64 results about "Vocal tract" patented technology

The invention relates to the technical field of speech recognition, in particular to a Chinese learner-oriented tone evaluation and improvement method, which comprises the following steps of: extracting a fundamental frequency F0 curve and sound channel parameters of a speech signal, constructing a turbid and clear adaptive fusion model, and fusing to generate fundamental frequency related characteristics and fundamental frequency unrelated characteristics; constructing a tone error corpus, carrying out tone classification labeling, generating a multi-level tone feature set through a turbid and clear adaptive fusion model, aligning a voice signal with a reference voice time domain, and extracting and decoupling a tone shape feature vector and a tone domain feature vector; and based on the decoupled feature vectors, establishing a dual-channel evaluation path, hierarchically calculating the tone type distance and the tone domain distance of the tones, carrying out weighted fusion, generating a final evaluation score, and finally generating feedback information for the Chinese learner. According to the method, a complete method from feature decoupling to two-channel evaluation is constructed, the pronunciation problem of the learner is quantitatively diagnosed, and the pertinence and efficiency of Chinese tone learning are improved.

The invention discloses a voice authentic identification method based on dual-track difference modeling, and belongs to the technical field of voice authentic identification. The method comprises the following steps: converting monaural audio into stereophonic sound through a pre-trained monaural and dual-channel voice conversion model, extracting left and right channel Mel spectrum features, and performing texture enhancement optimization processing on an absolute value of a difference between the left and right channel Mel spectrum features and an original monaural Mel spectrum feature. The method comprises the following steps: firstly, optimizing the voice authenticity, respectively inputting an optimization result into a left double-branch feature extractor and a right double-branch feature extractor as input of double-branch feature extractors, carrying out extraction, processing and information fusion on feature information to obtain a final attention map, and inputting the final attention map into an attention pooling layer and a final dichotomy layer to obtain a voice authenticity discrimination result. According to the invention, two-track differential modeling, a fine-grained texture enhancement method and a multi-head attention feature fusion method are adopted to carry out authentic identification on the voice, so that the method has higher accuracy, mobility and generalization.

The invention discloses a finite element identity recognition system based on voiceprint comparison. The finite element identity recognition system comprises a voiceprint extraction module, a sound channel finite element modal analysis module, a front balance processing module, a double-unit construction module, a pseudo measurement constraint module, an excitation unit parameter adaptive selection module, a representation unit correction module and an identity recognition module. The invention belongs to the field of voice identity recognition, and particularly relates to a finite element identity recognition system based on voiceprint comparison. According to the scheme, real glottis excitation and playing simulation signals are distinguished through finite element combination modality and double-unit joint estimation, and the detection capacity of anti-recording and anti-synthesis attacks is enhanced; pseudo measurement constraints are introduced, redundant excitation components are eliminated, and the distinguishing degree of different speakers is improved; through the correction of the characterization unit, the real-time self-adaptive capability is provided for the environment sudden change aiming at the non-stable speaking and environment, dynamic scaling measurement noise and process noise, the performance fluctuation caused by assumed mismatch is reduced, and the identity recognition effect is further improved.

The invention discloses a method and a system for constructing surface and subsurface sound channels influenced by ocean mesoscale vortexes. The method comprises the following steps: acquiring hydrological parameters of a specified area within specified time; calculating an Argo sound velocity profile and a sound velocity background field of a specified area, and performing quality control to obtain an Argo sound velocity profile data set; extracting the characteristics of the mesoscale vortex, and performing space-time matching and merging on the Argo sound velocity profile data set and the mesoscale vortex trajectory data set to generate a mesoscale vortex disturbance sound velocity field; after a mesoscale vortex underwater sound field and a sound velocity background field are interpolated respectively, the sound velocity background field is separated, a mesoscale vortex underwater sound velocity disturbance structure is extracted, and sound propagation loss abnormal data under the influence of mesoscale vortexes are obtained by combining a Bellhop acoustic toolbox to extract the characteristics of surface and subsurface sound channels. The invention aims to obtain the influence of the mesoscale vortex on the surface sound channel and accurately obtain the variability characteristics of the surface and subsurface sound channels corresponding to the mesoscale vortex.

An acoustic face mask reduces the distortion and muffling of speech sounds by a face mask wall. The distortion and muffling of speech sounds by a face mask wall may be reduced by reducing the acoustic coupling of the vocal tract and / or nasal cavity to the face mask chamber, which causes a reduction in the intelligibility of the speech. The acoustic coupling of the vocal tract and / or nasal cavity to the face mask chamber may be reduced, for example, by reducing the reflected sound energy caused by the face mask wall. For example, one or more sound absorbing members or acoustically transparent members reduce the reflected sound energy caused by the face mask wall.

The invention relates to the field of trachinotus ovatus swarm sound production identification, and discloses a method for identifying sound production of a trachinotus ovatus swarm in a net cage based on dual-channel time-varying Wiener filtering. According to the method, modeling and target signal counteracting are respectively carried out according to the conditions of the left sound channel and the right sound channel of the input signal so as to analyze the noise signal-to-noise ratio characteristics, enhancement of target biological sound and suppression of noise signals are realized in combination with a Wiener filtering method, trachinotus ovatus sound production can be identified in a complex noise environment, and the method has the target positioning capability. According to the method, by introducing a dual-channel signal equalization and interference compensation mechanism and fusing posterior and prior signal-to-noise ratio estimation, the sound production of the trachinotus ovatus swarm can be effectively identified under the complex noise condition without training data, the problems that the performance is poor and a large amount of training data is needed when non-stationary noise is separated by a traditional noise separation method are solved, and the method is suitable for popularization and application. Technical prerequisites are provided for net cage acoustic monitoring work, and possibility is created for rapid, efficient and accurate sound production analysis of subsequent net cage fish schools.

The invention discloses a solar sound-driven intelligent bird repeller which comprises a mounting mechanism, a buffer structure, a bird repeller body, bird repeller sound channels, a solar panel structure and a shielding structure, the buffer structure is located at the top end of the mounting mechanism, the bird repeller body is located at the top end of the buffer structure, the number of the bird repeller sound channels is four, and the number of the solar panel structure is four. The four bird repelling device sound channels are located at the four ends of the bird repelling device body respectively, the solar cell panel structure is located in the center of the top end of the bird repelling device body, and the shielding structure is located at the end of the top end of the bird repelling device body. The solar bird repelling device has the advantages that the device can be conveniently installed, vibration can be buffered, power can be generated through solar energy, and birds can be accurately repelled.

The present application discloses a method for extracting multiple types of sound velocity profile jump layers, belonging to the field of marine survey. Aiming at the problems that the extraction range is too deep, the spatial distribution is too scattered, and the extraction cannot be accurate in the case of multiple sound velocity jump layers when the traditional optimal segmentation algorithm is used to extract the sound velocity jump layer, the present application simplifies the sound velocity profile into a four-layer structure on the basis of the traditional optimal segmentation method; the accurate identification of the double jump layer is realized according to the principle of the number of sound channel axes that meet the gradient threshold; the upper limit of the sound velocity jump layer is obtained by repeated iterative calculations with the gradient mutation point as the starting point; in the case of a single sound channel axis, the optimal segmentation algorithm is used to determine the lower limit of the jump layer; in the case of multiple sound channel axes, the optimal three-segmentation method is used to determine the upper and lower limits of the jump layer. The method of the present application can effectively and accurately identify the multi-jump layer structure, and can also realize the feature extraction of multiple types of sound velocity profile jump layers, which is helpful to improve the reliability of underwater communication and the concealment performance of submarines.

The invention aims to provide the semi-closed sound channel training device for high pitch training, which is applied to semi-closed sound channel high pitch range training and can adjust the training airflow resistance, and the use method of the semi-closed sound channel training device. The sealing device comprises a sealing cover connected with the bottle body and a gas inlet pipe penetrating through the sealing cover and used for gas interaction. The air inlet end of the air inlet pipe is connected with a flow guide cover used for being tightly attached to the nose of a user, and an air flow cavity used for guiding air flow is formed in the flow guide cover. The flow guide cover covering the face of a user is adopted, the nasal cavity is communicated with the closed space in the bottle body through the air inlet pipe and the communicating pipe, the user can form airflow resistance by closing the mouth, using the nose to discharge air and produce sound and limiting the air flowing speed through the air pressure adjusting piece, and then the purpose of high-pitch semi-closed sound track training is achieved. The sound channel training device is applied to the technical field of sound channel training devices.

The invention discloses a digital human generation system and method based on a language model, and relates to the technical field of artificial intelligence. A complete digital human pronunciation modeling path from voice signal acquisition, formant frequency extraction, deviation analysis and pronunciation stability evaluation to motion compensation control and three-dimensional animation fusion is realized. Compared with the existing mode of driving the digital human pattern only based on energy envelope or mouth shape classification, the method has the advantages that the multi-frame formant frequency of the user voice is extracted and processed, so that the digital human can generate dynamic lingual surface and tongue tip actions according to the sound channel change in the real pronunciation process of the user; therefore, the physical correspondence and linguistic consistency of the pronunciation actions of the digital human are improved. According to the method, the direct mapping between the pronunciation characteristics and the three-dimensional tongue actions is realized, and the authenticity of the pronunciation animation of the digital human and the guidance in a teaching scene are remarkably improved.

The invention relates to the technical field of underwater acoustic communication, in particular to an underwater acoustic channel propagation loss distribution prediction method and device, and the method comprises the steps: obtaining a first sound source depth and a target sound source depth which are conjugated with respect to a sound channel axis in a water area based on the sound velocity profile of the water area; based on a resolution demand, determining underwater acoustic channel propagation loss distribution of the first sound source depth, and determining free space propagation loss distribution of the target sound source depth; carrying out amplitude limiting and normalization processing on the two, and combining on a channel dimension; and inputting the merged propagation loss distribution into a target neural network to output a normalized propagation loss distribution prediction result of the target sound source depth so as to determine the underwater acoustic channel propagation loss distribution of the target sound source depth. According to the method, the known propagation loss distribution under any sound source depth can be utilized to predict the propagation loss distribution of the sound source depth which is conjugated with the sound source depth about the sound channel axis, and the practical application capability and accuracy of the target neural network are also extremely high.

The invention discloses a time-difference method ultrasonic flowmeter installation detection method and equipment, and belongs to the technical field of flow detection, and the method comprises the following steps: obtaining uncertainty component information in flow measurement data, the uncertainty component information comprising a sound velocity time difference component, a linear flow velocity component, a sound path angle component and a water passing area component; a nominal precision table is created, the nominal precision table comprises submerged sound channel number, section shapes and flowmeter instantaneous flow errors, each sound channel number is associated with all the section shapes, and each sound channel number and section shape combination is associated with one flowmeter instantaneous flow error; before the ultrasonic flowmeter is installed and constructed, the water flow state in the section at the to-be-installed position needs to be analyzed, section type information needs to be obtained, the design and construction scheme is improved according to the nominal precision table, and the method has the effects that the metering precision of the flowmeter can be rapidly and accurately judged, and how to change the shape of the measured section to improve the metering precision can be known.

Techniques for processing measurement data streams for a person's speech apparatus are disclosed. The measurement data streams can be recorded by different sensor modalities, such as audio recordings or articulatory measurements, for example, which have an observable which quantifies a characteristic property of the vocal tract. The measurement data streams can be processed by means of a machine-learned model in order to obtain a temporal sequence of feature vectors in an articulatory feature space. These specific feature vectors enable a variety of applications, such as articulation training, therapeutic applications in the event of speech restrictions, for instance after a stroke, vocal training or speech synthesis.

The invention discloses a nasal sound recognition method based on acoustic feature decoupling, which is characterized in that a double-flow decoupling deep neural network is constructed, a mutual information minimization adversarial training strategy is introduced, and sound source features (vocal cord vibration) and sound channel features (oral cavity and nasal cavity adjustment) are forcibly separated from a single audio signal. On the basis, the high-resolution characteristic of CQT is used for accurately extracting acoustic characteristics related to the nasal sound, physiological offset correction is carried out in combination with personal acoustic fingerprints, and finally accurate quantification and recognition of different nasal sound states (normal, defect and skill) are achieved, so that an objective and visual feedback basis with artistic style adaptability is provided for vocal music teaching.

A kit of tools to aid in learning the sounds of the French language, designed to be applied to the vocal tract of a human subject, each tool comprising a functional part arranged at the end of a handle and shaped to make contact with a given place of articulation from among a plurality of predetermined places of articulation within the vocal tract and to constrain one or more organs of said vocal tract in a determined spatial configuration, each of said tools being associated with a set of consonants. See Figure 3

Techniques for sanitizing personally identifiable information (PII) from audio and visual data are provided. For instance, in a scenario where the data comprises an audio signal with speech uttered by a person P, these techniques can include removing / obfuscating / transforming speech-related PII in the audio signal such as pitch and acoustic cues associated with P's vocal tract shape and / or vocal actuators (e.g., lips, nasal air bypass, teeth, tongue, etc.) while allowing the content of the speech to remain recognizable. Further, in a scenario where the data comprises a still image or video in which a person P appears, these techniques can include removing / obfuscating / transforming visual PII in the image or video such as P's biological features and indicators of P's location / belongings / data while allowing the general nature of the image or video to remain discernable. Through this PII sanitization process, the privacy of individuals portrayed in the audio or visual data can be preserved.

Techniques for characterizing a vocal tract by using radar measurements. A radar marker that is arranged in or on the vocal tract is used.

This invention provides a tolerance-based distance estimation method and system for dual-channel acoustic imaging in Arctic ice regions based on normal mode-constrained modes. The method includes: using a normal mode model to simulate and predict the frequency domain sound pressure, phase velocity, group velocity, and mode covariance matrix of the experimental sea area in the Arctic ice region; using Fast Fourier Transform (FFT) to transform the time-domain signal received by the vertical array in the ice region into frequency-domain data; obtaining the sound velocity and depth corresponding to the upper boundary, channel axis, and lower boundary of the Beaufort waveguide based on the dual-channel sound velocity profile of the experimental sea area; obtaining the order index of the Beaufort waveguide modes based on the intersection of the phase velocity dispersion curve with the sound velocity at the upper boundary and channel axis of the Beaufort waveguide; performing mode separation using a matched filtering mode filtering method to extract the modes confined in the Beaufort waveguide; and constructing a tolerance-based distance estimation operator for ice and seabed based on the extracted modes in the Beaufort waveguide to achieve tolerance-based estimation of underwater target distances.

The invention relates to a multichannel ultrasonic flowmeter measuring method and application based on pressure correction, and the measuring method comprises the steps: uniformly and annularly arranging a plurality of ultrasonic transducers on the same circumferential section, and collecting signal data of a plurality of groups of sound channels; two pressure sensors are arranged at the front position and the rear position of the circumferential section respectively, and the pressure difference and the average pressure are calculated; correcting the instantaneous flow velocity of each group of sound channels based on the average pressure; determining a flow field model according to the pressure difference, and calculating the weighted average flow velocity of the plurality of groups of sound channels according to the instantaneous flow velocity; correcting a cross-sectional area of the circumferential cross-section based on the average pressure; and calculating the medium flow according to the corrected sectional area of the circumferential section and the weighted average flow velocity. The defect that an ultrasonic flowmeter is not high in measurement precision during measurement of a large-pipe-diameter and high-pressure pipeline is overcome, the measurement accuracy can be remarkably improved, the influence of errors of a single sensor or local flow field abnormity on the overall result is reduced, and stability is good.

An apparatus for encoding an audio signal having a plurality of channels is provided. The apparatus comprises a downmixer (1010) for down mixing the plurality of channels to obtain a downmix signal. Moreover, the apparatus comprises a residual signal calculator (1020) adapted for calculating a residual signal. Furthermore, the apparatus comprises a phase information calculator (1030) adapted for calculating information on a phase difference between the downmix and the residual signal to obtain phase information. Moreover, the apparatus comprises an output generator (1040) for outputting the phase information.

The invention discloses a touch screen voice wake-up system based on voiceprint recognition, which relates to the technical field of voice data processing and comprises an intelligent control terminal used for controlling data transmission and information interaction among modules. According to the method, various calculations are carried out on the wake-up voice data input by the user in advance, the sound channel shape, the acoustic structure feature and the feature vector independently belonging to the wake-up voice data of the user are determined, after the touch screen collects the surrounding voice data subsequently, feature emphasis is carried out on the first to-be-analyzed voice data so as to improve the feature intensity, and finally, the first to-be-analyzed voice data is analyzed. And verifying the sound channel shape, the acoustic structure feature and the feature vector of the to-be-verified voice through the sound channel shape, the acoustic structure feature and the feature vector of the wake-up voice data, determining whether the first to-be-analyzed voice data is sent by a user recording the wake-up voice data, if so, waking up the touch screen, and if not, not waking up the touch screen. And the privacy of the user is prevented from being leaked.

The application provides a voice detection method, comprising: obtaining a target voice, preprocessing the target voice, the preprocessing comprising pre-emphasis, framing and windowing; determining a first vocal tract feature, a first sound source wave feature and a plurality of first correlation features of the preprocessed target voice; determining the first principal component feature based on the first vocal tract feature, the first sound source wave feature and the plurality of first correlation features; inputting the first principal component feature into a trained classifier, and outputting a classification result, the classification result being a fake voice or a natural voice. The application utilizes the trace information left by the fake voice at the fundamental frequency, and utilizes the difference between the fake voice and the natural voice in the sound source and the vocal tract feature to realize fake voice detection. The method of principal component analysis is used to screen the sound source and the vocal tract feature respectively, the principal component with high correlation is selected as the feature, the feature dimension and the redundant feature are reduced, and the generalization ability and the efficiency of the model are improved.

An SOVT (semi-occluded vocal tract) exercise device is disclosed, consisting of a compact tubular body with an internal nonlinear airflow path that significantly extends the effective path length of exhaled air. In preferred embodiments a helical partition or insert inside the tube creates a tortuous channel, providing enhanced back-pressure and inertance for vocal training. The device further includes one or more adjustable vent apertures along the tube's wall, which can be selectively opened or closed (for example, by finger, plug, valve, or sleeve) to vary the resistance and / or effective tube length in real time, Certain embodiments incorporate a deformable or one-way valve element at the distal end of the channel that flexes or oscillates under airflow, introducing a pulsating resistance for additional therapeutic benefit. The modular design facilitates easy disassembly for cleaning, interchangeability of inserts or end elements, and portability for voice therapy, respiratory training, or musical applications.

The invention discloses a hearing aid. The hearing aid (2), in particular a behind-the-ear (BTE) hearing aid (2), comprises a base part (4) and an ear hook (6) which can be attached to the base part (4). In the assembled state, the base part (4) and the ear hook (6) are connected via a nozzle (14) and a socket (16), a sound channel (32) runs through the nozzle (14), the nozzle (14) has an outer peripheral wall (24), and the socket (16) has an inner peripheral wall (26), and, in the assembled state, a circumferential sealing ring (30) with a thickness (d) is arranged between the outer wall (24) and the inner wall (26). The sealing ring (30) is in particular located in an annular groove (28), and the inner wall (26) preferably has two reduced sections (36, 44) which compress the sealing ring (30).

The present invention relates to an earphone including a shell and a sound unit, wherein a housing is provided in the shell, a sound outlet is provided on the shell, the sound outlet is communicated with the housing, the sound unit includes a sound tube and a sound element, a mounting chamber and a sound channel are provided in the sound tube, the sound channel is communicated with the mounting chamber, the portion of the sound tube in the mounting chamber is arranged in the housing, the portion of the sound tube in the sound channel is exposed outside the sound outlet, the sound element is arranged in the mounting chamber, the sound outlet surface of the sound element is arranged toward the sound channel, the sound tube can play a better protective role on the sound element, reduce the probability of damaging the sound element due to the shell being squeezed, and can better protect the sound unit. An ear-sticking protrusion is provided outside the shell, the ear-sticking protrusion is used to hold against the inner side wall of the concha cavity, and by providing a flexible ear-sticking protrusion, the flexible shell can be more closely attached to the concha cavity of the human ear.

Methods, systems, and computer programs for encoding and decoding channel coherence parameters applied based on frequency bands, wherein the coherence parameters for each frequency band form a coherence vector. Coherence vectors are encoded and decoded using a prediction scheme, followed by variable bit rate entropy coding.

The application discloses a kind of early motor neuron disease diagnostic methods, belong to medical auxiliary technical field, specifically include: receiving single task and limb rhythm double task under the condition of knocking voice signal sequence and facial infrared thermography video stream data, after time stamp alignment, division analysis section is according to pre-set vocalization unit, extract sound channel micro-disturbance fluctuation sequence and lip movement trajectory curve from each analysis section, carry out instantaneous phase synchronization analysis and obtain coupling stability parameter value, compare the coupling stability parameter value of two kinds of task conditions under the same vocalization grade, calculate stability attenuation ratio value, the attenuation ratio value of different vocalization grade is composed of attenuation feature vector, input pre-constructed attenuation feature discriminant space, output classification mark by pre-division area in discriminant space, the application is thus realized to the analysis and use of sound channel micro-disturbance and lip movement coupling stability attenuation feature under double task load.

The invention aims to provide the semi-closed sound track training bottle convenient to adjust resistance, which is convenient to use in daily life and flexible to adjust air pressure resistance, and the use method of the semi-closed sound track training bottle convenient to adjust resistance. The bottle comprises a bottle body, and a connecting opening is formed in the top of the bottle body; the cover body is in locking fit with the connecting port, the cover body is provided with a first pipeline and a second pipeline, one end of the first pipeline penetrates through the cover body and extends into the bottom of the bottle body, one end of the second pipeline is communicated with the cover body, a reserved cavity is formed in the middle of the second pipeline, and the reserved cavity is communicated with the first pipeline. The reserved cavity is rotationally connected with an adjusting piece used for controlling the air flow. And each plug piece is correspondingly and detachably matched with the extending section of the first pipeline or the extending section of the second pipeline. The invention is applied to the technical field of semi-closed sound channel training bottles.

A speech analysis system that can accurately determine whether a predetermined sound in a spoken work has been correctly pronounced. The speech analysis system includes a machine learning algorithm that has been trained to consider temporal and spectral information about the frame-by-frame components of a target sound. The speech analysis system is personalized based on previous therapy recordings for a speaker's specific error and correct / mispronounced exemplars from speakers with matched vocal tracts. The speech analysis system may be integrated into an adaptive therapy program, such as Speech Motor Chaining, to provide an assessment of proper speech and personalized biofeedback in the place of a live clinician.