is compared with a predefined dance movement template or example for reference, so that the generated dance movement better meets the expected and artistic requirements. In addition, skills such as a noise mask and an initial image are combined, so that the generation effect of the dance video is further enhanced. The noise mask technology can help the model to better process uncertainty and randomness, and the diversity and innovativeness of generation actions are improved. The introduction of the initial image provides visual guidance and constraint for the generation of the dance video, so that the generated video better conforms to the expectation and aesthetic standard of the user.

The invention discloses an underwater glider acoustic data processing method based on Conformer and a storage medium, and belongs to the technical field of underwater acoustic signal processing. The method comprises the following steps: performing effective signal detection on original acoustic data acquired by a glider by utilizing an MCR-AAD model, outputting the probability that each frame is an effective signal, and judging the effective signal and a noise signal by adopting a dual-threshold post-processing method; and inputting the detected effective signal into a BGform noise reduction model with causality, and generating and inverting a noise mask through audio feature fusion, a DCC encoder, a Conform decoder and a label embedding module to obtain a noise-reduced audio result. The MCR-AAD model is composed of a Mel feature extractor, a multi-scale depth separable convolution, a dynamic position code, a Conformer encoder and a multi-task classification module. According to the method, an efficient and real-time underwater acoustic data processing scheme is constructed, and the application performance of the glider in ocean observation and target recognition is effectively improved.

The invention discloses a target voice extraction method and system and a medium, and belongs to the technical field of multi-mode voice signal processing. Comprising the following steps: giving a mixed voice and a lip video of a target speaker, and extracting a voice logarithmic power spectrum, a cross-channel phase difference and a visual time sequence feature; splicing multi-modal features, inputting the spliced multi-modal features into an improved DPCRN network, estimating a voice mask and deducing a noise mask; calculating a beam forming weight through the covariance matrix and generalized eigenvalue decomposition to obtain beam features of voice and noise; the logarithmic power spectrum, the visual features and the beam features are fused, and high-dimensional representation of voice and noise is constructed; feature mutual exclusion enhancement is realized by using a frame-level cross attention mechanism, residual noise is removed from voice features, and leaked voice is removed from noise features; and finally, an optimization mask is generated through a decoder, and pure target voice is output through spectrum reconstruction. According to the invention, complex noise environment target voice can be accurately separated.

This disclosure relates to spatiotemporal noise masks for image processing, and more particularly to apparatus, systems, and techniques for generating blue noise masks for real-time image rendering and enhancement. In at least one embodiment, a noise mask is generated and applied to one or more images to generate one or more enhanced images for image processing (e.g., real-time image rendering). In at least one embodiment, when images are rendered over multiple frames, the noise mask is capable of processing the temporal domain (e.g., adding time to the spatial domain) to improve image quality.

The described aspects and implementations enable privacy-respecting detection, separation, and localization of sounds in vehicle environments. The techniques include obtaining, using audio detector(s) of a vehicle, a sound recording that includes a plurality of elemental sounds (ESs) in a driving environment of the vehicle, and processing, using a sound separation model, the sound recording to separate individual ESs of the plurality of ESs. The techniques further include identifying a content of individual ESs and causing a driving path of the vehicle to be modified in view of the identified content of the individual ESs. Further techniques include rendering speech imperceptibly by redacting temporal portions of the speech, using sound recognition models to identify and discard recordings of speech, and driving at speeds that exceed threshold speeds at which speech becomes imperceptible from noise masking.