13 results about "Music identification" patented technology

This invention relates to the field of music signal recognition technology, and more particularly to a method for recognizing the degree of music enjoyment based on music-EEG feature fusion. The method includes: extracting the Mel-frequency cepstral coefficients of the music signal; preprocessing and performing Fast Fourier Transform on the Mel-frequency cepstral coefficients to obtain the spectral feature values ​​of the music signal; and sequentially filtering, reducing dimensionality, and decorrelating the spectral feature values ​​of the music signal; extracting the Mel-frequency cepstral coefficients of the EEG signal; optimizing and updating the order of the Mel-frequency cepstral coefficients of the EEG signal so that the comprehensive error value between the EEG feature reconstruction signal and the original EEG signal meets a preset condition; aligning the music feature coefficients and the EEG feature reconstruction signal in the feature dimension and time axis using the FastDTW algorithm; and recognizing the degree of music enjoyment of the subject based on the alignment result. This invention effectively improves the robustness and accuracy of music recognition.

The invention discloses a volume control method and device in a live broadcast process, equipment and a storage medium, and relates to the technical field of Internet. The method is executed by a terminal device of an anchor in a live broadcast room, and comprises the following steps: collecting a live broadcast picture stream of the anchor; displaying the live broadcast picture stream in a live broadcast interface of the live broadcast room; a live broadcast interface of the live broadcast room has background music; identifying limb actions of an anchor in the live broadcast picture stream; determining a volume adjustment operation corresponding to the limb movement; and adjusting the volume of background music in the live broadcast room based on the volume adjustment operation. The method can simplify the volume adjustment operation of the anchor and improve the live broadcast effect.

The invention provides a music time coding method for artificial neural network modeling. The method aims at the coding requirement of music beat time, molecules are decomposed according to a fixed denominator for multi-bit independent classification coding, and the method comprises the following steps: setting the fixed denominator as a composite number and decomposing the composite number into a prime factor combination; the molecules are mapped into vectors, the value range of each bit is determined by a cardinal number of a prime factor combination, and the vectors are converted into one-hot coding sub-vectors with redundant components removed. The technical scheme of the invention solves the problem that the traditional mean square error (MSE) ignores the internal correlation of scores and has high sparsity in the topology recognition task of an optical music recognition system (OMR), and compared with a traditional coding method adopting floating point numerical output and a mean square error loss function under the conditions of the same data set, model architecture and training hyper-parameters, the method provided by the invention has the advantages that the coding efficiency is greatly improved; according to the method, the comprehensive error index is reduced by about 21%, and the model precision and the training efficiency are improved.

The invention relates to the technical field of music recognition education, and discloses a music rhythm recognition and evaluation method and system based on multi-modal feedback. The music rhythm identification and evaluation method is applied to the music rhythm identification and evaluation electronic equipment, and specifically comprises the following steps: S101, receiving a music rhythm identification and evaluation request of a user side, and collecting time sequence data and force information of a pressing action of a user in real time through a sensor arranged in the music rhythm identification and evaluation electronic equipment, meanwhile, audio signals are collected and captured through a high-fidelity microphone or a direct audio interface. According to the method, objective quantitative evaluation of music rhythm capability is realized through a multi-modal data fusion analysis mechanism, three-modal synchronous acquisition and fusion analysis of time sequence data, strength information and audio signals are adopted, and a multi-level fusion strategy of a data level, a feature level and a decision level is adopted; an advanced algorithm combining dynamic time warping and a hidden Markov model overcomes the limitation of traditional single-mode evaluation.

A video editing method is performed by a computer device. The method includes: performing music recognition on audio data in first video data to obtain a recognition result of each of audio frames in the audio data, the recognition result indicating whether the audio frame belongs to a music audio frame; determining a music segment in the audio data based on the recognition results of the audio frames, the music segment comprising a plurality of music audio frames; and extracting, from the first video data, a video clip with a same playback period as the music segment as second video data comprising the music segment.

A computer-implemented method for providing a holistic music experience is disclosed. The method includes receiving musical input from users, transforming the input into an identifiable melody using a unique identification algorithm, and matching the melody against a database of musical compositions. Information about the matched musical composition is retrieved and displayed. User instructions to alter the tone or speed of the matched musical composition are received and implemented. A specific musical device can be silenced based on user instructions. The method also includes receiving user-submitted written music, capturing personalized audio and visual recordings, and displaying these recordings to a global audience.

According to the sensitive music recognition device and method, the problems that generalization ability is poor and mismatching is prone to occurring can be solved, firstly, HPCP features are used as audio fingerprints, the HPCP features are generally not greatly changed for different sensitive music turning versions, and therefore the generalization ability of the fingerprint features can be effectively improved; and secondly, continuous multi-frame audio fingerprints are spliced into a relatively long audio fingerprint by using a sliding window algorithm, and the relatively long audio fingerprint comprises change information of an audio time sequence, so that the mismatching rate during large-scale search can be effectively reduced.

A sound recording device and system able to automatically record and transmit sound data corresponding to a preset recording interval time and match the sound data to specific sounds and songs. A sound recording device includes housing, a processor, a microphone, cellular and wireless communication devices. The processor includes an interval timer module and location module. The interval timer module is configured to count down a predetermined time interval corresponding to a recording length of time during which the microphone is set to record at least one input sound signal. The location module is configured to determine or receive location data and store in in a location database. The wireless communication device is configured to receive the at least one input sound signal from the processor and transmit the at least one sound signal and the location data to a network or computer device.

The application provides a vehicle-mounted music recognition method, which comprises the following steps: step 1, determining whether a command for starting music recognition is received; if yes, the next step is performed; otherwise, the current step is repeatedly performed; step 2, starting audio recording, recording current music with a preset time length, generating a recording file after pretreatment and saving the recording file in a cache; step 3, uploading the recording file to a server, receiving recognition information returned by the server and determining whether the current music is free to download according to the recognition information; if yes, step 5 is performed; otherwise, the next step is performed; step 4, determining whether a command for agreeing to pay for downloading is received; if yes, the server is accepted and the next step is performed; otherwise, step 7 is performed; step 5, downloading corresponding music and performing the next step; step 6, playing the downloaded music; and step 7, exiting the music recognition mode. The application improves the intelligent degree of vehicle-mounted music recognition.

The present disclosure provides a music recognition method and a training method and device of a music feature extraction model, and relates to the technical field of artificial intelligence, in particular to the technical field of audio processing and intelligent recommendation. The implementation scheme is as follows: determining a music segment in a to-be-recognized audio, the music segment being an audio segment containing a music component; extracting a first feature vector of the music segment; and determining at least one target music matching the to-be-recognized audio from a music library based on the first feature vector.

A system for closed-loop piano learning in augmented reality (XR), consisting of: an optical music recognition (OMR) engine configured to receive note inputs, preprocess the note inputs through rotation correction and image enhancement, segment the preprocessed notes into staves and musical symbols using dual neural network architectures, analyze the position and relationship of the segmented musical symbols, and generate a structured MusicXML file based on the analyzed musical symbols; a virtual reality (VR) headset configured to display a rendered virtual environment including a virtual piano, track a user's hand movements, and register the user's tracked hand movements; a virtual reality piano interface module connected to the VR headset, the virtual reality piano interface module configured to: render a three-dimensional virtual piano model with individual interactive keys; dynamically highlight specific piano keys based on the sequential piano key instructions derived from the MusicXML file generated by the OMR engine; detect collisions between virtual representations of the user's fingers and the virtual piano keys as registered via the VR headset; provide real-time auditory, visual, and haptic feedback based on the detected collisions; and adjust the difficulty level of the piano lesson based on the user's performance; and a computing device comprising a processor and a GPU, connected to a memory, and configured to control the operation of a virtual reality piano interface module and a VR headset, wherein the computing device is configured to implement a virtual reality piano interface, process the MusicXML file to create sequential piano key instructions, transmit visual rendering data to the VR headset via the interface module, and process hand tracking data received from the VR headset.

The invention relates to the technical field of music signal identification, in particular to a music enjoyment degree identification method based on music-electroencephalogram feature fusion, which comprises the following steps: extracting a Mel-frequency cepstrum coefficient of a music signal, carrying out preprocessing and fast Fourier transform on the Mel-frequency cepstrum coefficient to obtain a music signal frequency spectrum feature value, and identifying the music enjoyment degree based on the music-electroencephalogram feature fusion. Carrying out filtering, dimension reduction and decorrelation processing on the music signal frequency spectrum characteristic value in sequence; extracting the Mel-frequency cepstral coefficient of the electroencephalogram signal, and optimizing and updating the order of the Mel-frequency cepstral coefficient of the electroencephalogram signal, so that a comprehensive error value of the electroencephalogram feature reconstruction signal of the Mel-frequency cepstral coefficient of the electroencephalogram signal and the original electroencephalogram signal meets a preset condition; a FastDTW algorithm is adopted to align the music feature coefficient and the electroencephalogram feature reconstruction signal on the feature dimension and the time axis; and identifying the music enjoyment degree of the to-be-identified person based on the alignment result. According to the invention, the robustness and accuracy of music identification are effectively improved.

A music recognition model training method executed by a computer device, comprising: acquiring a sample data set, in which each piece of sample data comprises a sample music clip and a sample music description corresponding thereto, the sample music description comprising a plurality of sample music attributes of the sample music clip (S201); and, on the basis of the sample data set, iteratively training a plurality of rounds a music recognition model to be trained (S202), the music recognition model comprising: a trained feature extraction sub-model and text recognition sub-model and a feature mapping sub-model to be trained, and each round of the iterative training comprising: using the feature extraction sub-model to extract a spectrum feature presented by the sample music clip (S2021); using the feature mapping sub-model to map the obtained spectrum feature to a description text feature, the description text feature being used for describing a plurality of training music attributes of the sample music clip (S2022); using the text recognition sub-model to recognize the description text feature so as to obtain a training music description (S2023); and, on the basis of the difference between the obtained training music description and the sample music description corresponding thereto, adjusting model parameters of the feature mapping sub-model (S2024).