289 results about "Letter to sound" patented technology

A method and apparatus are provided for segmenting words into component parts. Under the invention, mutual information scores for pairs of graphoneme units found in a set of words are determined. Each graphoneme unit includes at least one letter. The graphoneme units of one pair of graphoneme units are combined based on the mutual information score. This forms a new graphoneme unit. Under one aspect of the invention, a syllable n-gram model is trained based on words that have been segmented into syllables using mutual information. The syllable n-gram model is used to segment a phonetic representation of a new word into syllables. Similarly, an inventory of morphemes is formed using mutual information and a morpheme n-gram is trained that can be used to segment a new word into a sequence of morphemes.

A personal alerting device and method for detecting an approaching sound source includes a sound detector for detecting environmental sounds and for providing an electrical signal to a sound analyzer. The sound signal is analyzed to determine a baseline sound pattern comprising a plurality of distinct sounds corresponding to sounds emitted from a reference sound source. The distinct sounds in the baseline sound pattern may have substantially the same amplitude and time interval. The sound signal is monitored and compared against the baseline sound pattern to determine whether a target sound pattern is present in the sound signal, the target sound pattern corresponding to sounds emitted by the approaching sound source. When it is determined that the target sound is present in the sound signal, one or more of an audible, visual and tactile alert may be emitted to provide warning of the approaching sound source.

The invention discloses a mutually translating system of sign language and speech, a gesture image collecting module 101 is used for collecting the video data of gestures, an input image preprocessing module 102 is used for image preprocessing, an image characteristic extracting module 103 is adopted for image characteristic extraction of the video data after image preprocessing and then outputs 56-dimension characteristic vectors, the 56-dimension characteristic vectors are used for constructing a sign language model 104, a continuous and dynamic sign language recognizing module 105 is used for recognizing the sign language model 104, and recognition results are output and translated into Chinese speech through a Chinese sounding module 106; voice signals collected by a voice signal collecting device are input in a speech recognition programming interface of Microsoft Speech SDK 5.1 and converted into characters to be output; three-dimensional models and three-dimensional animation are established through three-dimensional modeling software; the information of the three-dimensional models and the three-dimensional animation is output into .x formatted files through a Panda plug-in; and DirectX 3D is utilized to load the .x formatted three-dimensional models and the three-dimensional animation and then output sign language animation.

A media capture device has an audio input receptive of user speech relating to a media capture activity in close temporal relation to the media capture activity. A plurality of focused speech recognition lexica respectively relating to media capture activities are stored on the device, and a speech recognizer recognizes the user speech based on a selected one of the focused speech recognition lexica. A media tagger tags captured media with generated speech recognition text, and a media annotator annotates the captured media with a sample of the user speech that is suitable for input to a speech recognizer. Tagging and annotating are based on close temporal relation between receipt of the user speech and capture of the captured media. Annotations may be converted to tags during post processing, employed to edit a lexicon using letter-to-sound rules and spelled word input, or matched directly to speech to retrieve captured media.

A behavior control system of a robot for learning a phoneme sequence includes a sound inputting device inputting a phoneme sequence, a sound signal learning unit operable to convert the phoneme sequence into a sound synthesis parameter and to learn or evaluate a relationship between a sound synthesis parameter of a phoneme sequence that is generated by the robot and a sound synthesis parameter used for sound imitation, and a sound synthesizer operable to generate a phoneme sequence based on the sound synthesis parameter obtained by the sound signal learning unit.

An apparatus and a method that achieve physical separation of sound sources by pointing directly a beam of coherent electromagnetic waves (i.e. laser). Analyzing the physical properties of a beam reflected from the vibrations generating sound source enable the reconstruction of the sound signal generated by the sound source, eliminating the noise component added to the original sound signal. In addition, the use of multiple electromagnetic waves beams or a beam that rapidly skips from one sound source to another allows the physical separation of these sound sources. Aiming each beam to a different sound source ensures the independence of the sound signals sources and therefore provides full sources separation.

The invention discloses a city noise identification method based on hybrid deep neural network models. The city noise identification method comprises the following steps that 1, city noise is collected, and a sound sample database is built; 2, sound signals in the sound sample database are converted into a speech spectrum; 3, the obtained speech spectrum is clipped, and then feature extracting isconducted by using the multiple pre-trained deep neural network models; 4, features extracted by the multiple models are spliced; 5, the spliced fusion feature serves as final input of a classifier, and a prediction model is trained; and 6, as for unknown sound, the sound is converted into the speech spectrum firstly, feature extracting is conducted by using the multiple pre-trained deep neural network models, the extracted features are spliced, then prediction is conducted by using the trained prediction model, and the final sound type is obtained. A large quantity of datasets are not needed,the operating rate is higher, and needed resources are fewer.

The invention discloses a video interaction control method and device, and belongs to the technical field of monitoring. The device comprises a PTZ camera (1), a first direction sound collection unit (2), a second direction sound collection unit (3), a processing unit (4), and a communication unit (5). The processing unit (4) is used for recognizing the obtained sound information, calculating the position information of a sound source according to first sound information and second sound information when the obtained sound information belongs to preset sound sample data, and controlling the PTZ camera (1) to turn to the sound source according to the calculated position information. Through the recognition of the sound generated by the sound source, the device controls the PTZ camera (1) to turn to the sound source when the sound generated by the sound source belongs to the preset sound sample data, and carries out the interaction of the obtained video information of the sound source and a preset video interaction object, thereby enabling the emergency of the sound source to be timely fed back to the video interaction object, and enabling the sound source to be rescued or cared timely.

An improved method and arrangement for altering musical dynamics of a sound S included in a sound signal is disclosed. The altering of the musical dynamics is performed by filtering and amplification of the sound signal. The filtering is performed by the use of a parametric equalizer, the parametric equalizer having a first gain G1 and a resonance frequency f, being related to a pitch frequency fp of said sound S. The amplification is performed by an amplifier amplifying the sound signal with a second gain G2, the second gain G2 being dependent on the first gain G1.

The invention relates to the field of language learning, in particular to a language learning device, comprising an input module for receiving the voice information inputted by a user, a standard mandarin word stock with the storage of spellings of commonly used characters and words and standard mandarin audio data, a comparison module for comparing the inputted voice information with the standard mandarin audio data, a display module for displaying a comparison result and a playing module for playing the standard mandarin audio data. The language learning device has the beneficial effect of finding wrong pronunciation by comparing the inputted voice information with the standard mandarin audio data, and helping to correct the pronunciation habit of a user by correct pronouncing demonstration and promotion.

Provided is a sound event recognition method that may improve a sound event recognition performance using a correlation between difference sound signal feature parameters based on a neural network, in detail, that may extract a sound signal feature parameter from a sound signal including a sound event, and recognize the sound event included in the sound signal by applying a convolutional neural network (CNN) trained using the sound signal feature parameter.

A method is disclosed for discriminating voiced and unvoiced sounds in speech. The method detects characteristic waveform features of voiced and unvoiced sounds, by applying integral and differential functions to the digitized sound signal in the time domain. Laboratory tests demonstrate extremely high reliability in separating voiced and unvoiced sounds. The method is very fast and computationally efficient. The method enables voice activation in resource-limited and battery-limited devices, including mobile devices, wearable devices, and embedded controllers. The method also enables reliable command identification in applications that recognize only predetermined commands. The method is suitable as a pre-processor for natural language speech interpretation, improving recognition and responsiveness. The method enables realtime coding or compression of speech according to the sound type, improving transmission efficiency.

An apparatus and a method that achieve physical separation of sound sources by pointing directly a beam of coherent electromagnetic waves (i.e. laser). Analyzing the physical properties of a beam reflected from the vibrations generating sound source enable the reconstruction of the sound signal generated by the sound source, eliminating the noise component added to the original sound signal. In addition, the use of multiple electromagnetic waves beams or a beam that rapidly skips from one sound source to another allows the physical separation of these sound sources. Aiming each beam to a different sound source ensures the independence of the sound signals sources and therefore provides full sources separation.

A method and apparatus are provided for segmenting words into component parts. Under the invention, mutual information scores for pairs of graphoneme units found in a set of words are determined. Each graphoneme unit includes at least one letter. The graphoneme units of one pair of graphoneme units are combined based on the mutual information score. This forms a new graphoneme unit. Under one aspect of the invention, a syllable n-gram model is trained based on words that have been segmented into syllables using mutual information. The syllable n-gram model is used to segment a phonetic representation of a new word into syllables. Similarly, an inventory of morphemes is formed using mutual information and a morpheme n-gram is trained that can be used to segment a new word into a sequence of morphemes.

The invention relates to a speech emotion recognition method based on a multistage residual convolutional neural network, and belongs to the technical field of speech signal analysis, image processingand the like. The method comprises the following steps: 1) a training process: collecting and preprocessing sound signals with all emotions to generate a spectrogram; constructing a multi-stage residual convolutional neural network, and inputting the spectrogram into the multi-stage residual convolutional neural network for training; 2) a test process: acquiring and preprocessing a to-be-identified sound signal, and generating a to-be-identified spectrogram; and then inputting the to-be-identified spectrogram into the trained multistage residual convolutional neural network to obtain a recognition result. According to the method, the CNN is subjected to feature compensation by crossing multi-stage residual blocks, so that the problem of feature loss of the CNN along with deepening of a convolution layer is solved, and the recognition rate is increased.

The embodiment of the invention provides an entrance guard control method and device. The entrance guard control method comprises the steps that a sound signal is collected; the sound signal is subjected to voice recognition, a recognition result is obtained, and whether the recognition result is matched with preset keywords or not is judged; and if the recognition result is matched with the preset keywords, the sound signal is subjected to voiceprint recognition, and whether the sound corresponding to the sound signal is derived from a target user or not is confirmed; and if yes, an entranceguard is opened. Through the entrance guard control method and device, opening of the entrance guard is controlled through voice recognition and voiceprint recognition, the trouble that unlocking needs to be operated by hand is avoided, and thus convenience is improved.

In a method and system for identifying speech sound and non-speech sound in an environment, a speech signal and other non-speech signals are identified from a mixed sound source having a plurality of channels. The method includes the following steps: (a) using a blind source separation (BSS) unit to separate the mixed sound source into a plurality of sound signals; (b) storing spectrum of each of the sound signals; (c) calculating spectrum fluctuation of each of the sound signals in accordance with stored past spectrum information and current spectrum information sent from the blind source separation unit; and (d) identifying one of the sound signals that has a largest spectrum fluctuation as the speech signal.

The invention provides a conversational speech analyzer which analyzes whether utterances in a meeting are of interest or concern. Frames are calculated using sound signals obtained from a microphone and a sensor, sensor signals are cut out for each frame, and by calculating the correlation between sensor signals for each frame, an interest level which represents the concern of an audience regarding utterances is calculated, and the meeting is analyzed.

A sound event detection method includes receiving a sound signal and determining and outputting whether a sound event is present in the sound signal by applying a trained neural network to the received sound signal, and performing post-processing of the output to reduce an error in the determination, wherein the neural network is trained to early stop at an optimal epoch based on a different threshold for each of at least one sound event present in a pre-processed sound signal. That is, the sound event detection method may detect an optimal epoch to stop training by applying different characteristics for respective sound events and improve the sound event detection performance based on the optimal epoch.

The invention discloses a sound box control method, a sound box and a sound box system, and relates to the fields of intelligent terminals, man-machine interaction and the like. The method comprises the steps that a first sound box collects a first sound signal, and a second sound box collects a second sound signal; the first sound box determines the position of a sound source according to the first sound signal and the second sound signal, and determines a first distance between the sound source and the first sound box and a second distance between the sound source and the second sound box; the first sound box determines a time delay difference based on the first distance and the second distance; the first sound box indicates the second sound box to make a sound at a second moment, the second moment is determined according to the first moment and the time delay difference, and the first moment is the sound making time of the first sound box; the first sound box sends out a third soundsignal at a first moment, and the second sound box sends out a fourth sound signal at a second moment, the third sound signal and the fourth sound signal are signals of different sound channels of the same audio file. In this way, a user can obtain a good three-dimensional sound effect at different positions in a room.

The invention relates to a multiway electronic switch which comprises a multi-channel interface, a single channel interface, a sound collecting unit and a control unit; wherein, the sound collecting unit is used for collecting sound signals to carry out the sound-electric transition and generate sound-electric signals; and the control unit is used for communicating each channel in the multi-channel interface with the single channel interface one by one according to the sound-electric signals; the multiway electronic switch is used as the medium for connecting an electronic device to be detected with a test device and controls the channel switching of the multiway electronic switch through the sound to realize the connection of the multi-channel of the electronic device to be tested with the detection device one by one. Compared with the previous channel switching method for replacing the interface, the test operation of the invention is more convenient, and the physical damage which is brought by the insertion and pulling of the interface is avoided. The invention also provides a test device adopting the multiway electronic switch.

The provided method for protecting audition comprises: detecting the volume degree of environment to obtain the reference energy value; receiving analog sound signal from a source, conversing the analog signal into digital signal; sampling the signal with pre-set sampling frequency to obtain multiple amplitude values; taking these amplitude values as one parameter to calculate the sound energy value in a pre-set time interval; when the sound energy value up to the former reference value, generating a protective signal and reducing the gain value.

The invention discloses an intelligent monitoring system based on sound source positioning. Abnormal sound can be precisely judged and positioned in daily patrol processes, and the orientation of the abnormal sound is obtained at the first time, so the abnormal event can be timely processed. The intelligent monitoring system comprises a sound acquisition unit, a sound position discrimination unit, an abnormal sound discrimination unit and an execution unit. The sound acquisition unit acquires sound signals in real time and sends the acquired sound to the sound position discrimination unit so as to determine the specific orientation of the sound. The abnormal sound discrimination unit discriminates the sound signals, judges whether the sound is abnormal sound, and if the sound is abnormal sound, controls the execution mechanism to give out an alarm signal to a main control chamber and rotate a holder camera to pass the image of a generation place of the abnormal sound back the main control chamber for reference of workers in the main control chamber.

The invention provides a method, a device and a system for determining a relative angle between intelligent devices and the intelligent devices. The method is suitable for a first intelligent device,the first intelligent device comprises a first sound detection module and a second sound detection module, and the method comprises the following steps: enabling the first sound detection module to detect a first sound signal sent by a second intelligent device and directly reaching the first sound detection module, enabling a second sound detection module to detect a second sound signal sent by the second intelligent device and directly reaching the second sound detection module, wherein the first sound signal and the second sound signal are sent by the second intelligent device at the same time; determining the time difference between the receiving moment of the first sound signal and the receiving moment of the second sound signal; and determining the relative angle between the first intelligent device and the second intelligent device based on the distance between the first sound detection module and the second sound detection module and the time difference. The relative angle between the intelligent devices can be determined quickly, simply, conveniently and accurately.

Presented herein are systems and methods for processing sound signals for use with electronic speech systems. Sound signals are temporally parsed into frames, and the speech system includes a speech codebook having entries corresponding to frame sequences. The system identifies speech sounds in an audio signal using the speech codebook.

Focusing sound signals in a shared 3D space uses an array of physical microphones, preferably disposed evenly across a room to provide even sound coverage throughout the room. At least one processor coupled to the physical microphones does not form beams, but instead preferably forms 1000's of virtual microphone bubbles within the room. By determining the processing gains of the sound signals sourced at each of the bubbles, the location(s) of the sound source(s) in the room can be determined. This system provides not only sound improvement by focusing on the sound source(s), but with the advantage that a desired sound source can be focused on more effectively (rather than steered to) while un-focusing undesired sound sources (like reverb and noise) instead of rejecting out of beam signals. This provides a full three dimensional location and a more natural presentation of each sound within the room.

The invention provides a sound signal separation method of double sound sources and a sound pickup. The method comprises the followings steps: dividing the mixed sound signals into voice frames, estimating the delay inequality of the voice frames reaching different array element combinations in the microphone array, judging the propagation direction of the voice frames according to the determineddelay inequality, separating sound signals corresponding to different sound sources in real time according to the propagation direction, and outputting the sound signals. Time delay estimation is carried out through a generalized cross-correlation algorithm, time delay can be accurately estimated, the calculation amount of the algorithm is low, the algorithm can track the sound source orientationmore accurately and efficiently in a real-time system, and therefore automatic separation of sound signals of a first sound source and a second sound source is achieved.

The invention provides a disease detection method based on cough sound recognition, which comprises the following steps of: acquiring sound information to be detected or training sound information, and extracting a characteristic Mel-frequency cepstrum coefficient of the sound information to be detected or extracting a training Mel-frequency cepstrum coefficient of the training sound information;drawing a characteristic Mel spectrogram by taking time and frequency as axes according to the characteristic Mel frequency cepstrum coefficient, or respectively drawing a plurality of training Mel spectrograms by taking time and frequency as axes according to the plurality of training Mel frequency cepstrum coefficients; and training the plurality of training Mel spectrograms through a preset convolutional neural network to obtain a feature convolutional neural network model. Compared with the prior art, the method has the advantages that the detection process of related diseases is simple, and the detection efficiency is high.

The invention discloses a speech synthesis device supporting styles of multiple speakers, language switching and controllable rhythm, and belongs to the field of speech synthesis. The device comprises a text acquisition unit and a text preprocessing unit which are used for acquiring and preprocessing different text data; a language switching unit used for storing and displaying speaker tags corresponding to the training data of different language types and automatically identifying the language type of the text to be synthesized; a style switching unit used for specifying a speech synthesis style according to the language type; a speaker switching unit for specifying a speaker; a coding-decoding unit for obtaining a predicted Mel spectrum; a training unit for training the encoding-decoding unit; and a voice synthesis unit which is used for generating the predicted Mel frequency spectrum and converting the predicted Mel frequency spectrum into a sound signal for voice playing. According to the invention, the speaker and the style of the speaker can be respectively controlled while the voice with richer rhythm change is generated.

The invention discloses a sound enhancement method and a sound enhancement system. The method comprises the following steps: obtaining a sound signal, and converting the sound signal into a digital signal; decomposing the digital signal to obtain a plurality of intrinsic mode functions or a plurality of similar intrinsic mode functions; selectively amplifying the amplitudes of the plurality of obtained intrinsic mode functions or the plurality of similar intrinsic mode functions; integrating the selectively amplified intrinsic mode functions or similar intrinsic mode functions to obtain an integrated reconstruction signal; and converting the integrated reconstructed signal into an analog signal. Based on Hilbert-Huang transform, sound can be effectively and selectively enhanced, only high-frequency consonants in the sound are amplified instead of amplifying vowels, the method can effectively improve the sharpness of the amplified sound, and the problem that only the loudness of the sound is increased but the sharpness is not increased in an existing sound enhancement method is solved.