Audio correction method, system and storage medium

Abstract

The present disclosure provides an audio correction method, system and storage medium; in this method, after a test signal is played and environmental sound is collected by a common terminal device with signal playback and sound collection, it is uploaded to a server, and the computing power resources of the server are used to quickly and accurately identify environmental acoustic parameters, and digital equalizer parameters for users to correct environmental acoustic defects are generated based on the environmental acoustic parameters. The correction parameters generated by the server are then sent to an audio playback device through a terminal, and the audio playback device applies the correction parameters to correct the digital equalizer. Playing audio with the corrected parameters can improve the adaptability of audio playback to the acoustic environment and improve the user's listening experience. The whole process is simple to operate, no professional intervention is required, and the audio operation cost is also reduced.

Description

Technical Field

[0001] The present disclosure relates to the technical field of audio processing, and in particular to an audio correction method, system and storage medium. Background Art

[0002] When audio playback devices are played in different environments, the sound quality varies and the user experience is poor due to differences in the environment, such as the size of the room, the material of the sound reflector, the placement of the audio playback device, etc. Existing audio correction technology usually requires professional equipment and technicians to operate, which is costly and the correction process is cumbersome, making it difficult to adapt to diverse acoustic environments. Summary of the invention

[0003] In view of this, an object of the present disclosure is to provide an audio correction method, system and storage medium to reduce the operating cost of audio correction.

[0004] In a first aspect, an embodiment of the present disclosure provides an audio correction method, which is applied to an audio correction system, wherein the audio correction system includes a terminal, a server and an audio playback device, and the method includes: the terminal responds to a sound collection instruction, plays at least one test signal, and collects ambient sound to obtain target audio data; the terminal uploads the target audio data to the server; the server identifies ambient acoustic parameters of the target audio data to obtain target ambient acoustic parameters; the server generates correction parameters of a digital equalizer based on the target ambient acoustic parameters to obtain target correction parameters; the server sends the target correction parameters to the terminal, and the terminal transmits the received target correction parameters to the audio playback device; the audio playback device calibrates a preset digital equalizer through the received target correction parameters, and plays audio through the corrected digital equalizer.

[0005] Optionally, the terminal responds to the sound collection instruction, plays at least one test signal, and collects ambient sound to obtain the target audio data, including: the terminal responds to the sound collection instruction, obtains at least one pre-generated test signal, and plays the at least one test signal in a preset order; while playing the at least one test signal, the terminal records the ambient sound; when the at least one test signal is played, the terminal saves the recorded ambient sound as the target audio data.

[0006] Optionally, the at least one test signal includes an impulse signal, a logarithmic swept sine signal and pink noise; the target audio data includes a response signal corresponding to each test signal; the server identifies environmental acoustic parameters of the target audio data to obtain target environmental acoustic parameters, comprising: the server converts the response signal corresponding to each test signal into a pulse signal to obtain a pulse signal corresponding to each test signal; the server calculates the reverberation time, early decay time and clarity of the pulse signal corresponding to each test signal to obtain the target environmental acoustic parameters.

[0007] Optionally, the server converts the response signal corresponding to each test signal into a pulse signal to obtain the pulse signal corresponding to each test signal, including: the server determines the response signal corresponding to the impulse signal as the pulse signal corresponding to the impulse signal; the server performs pulse signal extraction on the response signal corresponding to the logarithmic swept sine signal to obtain the pulse signal corresponding to the logarithmic swept sine signal; the server performs cross-correlation on the pink noise and the response signal corresponding to the pink noise to obtain the pulse signal corresponding to the pink noise.

[0008] Optionally, the server extracts a pulse signal from a response signal corresponding to the logarithmic swept-frequency sine signal to obtain a pulse signal corresponding to the logarithmic swept-frequency sine signal, comprising: the server performs Fourier transform on the response signal corresponding to the logarithmic swept-frequency sine signal to obtain a first frequency response curve; the server filters the frequency domain characteristics of the logarithmic swept-frequency sine signal on the first frequency response curve through a preset frequency domain filter to obtain a second frequency response curve; the server performs inverse filtering on the second frequency response curve to obtain a pulse signal corresponding to the logarithmic swept-frequency sine signal;

[0009] Optionally, the server calculates the reverberation time, early decay time and clarity of the pulse signal corresponding to each test signal to obtain the target environment acoustic parameters, including: the server calculates the energy decay curve, the energy before a preset moment, and the total energy of the pulse signal corresponding to each test signal according to the pulse signal corresponding to each test signal; the server calculates the time required for the energy to decay to a first preset value according to the energy decay curve corresponding to each test signal, and obtains the reverberation time corresponding to each test signal; the server calculates the time required for the energy to decay from a maximum value to a second preset value according to the energy decay curve corresponding to each test signal, and obtains the early decay time corresponding to each test signal; the server calculates the clarity corresponding to each test signal according to the ratio between the energy before a preset moment corresponding to each test signal and the total energy; the server obtains the target environment acoustic parameters in combination with the reverberation time, early decay time and clarity corresponding to each test signal.

[0010] Optionally, the step in which the server obtains the target environment acoustic parameters by combining the reverberation time, early decay time and clarity corresponding to each test signal includes: the server performs weighted averaging on the reverberation time, early decay time and clarity corresponding to each test signal to obtain the target environment acoustic parameters.

[0011] Optionally, the server generates correction parameters of the digital equalizer based on the target environment acoustic parameters, and the step of obtaining the target correction parameters includes: the server extracts sound quality features of the target environment acoustic parameters through a pre-trained sound quality feature extraction model to obtain target sound quality features; the server obtains preset digital equalizer parameters corresponding to the target sound quality features to obtain target digital equalizer parameters; the server converts the target digital equalizer parameters into digital filter parameters to obtain target correction parameters.

[0012] Optionally, the server extracts sound quality features of the target environment acoustic parameters through a pre-trained sound quality feature extraction model to obtain the target sound quality features, comprising: the server identifies abnormal sound quality factors of the target environment acoustic parameters through a pre-trained sound quality feature extraction model to obtain the target sound quality features.

[0013] Optionally, the server obtains the preset digital equalizer parameters corresponding to the target sound quality characteristics, and the step of obtaining the target digital equalizer parameters includes: the server obtains the preset digital equalizer parameters and correction parameter values ​​corresponding to the target sound quality characteristics according to the mapping relationship between the preset sound quality abnormality factors and the digital equalizer parameters, and obtains the target digital equalizer parameters.

[0014] Optionally, the server converts the target digital equalizer parameters into digital filter parameters to obtain target correction parameters, including: the server determines the target digital filter corresponding to the target digital equalizer parameters based on a mapping relationship between preset digital equalizer parameters and digital filters; the server calculates the coefficients of the target digital filter based on the target digital equalizer parameters to obtain target correction parameters.

[0015] Optionally, the audio playback device calibrates a preset digital equalizer through the received target correction parameters, and performs audio playback through the corrected digital equalizer, comprising: the audio playback device uses the received target correction parameters as digital filter parameters and superimposes them with digital filter parameters in the audio playback channel, and performs audio playback through the superimposed digital filter parameters.

[0016] In a second aspect, an embodiment of the present disclosure provides an audio correction system, characterized in that the audio correction system includes a terminal, a server and an audio playback device, and the audio correction system executes the above-mentioned audio correction method.

[0017] In a third aspect, an embodiment of the present disclosure provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the above-mentioned audio correction method.

[0018] The embodiments of the present disclosure bring the following beneficial effects:

[0019] The above-mentioned audio correction method, system and storage medium, after playing the test signal and collecting the ambient sound through an ordinary terminal device with signal playback and sound collection, upload it to the server, use the server's computing power resources to quickly and accurately identify the ambient acoustic parameters, and generate digital equalizer parameters for the user to correct the ambient acoustic defects based on the ambient acoustic parameters, and then send the correction parameters generated by the server to the audio playback device through the terminal. The audio playback device applies the correction parameters to calibrate the digital equalizer. Playing audio with the corrected parameters can improve the adaptability of audio playback to the acoustic environment and improve the user's listening experience. The whole operation is simple, no professional intervention is required, and the audio operation cost is also reduced.

[0020] Other features and advantages of the present disclosure will be described in the following description, and partly become apparent from the description, or understood by practicing the present disclosure. The purpose and other advantages of the present disclosure are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0021] In order to make the above-mentioned objectives, features and advantages of the present disclosure more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the specific embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the specific embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.

[0023] Figure 1 is a flow chart of an embodiment of the audio correction method in the embodiment of the present disclosure;

[0024] Figure 2 The present invention is a flowchart of another embodiment of the audio correction method in the embodiment of the present disclosure. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present disclosure, rather than all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present disclosure.

[0026] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0027] For ease of understanding, the specific process of the embodiment of the present disclosure is described below. Figure 1 The embodiment of the present disclosure is applied to an audio correction system, the audio correction system includes a terminal, a server and an audio playback device, and an embodiment of the audio correction method in the embodiment of the present disclosure includes:

[0028] Step S10: The terminal responds to the sound collection instruction, plays at least one test signal, and collects ambient sound to obtain target audio data;

[0029] It is understandable that the audio correction method in the embodiment of the present disclosure is triggered and executed by the terminal, and the terminal provides a graphical user interface, and the graphical user interface displays a trigger control of a sound collection instruction. The user triggers the sound collection instruction by operating the trigger control, and the terminal responds to the sound collection instruction and starts playing at least one test signal, and collects ambient sound at the same time, thereby obtaining the collected target audio data. In one embodiment, the terminal can also establish a connection with an audio playback device and control the audio playback device to play at least one test signal. When the audio playback device sends a play confirmation instruction to the terminal, the terminal starts collecting ambient sound, thereby obtaining the target audio data.

[0030] In one embodiment, the graphical user interface of the terminal also includes an input control for environmental parameters, wherein the environmental parameters can be any parameters that affect the audio playback effect, such as the size, shape, wall material, location of the audio playback device, etc. of the environmental space. The user can input the environmental parameters through the input control. After the terminal receives the submitted environmental parameters, the environmental parameters can be uploaded to the server, and the server determines the placement position of the terminal based on the environmental parameters, or the terminal directly determines the placement position based on the environmental parameters, wherein the placement position is used to indicate the optimal position for the terminal to play the test signal and collect the ambient sound. For example, the placement position can be the center of the room, the corner of the wall, etc., and the specifics are not limited here.

[0031] In one embodiment, before step S10, the terminal may also control the audio playback device to play simulated sound waves. When the audio playback device sends a play confirmation instruction to the terminal, the terminal starts to collect response signals of the simulated sound waves and uploads the collected response signals to the server. After the server receives the response signals of the simulated sound waves, it performs a simulation calculation of the sound field on the response signals of the simulated sound waves to obtain environmental parameters of the environment in which the audio playback device is located, such as the position of the wall / sound wave reflector in the environment in which the audio playback device is located, the reflection intensity of the wall / sound wave reflector and other environmental parameters, which can be used to determine and select the terminal placement position.

[0032] In one embodiment, after obtaining the environmental parameters, a simulated plan view of the environment can also be drawn, and the user can select any position on the simulated plan view as the placement position through the terminal, or the calculated placement position mark can be displayed on the simulated plan view through the terminal to guide the user to place the terminal at the specified position for sound collection. The placement position can be one or more, and is not specifically limited here.

[0033] In one embodiment, at least one placement position can be generated based on the input environmental parameters, and the graphical user interface of the terminal provides selection controls for all placement positions. After the user places the terminal at any placement position, the user operates the selection control to input the selected placement position into the terminal, and the terminal starts playing at least one test signal and simultaneously collects environmental sound. After the terminal completes playing the test signal of a placement position, the graphical user interface of the terminal displays the selection controls for unselected placement positions, and the user continues to select placement positions. The terminal collects different audio data at different placement positions, thereby obtaining at least one target audio data. In subsequent steps, the terminal uploads all target audio data to the server, and the server identifies environmental acoustic parameters for all target audio data to obtain target environmental acoustic parameters.

[0034] In one embodiment, if there is more than one placement position, the terminal executes the above step S10 at each placement position, and uploads the target audio data corresponding to each placement position obtained after each placement position completes the above step S10 to the server, so that the server can perform environmental acoustic parameter identification based on the ambient sounds collected by multiple placement positions, thereby improving the adaptability of audio correction to the actual audio playback environment, thereby improving the user experience.

[0035] Based on the above, after the terminal obtains the placement position returned by the server or determined by itself, it can also display the placement position guidance through the graphical user interface. The guidance can be one or a combination of more than one of the forms of graphics, text, voice, etc., which are not limited here. The user can place the terminal at the placement position through the placement position guidance displayed by the terminal, trigger the trigger control of the sound collection instruction, and the terminal starts to play at least one test signal and collect ambient sound at the same time. It should be noted that the test signal played by the terminal is pre-generated, not pre-recorded. Therefore, the terminal plays the test signal and collects the ambient sound at the same time, and there will be no howling problem due to signal superposition, and the played test signal will not interfere with the collection of ambient sound.

[0036] In this embodiment, the type of the test signal can be an impulse signal, a logarithmic swept sine signal, pink noise, white noise, a sine wave, a square wave, a sawtooth wave, a triangle wave, a polyphonic signal, a stepped sine wave, a pulse signal or a frequency modulated signal. The test signal can be generated by a preset function / algorithm, which is not specifically limited here.

[0037] It can be understood that the playback duration of each test signal and the playback order of all test signals have been determined in advance. Therefore, in the target audio data containing the collected ambient sound, audio data in different time periods correspond to different types of test signals. For example, assuming that the terminal first plays a 20-second impact signal, then in the target audio data, the audio data from 0s to 20s is the audio data corresponding to the impact signal. No specific limitation is made here.

[0038] Step S20: The terminal uploads the target audio data to the server;

[0039] In this embodiment, the terminal uploads the target audio data including the collected ambient sound to the server, so as to use the computing resources of the server to quickly and accurately generate the equalizer correction parameters, thereby improving the efficiency and accuracy of the audio correction. If the terminal collects target audio data corresponding to more than one placement position, the terminal uploads all the collected target audio data, so that the generation of the equalizer correction parameters is more accurate.

[0040] Step S30: the server performs environmental acoustic parameter recognition on the target audio data to obtain target environmental acoustic parameters;

[0041] In one embodiment, if the terminal collects target audio data corresponding to more than one placement position, the server first integrates all target audio data, obtains the integrated target audio data, and then identifies the environmental acoustic parameters of the target audio data to obtain the target environmental acoustic parameters. Specifically, when the server integrates all target audio data, it can calculate the average value of all target audio data to obtain the integrated target audio data, or it can calculate the weighted average value of all target audio data according to the weights corresponding to different placement positions to obtain the integrated target audio data. It can also cluster all placement positions through a clustering algorithm, select representative positions from each category, and perform weighted average of the target audio data corresponding to all representative positions to obtain the integrated target audio data. The server can also integrate all target audio data through other methods, which will not be described in detail here.

[0042] It is understandable that environmental acoustic parameters are important indicators for describing and evaluating environmental sound characteristics. Environmental acoustic parameters may include: Reverberation Time (RT), Early Decay Time (EDT), Clarity (C50 / C80), Center Time (Ts), Definition (D50), Sound Pressure Level (SPL), Background Noise Level (BNL), etc. The server can extract the features required for environmental acoustic parameter identification from the target audio data, and then calculate the environmental acoustic parameters based on the extracted features to obtain the target environmental acoustic parameters, which are used to indicate the sound characteristics of the ambient sound collection space. Specifically, in one embodiment, the server can identify the environmental acoustic parameters of the target audio data through a pre-trained artificial intelligence model to obtain the target environmental acoustic parameters.

[0043] In one implementation, after receiving the target audio data uploaded by the terminal, the server may also pre-process the target audio data, such as format conversion, noise reduction, etc., to improve the accuracy of digital equalizer parameter correction.

[0044] Step S40: The server generates correction parameters of the digital equalizer based on the target environment acoustic parameters to obtain target correction parameters;

[0045] It can be understood that a digital equalizer is a virtual device used to adjust the gain of different frequency bands of audio signals in digital signal processing. It achieves the purpose of improving sound quality, compensating frequency response, etc. by changing the amplitude of each frequency component in the audio signal, so as to make the audio playback effect better. The digital equalizer performs gain / attenuation adjustment on the audio signal in different frequency bands by adjusting the parameters of the filter of the corresponding frequency band. The gain attenuation adjustment of each frequency band can be understood as applying a specific filter to the frequency band, and the frequency response of the filter is adjusted to increase or decrease the amplitude of the signal in the frequency band. Therefore, in this embodiment, the target correction parameters include the filter parameters of at least one frequency band, and the frequency band to be adjusted and the filter parameters corresponding to the frequency band are determined according to the acoustic parameters of the target environment.

[0046] It should be noted that, according to the frequency of the audio signal, the audio signal can be divided into different frequency bands, such as high frequency, medium frequency and low frequency, and the amplitude of different frequency bands can be adjusted by corresponding filters. By adjusting the amplitude of different frequency bands, the environmental acoustic parameters can be changed, so as to achieve the result of audio playback and acoustic environment adaptation, and enhance the user's audio playback experience. In one embodiment, the server generates correction parameters of the digital equalizer for the target environmental acoustic parameters through a pre-trained artificial intelligence model to obtain the target correction parameters.

[0047] Step S50: the server sends the target correction parameter to the terminal, and the terminal transmits the received target correction parameter to the audio playback device;

[0048] In this embodiment, the terminal acts as a medium between the server and the audio playback device, and sends the target correction parameters generated by the server to the audio playback device to make the audio correction of the audio playback device easier. The mobile phone can be used as a terminal to communicate with the audio playback device. The communication connection method is not limited to cable connection and wireless connection. The cable connection includes USB / Lightning interface connection, and the wireless connection includes Wi-Fi / Bluetooth connection. It should be noted that the target correction parameters can be transmitted to one or more audio playback devices, which is not limited here.

[0049] Step S60: The audio playback device calibrates a preset digital equalizer using the received target calibration parameters, and plays audio using the calibrated digital equalizer.

[0050] It should be noted that the audio playback device includes a player with a digital equalizer. After the audio player receives the target correction parameters through the terminal, the target correction parameters are imported into the player with the digital equalizer to correct the original parameters of the digital equalizer of the player, so as to play the audio through the corrected digital equalizer, thereby achieving the result of mutual adaptation between the audio playback and the acoustic environment, and providing the user with a better listening experience.

[0051] The audio correction method provided in the above-mentioned embodiment plays a test signal and collects ambient sound through an ordinary terminal device with signal playback and sound collection, and then uploads it to a server. The computing power resources of the server are used to quickly and accurately identify ambient acoustic parameters, and generate digital equalizer parameters for users to correct ambient acoustic defects based on the ambient acoustic parameters. The correction parameters generated by the server are then sent to the audio playback device through the terminal. The audio playback device applies the correction parameters to calibrate the digital equalizer. Playing audio with the corrected parameters can improve the adaptability of audio playback to the acoustic environment and improve the user's listening experience. The whole operation is simple, no professional intervention is required, and the audio operation cost is also reduced.

[0052] See also Figure 2 Another embodiment of the audio correction method in the disclosed embodiment includes:

[0053] Step S201, the terminal responds to the sound collection instruction, plays at least one test signal, and collects ambient sound to obtain target audio data; the at least one test signal includes an impulse signal, a logarithmic swept sine signal, and pink noise; the target audio data includes a response signal corresponding to each test signal;

[0054] In one embodiment, the terminal responds to a sound collection instruction, obtains at least one pre-generated test signal, and plays at least one test signal in a preset order; while playing at least one test signal, the terminal records ambient sound; when at least one test signal is played, the terminal saves the recorded ambient sound as target audio data.

[0055] In this embodiment, the terminal plays three test signals in order, including an impulse signal, a logarithmic swept sine signal and pink noise. The playing duration of each test signal is predetermined. Therefore, the response signal corresponding to each test signal can be determined in the ambient sound recorded by the terminal. Different environmental acoustic parameter calculation methods are used for the response signals corresponding to different test signals to improve accuracy.

[0056] It should be noted that an impulse signal is a high-energy pulse signal that occurs in a very short period of time and can be used to measure the impulse response of the environment; a logarithmic sine sweep signal (Log Sweep) is a sine wave signal whose frequency changes logarithmically over time. The frequency change of this signal covers a wide frequency band, gradually increasing from low frequency to high frequency. The logarithmic swept signal can evenly excite each frequency component, thereby making it possible to more comprehensively measure the frequency response of the environment; pink noise is a noise signal whose power spectral density is inversely proportional to the frequency, that is, the energy in each octave is equal. Pink noise can be used for environmental noise simulation and can provide a uniform spectrum distribution.

[0057] Step S202: The terminal uploads the target audio data to the server;

[0058] The execution process of step S202 is similar to the execution process of step S20, and will not be described in detail here.

[0059] Step S203: The server converts the response signal corresponding to each test signal into a pulse signal to obtain a pulse signal corresponding to each test signal;

[0060] In one embodiment, the server determines the response signal corresponding to the impact signal as the pulse signal corresponding to the impact signal; the server performs pulse signal extraction on the response signal corresponding to the logarithmic swept sine signal to obtain the pulse signal corresponding to the logarithmic swept sine signal; the server performs cross-correlation on pink noise and the response signal corresponding to the pink noise to obtain the pulse signal corresponding to the pink noise.

[0061] In this embodiment, different types of test signals correspond to different pulse signal extraction methods. First, for the impulse signal, since the impulse signal itself is a signal that is close to an ideal pulse, its response can be approximately considered as the pulse response of the ambient space; and for the logarithmic swept sine signal, the pulse signal can be extracted by deconvolution, which is not limited here; for pink noise, the input pink noise signal and the collected response signal can be cross-correlated to obtain the pulse signal of the ambient space. Different test signals have different characteristics. Using different pulse signal extraction methods for different test signals can more accurately retain the pulse characteristics of the test signal and improve the accuracy of audio correction.

[0062] In one embodiment, the server extracts a pulse signal from a response signal corresponding to a logarithmic swept frequency sine signal to obtain a pulse signal corresponding to the logarithmic swept frequency sine signal, comprising: the server performs Fourier transform on the response signal corresponding to the logarithmic swept frequency sine signal to obtain a first frequency response curve; the server filters the frequency domain characteristics of the logarithmic swept frequency sine signal on the first frequency response curve through a preset frequency domain filter to obtain a second frequency response curve; the server performs inverse filtering on the second frequency response curve to obtain a pulse signal corresponding to the logarithmic swept frequency sine signal;

[0063] In this implementation, for the logarithmic swept frequency sine signal, the response signal can be first Fourier transformed to obtain the frequency response curve of the ambient space, and then the frequency characteristics of the logarithmic swept frequency sine signal itself can be removed by a preset frequency domain filter, leaving only the influence of the ambient space, and finally the impulse response of the ambient space is calculated by an inverse filtering algorithm to obtain the pulse signal corresponding to the logarithmic swept frequency sine signal. This implementation can more accurately obtain the pulse signal corresponding to the logarithmic swept frequency sine signal, thereby improving the accuracy of audio correction.

[0064] Step S204: The server calculates the reverberation time, early decay time and clarity of the pulse signal corresponding to each test signal to obtain the target environment acoustic parameters;

[0065] In one embodiment, the server calculates the energy decay curve of the pulse signal corresponding to each test signal, the energy before a preset moment, and the total energy based on the pulse signal corresponding to each test signal; the server calculates the time required for the energy to decay to a first preset value based on the energy decay curve corresponding to each test signal, and obtains the reverberation time corresponding to each test signal; the server calculates the time required for the energy to decay from a maximum value to a second preset value based on the energy decay curve corresponding to each test signal, and obtains the early decay time corresponding to each test signal; the server calculates the clarity corresponding to each test signal based on the ratio between the energy before a preset moment corresponding to each test signal and the total energy; the server obtains the target environment acoustic parameters by combining the reverberation time, early decay time, and clarity corresponding to each test signal.

[0066] In this embodiment, the server can calculate the reverberation time, early decay time and clarity as environmental acoustic parameters based on the pulse signal corresponding to each test signal. Specifically, the server can obtain the energy decay curve corresponding to each test signal by calculating the square of the pulse signal corresponding to each test signal and then taking the logarithm. The reverberation time and early decay time can be calculated through the energy decay curve, wherein the reverberation time is generally represented by the time required for the sound pressure level to drop by 60 dB. Therefore, according to the energy decay curve corresponding to each test signal, the time required for the energy to decay by 60 dB is calculated to obtain the reverberation time corresponding to each test signal. The early decay time is generally represented by the time required for the sound pressure level to drop to 10 dB. Therefore, according to the energy decay curve corresponding to each test signal, the time required for the energy to decay from the maximum value to 10 dB is calculated to obtain the early decay time corresponding to each test signal. Clarity is a measure of the ratio of early energy to later energy. Common ones are C50 and C80, which represent the ratio of the energy in the first 50 milliseconds and the first 80 milliseconds to the total energy, respectively. Taking C50 as an example, the server calculates the energy in 0-50 milliseconds, then calculates the total energy of the entire response signal, and finally calculates the ratio of the energy in 0-50 milliseconds to the total energy, and then takes the logarithm to obtain the clarity corresponding to each test signal.

[0067] In one implementation, the reverberation time, early decay time, and clarity corresponding to each test signal may be directly combined to obtain the target environment acoustic parameters, or the reverberation time, early decay time, and clarity corresponding to each test signal may be fused according to a preset algorithm to obtain the target environment acoustic parameters, which is not specifically limited here.

[0068] In one embodiment, the step of obtaining the target environment acoustic parameters by combining the reverberation time, early decay time and clarity corresponding to each test signal by the server includes: the server performs weighted averaging on the reverberation time, early decay time and clarity corresponding to each test signal to obtain the target environment acoustic parameters. In this embodiment, different test signals may correspond to different weights. According to the weight value corresponding to each test signal, the reverberation time, early decay time and clarity corresponding to each test signal are weighted averaged to obtain the target reverberation time, target early decay time and target clarity, which are the target environment acoustic parameters.

[0069] Step S205: The server generates correction parameters of the digital equalizer based on the target environment acoustic parameters to obtain target correction parameters;

[0070] In one embodiment, the server extracts sound quality features of the target environment acoustic parameters through a pre-trained sound quality feature extraction model to obtain target sound quality features; the server obtains preset digital equalizer parameters corresponding to the target sound quality features to obtain target digital equalizer parameters; based on the deviation between the preset digital equalizer parameter values ​​corresponding to the target digital equalizer parameters and the corresponding standard digital equalizer parameter values, the server calculates the correction parameter values ​​corresponding to the target digital equalizer parameters to obtain the target digital equalizer parameter values; wherein the preset digital equalizer parameter values ​​are the parameter values ​​corresponding to the target digital equalizer parameters of the audio playback device; the server converts the target digital equalizer parameter values ​​into digital filter parameter values ​​to obtain target correction parameter values.

[0071] In this implementation, the sound quality features in the target environment parameters are first identified through an artificial intelligence model, and then the digital equalizer parameters that need to be adjusted are obtained based on the sound quality features to obtain the target digital equalizer parameters. Next, the parameter values ​​corresponding to the target digital equalizer parameters of the audio playback device are obtained, that is, the preset digital equalizer parameter values ​​corresponding to the target digital equalizer parameters. The difference between the preset digital equalizer parameter values ​​and the corresponding standard digital equalizer parameter values ​​is calculated to obtain the correction parameter values ​​corresponding to the target digital equalizer parameters, that is, the target digital equalizer parameter values. Finally, the target digital equalizer parameter values ​​are converted into parameter values ​​of digital filter parameters to obtain the target correction parameter values ​​that can be directly applied to the audio playback device.

[0072] In this embodiment, the pre-trained sound quality feature extraction model is an artificial intelligence model, wherein the model may include multiple long short-term memory (LSTM) networks and fully connected layers, wherein the LSTM network is used to capture the time dependency of the target environment acoustic parameters, and the fully connected layer is used to classify and output the target sound quality features. The sound quality feature extraction model may also be composed of other networks. Specifically, in one embodiment, the pre-trained sound quality feature extraction model includes a convolutional neural network, a recurrent neural network, and a deep neural network; when the server extracts the sound quality features of the target environment acoustic parameters through the pre-trained sound quality feature extraction model to obtain the target sound quality features, the server extracts the local audio features of the target environment acoustic parameters through the convolutional neural network to obtain the local audio feature information; the server identifies the temporal dependency of the local audio feature information through the recurrent neural network to obtain the temporal audio feature information; the server classifies the sound quality features of the temporal audio feature information through the deep neural network to obtain the target sound quality features.

[0073] In this embodiment, the convolutional neural network (CNN) has the characteristics of weight sharing and local connection, which helps to reduce the number of parameters and the amount of calculation of the model. By performing convolution operations on the convolution kernel in the CNN and the acoustic parameters of the target environment, information such as the reverberation time characteristics, early decay time characteristics and clarity characteristics of the local audio can be obtained; and the recurrent neural network (RNN) can capture the temporal dependencies in the local audio feature information, and can extract the temporal audio feature information of the sound quality characteristics that change over time; the deep neural network (DNN) has a strong nonlinear mapping ability, and can learn and recognize complex mapping relationships. By increasing the number of network layers and neurons of the DNN, the classification performance of the DNN can be improved. By classifying the sound quality characteristics of the temporal audio feature information by the DNN, the target sound quality characteristics can be obtained.

[0074] In this embodiment, different sound quality features are pre-corresponded to different digital equalizer parameters, wherein the target digital equalizer parameters are digital equalizer parameters corresponding to the target sound quality features. In one embodiment, the server obtains the preset digital equalizer parameters corresponding to the target sound quality features. When obtaining the target digital equalizer parameters, the server obtains the preset digital equalizer parameters corresponding to the target sound quality features according to the preset mapping relationship between the sound quality features and the digital equalizer parameters, and obtains the target digital equalizer parameters. For example, if the target sound quality feature is that the reverberation is too low, the gain values ​​of the intermediate and low frequencies can be increased to increase the reverberation time. Then, the digital equalizer parameters corresponding to the target sound quality feature of too low reverberation are the gain values ​​of the intermediate and low frequencies.

[0075] In one implementation, the correction parameter values ​​corresponding to the target digital equalizer parameters may also be pre-bound through a mapping relationship. For example, when the reverberation time is higher than the preset reverberation time of 0s-0.5s, the gain values ​​of the mid-frequency and low-frequency are increased by 1dB. The specifics are not limited here.

[0076] In one embodiment, the server determines the target digital filter corresponding to the target digital equalizer parameter according to the mapping relationship between the preset digital equalizer parameter and the digital filter; the server calculates the coefficient of the target digital filter according to the target digital equalizer parameter value to obtain the target correction parameter. It can be understood that different frequency bands correspond to different digital filters, and the corresponding target digital filter is determined according to the frequency band corresponding to the target digital equalizer parameter, and then the coefficient of the target digital filter corresponding to the target digital equalizer parameter value is calculated through the coefficient calculation function corresponding to the target digital filter to obtain the target correction parameter, that is, the coefficient of the target digital filter can be used to apply the target digital filter to play the audio signal in real-time signal processing.

[0077] Step S206: the server sends the target correction parameter to the terminal, and the terminal transmits the received target correction parameter to the audio playback device;

[0078] The execution process of step S206 is similar to the execution process of step S50, and will not be described in detail here.

[0079] Step S207: The audio playback device calibrates a preset digital equalizer using the received target calibration parameters, and plays audio using the calibrated digital equalizer.

[0080] In one embodiment, the audio playback device uses the received target correction parameters as digital filter parameters and superimposes them with the digital filter parameters in the audio playback channel, and performs audio playback using the superimposed digital filter parameters. In this embodiment, the audio playback device superimposes the target correction parameters into the original audio processing path, and thereafter, all audio signals played by the audio playback device will be corrected by the target correction parameters to adapt to the current ambient acoustic environment.

[0081] The audio correction method provided in the above-mentioned implementation manner plays a test signal and collects ambient sound through an ordinary terminal device with signal playback and sound collection, and then uploads it to a server. The computing power resources of the server are used to quickly and accurately identify the reverberation time, early decay time and clarity of the environment, and generate digital equalizer parameters for the user to correct the acoustic defects of the environment based on these parameters. The correction parameters generated by the server are then sent to the audio playback device through the terminal. The audio playback device applies the correction parameters to calibrate the digital equalizer. Playing audio with the corrected parameters can improve the adaptability of audio playback to the acoustic environment and improve the user's listening experience. The whole process is simple to operate, without the intervention of professionals, and the audio operation cost is also reduced.

[0082] This embodiment also provides an audio correction system, characterized in that the audio correction system includes a terminal, a server and an audio playback device, and the audio correction system executes the above-mentioned audio correction method.

[0083] This embodiment also provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the above-mentioned audio correction method.

[0084] The audio correction method, system and computer program product of the storage medium provided in the embodiments of the present disclosure include a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the previous method embodiments. The specific implementation can be found in the method embodiments, which will not be repeated here.

[0085] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the system and device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0086] In addition, in the description of the embodiments of the present disclosure, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present disclosure can be understood according to specific circumstances.

[0087] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present disclosure, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present disclosure. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.

[0088] In the description of the present disclosure, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present disclosure and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation of the present disclosure. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0089] Finally, it should be noted that the above embodiments are only specific implementation methods of the present disclosure, which are used to illustrate the technical solutions of the present disclosure, rather than to limit them. The protection scope of the present disclosure is not limited thereto. Although the present disclosure is described in detail with reference to the above embodiments, those skilled in the art should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above embodiments within the technical scope disclosed in the present disclosure, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present disclosure, and should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be based on the protection scope of the claims.

Claims

1. A method for correcting audio, characterized in that: A correction system for audio, the correction system for audio comprising a terminal, a server and an audio playback device, the method comprising: The terminal responds to the sound collection instruction, plays at least one test signal, and collects ambient sound to obtain target audio data; The terminal uploads the target audio data to the server; The server performs environmental acoustic parameter recognition on the target audio data to obtain target environmental acoustic parameters; The server extracts the sound quality features of the target environment acoustic parameters through a pre-trained sound quality feature extraction model to obtain target sound quality features; The server obtains the preset digital equalizer parameters corresponding to the target sound quality features according to the preset mapping relationship between the sound quality features and the digital equalizer parameters, and obtains the target digital equalizer parameters; According to the deviation between the preset digital equalizer parameter value corresponding to the target digital equalizer parameter and the corresponding standard digital equalizer parameter value, the correction parameter value corresponding to the target digital equalizer parameter is calculated to obtain the target digital equalizer parameter value; wherein the preset digital equalizer parameter value is the parameter value corresponding to the target digital equalizer parameter of the audio playback device; The server converts the target digital equalizer parameter value into a digital filter parameter value to obtain a target correction parameter value; The server sends the target correction parameter to the terminal, and the terminal transmits the received target correction parameter to the audio playback device; The audio playback device calibrates a preset digital equalizer using the received target correction parameters, and plays audio using the calibrated digital equalizer.

2. The method according to claim 1, characterized in that The terminal responds to the sound collection instruction, plays at least one test signal, and collects ambient sound to obtain target audio data, including: The terminal responds to the sound collection instruction, obtains at least one pre-generated test signal, and plays the at least one test signal in a preset order; While playing the at least one test signal, the terminal records the ambient sound; When the at least one test signal is played, the terminal saves the recorded ambient sound as target audio data.

3. The method according to claim 1, characterized in that The at least one test signal includes an impulse signal, a logarithmic swept sine signal and pink noise; the target audio data includes a response signal corresponding to each test signal; The server performs environmental acoustic parameter recognition on the target audio data to obtain the target environmental acoustic parameter, comprising: The server converts the response signal corresponding to each test signal into a pulse signal to obtain a pulse signal corresponding to each test signal; The server calculates the reverberation time, early decay time and clarity of the pulse signal corresponding to each test signal to obtain the target environment acoustic parameters.

4. The method according to claim 3, characterized in that: The server converts the response signal corresponding to each test signal into a pulse signal to obtain the pulse signal corresponding to each test signal, including: The server determines the response signal corresponding to the impact signal as a pulse signal corresponding to the impact signal; The server extracts a pulse signal from a response signal corresponding to the logarithmic swept frequency sine signal to obtain a pulse signal corresponding to the logarithmic swept frequency sine signal; The server performs cross correlation on the pink noise and a response signal corresponding to the pink noise to obtain a pulse signal corresponding to the pink noise.

5. The method according to claim 4, characterized in that The server extracts a pulse signal from a response signal corresponding to the logarithmic swept frequency sine signal to obtain a pulse signal corresponding to the logarithmic swept frequency sine signal, comprising: The server performs Fourier transform on the response signal corresponding to the logarithmic swept frequency sine signal to obtain a first frequency response curve; The server filters the frequency domain characteristics of the logarithmic swept frequency sine signal on the first frequency response curve through a preset frequency domain filter to obtain a second frequency response curve; The server performs inverse filtering on the second frequency response curve to obtain a pulse signal corresponding to the logarithmic swept frequency sine signal.

6. The method according to claim 3, characterized in that The server calculates the reverberation time, early decay time and clarity of the pulse signal corresponding to each test signal to obtain the acoustic parameters of the target environment, including: The server calculates the energy attenuation curve of the pulse signal corresponding to each test signal, the energy before a preset time, and the total energy according to the pulse signal corresponding to each test signal; The server calculates the time required for energy decay to a first preset value according to the energy decay curve corresponding to each test signal, and obtains the reverberation time corresponding to each test signal; The server calculates the time required for the energy to decay from the maximum value to the second preset value according to the energy decay curve corresponding to each test signal, and obtains the early decay time corresponding to each test signal; The server calculates the clarity corresponding to each test signal according to the ratio between the energy before the preset time corresponding to each test signal and the total energy; The server obtains the target environment acoustic parameters by combining the reverberation time, early decay time and clarity corresponding to each test signal.

7. The method according to claim 6, characterized in that The server obtains the target environment acoustic parameters by combining the reverberation time, early decay time and clarity corresponding to each test signal, including: The server performs weighted averaging on the reverberation time, early decay time and clarity corresponding to each test signal to obtain the target environment acoustic parameters.

8. The method according to claim 1, characterized in that The pre-trained sound quality feature extraction model includes a convolutional neural network, a recurrent neural network and a deep neural network; The server extracts the sound quality features of the target environment acoustic parameters through a pre-trained sound quality feature extraction model to obtain the target sound quality features, including: The server extracts local audio features from the target environment acoustic parameters through the convolutional neural network to obtain local audio feature information; The server identifies the temporal dependency relationship of the local audio feature information through the recurrent neural network to obtain temporal audio feature information; The server performs sound quality feature classification on the time series audio feature information through the deep neural network to obtain target sound quality features.

9. The method according to claim 1, characterized in that: The server converts the target digital equalizer parameter value into a digital filter parameter value to obtain a target correction parameter value, comprising: The server determines the target digital filter corresponding to the target digital equalizer parameter according to the mapping relationship between the preset digital equalizer parameter and the digital filter; The server calculates the coefficients of the target digital filter according to the target digital equalizer parameter values ​​to obtain target correction parameters.

10. The method according to claim 1, characterized in that The step of the audio playback device correcting a preset digital equalizer by using the received target correction parameter and playing audio by using the corrected digital equalizer comprises: The audio playback device uses the received target correction parameters as digital filter parameters and superimposes them with the digital filter parameters in the audio playback channel, and performs audio playback using the superimposed digital filter parameters.

11. An audio correction system, characterized in that: The audio correction system comprises a terminal, a server and an audio playback device, and the audio correction system executes the audio correction method as described in any one of claims 1-10.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the audio correction method according to any one of claims 1 to 10.

Images