Reverberation method and device, vehicle machine, storage medium and program product
By using the learning model in the Moorer model reverb algorithm to obtain reverb parameters and adaptively adjust the number of reverb delay arrays and comb filters, the problem of insufficient real-time performance of the Moorer model reverb algorithm in car karaokes is solved, achieving lower reverb delay and higher real-time performance.
Patent Information
- Application Number
- CN202510028650.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-08
- Publication Date
- 2025-04-11
AI Technical Summary
The existing Moorer model reverb algorithm lacks real-time performance during the on-board karaoke process, resulting in large reverb delays and high computational complexity, which cannot meet the needs of high-quality sound effects.
By inputting the audio signal to be processed into the trained learning model, obtaining the corresponding reverb parameters, and adaptively adjusting the number of reverb delay arrays and comb filters involved in the processing in the Moorer reverb model based on these parameters, reducing unnecessary processing volume and improving real-time performance.
It improves the real-time performance of the Moorer reverb model, reduces reverb delay, reduces the occupation of vehicle resources, realizes the reverb processing on the vehicle side, and reduces the dependence on cloud servers.
Smart Images
Figure CN119811353A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of audio processing, and in particular to a reverberation method, a device, a vehicle computer, a storage medium and a program product. Background Art
[0002] In the car cabin, the application scenarios of reverberation technology are mainly focused on providing a better and more personalized audio experience. The following are several major application scenarios:
[0003] (1) Multimedia entertainment: In the in-vehicle multimedia entertainment system, reverberation technology can enhance the layering and spatial sense of music, bringing passengers a more immersive music experience. For example, when playing music, the reverberation effect can simulate the echo and reverberation of music in different spatial environments, making the music sound more three-dimensional and full.
[0004] (2) Navigation broadcast: When driving, the mixing of navigation broadcast sound with music or other media sound is a common challenge. Reverberation technology can optimize the mixing effect of navigation broadcast and music, making the two independent and non-interfering with each other, ensuring the clarity of navigation information without affecting the passengers' music experience.
[0005] (3) Call privacy protection: When making calls in the car, reverberation technology can be used to protect the privacy of the call content. For example, by setting the sound emission position, the call sound can be heard from the driver's headrest speaker, while other passengers can continue to listen to music, thus achieving the privacy of the call.
[0006] (4) Noise optimization: During driving, noises such as tire noise and horns may affect the clarity of the sound of the vehicle speakers. Reverberation technology can improve these problems to a certain extent by optimizing the propagation and reflection of sound, reducing the impact of noise on audio quality, and improving the passenger experience.
[0007] In addition, with the development of technology, the introduction of in-vehicle spatial audio solutions has also provided more possibilities for the application of reverberation technology in the vehicle cockpit. By combining the sound field layout of the cockpit with professional adjustments, a more accurate and personalized audio experience can be achieved. In short, the application of reverberation technology in the vehicle cockpit is aimed at enhancing the audio experience of passengers, optimizing the sound quality, and meeting different usage needs. With the continuous advancement and innovation of technology, the application of reverberation technology in the automotive field will continue to expand and deepen in the future.
[0008] At present, the entertainment functions in the car cabin are becoming more and more abundant, and car companies are also constantly breaking through and updating. With the continuous upgrading and improvement of the in-car entertainment system, the karaoke function has gradually become a feature of modern cars, and reverberation technology is the key to achieving a high-quality karaoke experience.
[0009] Digital signal processing technology plays a core role in the reverberation of in-car karaoke. It can process and analyze the collected data in real time and adjust the reverberation effect of the audio through algorithms. These algorithms can accurately adjust the reverberation parameters according to factors such as the acoustic environment inside the car, the position of the microphone, and the user's preferences to achieve the best sound effect.
[0010] Audio engineering also provides a theoretical basis for in-car karaoke reverberation. It studies the generation, propagation and reception of sound, as well as its processing and conversion. In in-car karaoke, audio engineers need to comprehensively consider factors such as the spatial layout of the car, the sound absorption performance of the materials, and the performance of the audio equipment to design a suitable reverberation solution to create a sound environment suitable for karaoke.
[0011] Among the existing reverberation algorithms, the more popular algorithms are convolution method, comb filter reverberation model, Schroeder reverberation model and Moorer model reverberation. Among them, the convolution method requires the design of the impact response of the size and material of the interior space in advance, and is rarely used in actual scenarios. The comb filter reverberation model has a relatively low echo density, so it will have a metallic texture, which sounds uncomfortable and unnatural. The Schroeder reverberation model needs to be improved by using an all-pass filter to solve the problems of uneven spectrum and insufficient echo density. Moorer reverberation is optimized based on the former, considering the high-frequency attenuation factor of the signal propagating in space, adding a low-pass filter to attenuate the high frequency, and nesting the all-pass filter behind it, which has better group delay. However, the implementation of the reverberation algorithm during car karaoke also needs to consider factors such as real-time performance and computational complexity to ensure that a smooth and high-quality sound experience can be provided during karaoke. The current Moorer model reverberation real-time performance needs to be improved. Summary of the invention
[0012] In view of this, the present invention provides a reverberation method, device, vehicle computer, storage medium and program product to improve the real-time performance of Moorer model reverberation.
[0013] In a first aspect, the present invention provides a reverberation method, comprising the following steps: obtaining an audio signal to be processed; inputting the audio signal to be processed into a trained learning model to obtain reverberation parameters corresponding to the audio signal to be processed; inputting the reverberation parameters into a preset Moorer reverberation model to obtain a first array set, wherein the number of reverberation delay arrays in the first array set is a first number; determining the reverberation delay arrays participating in the reverberation processing according to the first number and the first array set to obtain a second array set, wherein the number of reverberation delay arrays in the second array set is a second number, and the second number is less than or equal to the first number; determining the comb filter participating in the reverberation processing in the Moorer reverberation model according to the second number to obtain a target comb filter; and processing the second array set using the target comb filter to obtain a target reverberation signal.
[0014] The present invention provides a method of inputting an audio signal to be processed into a trained learning model to obtain reverberation parameters corresponding to the audio signal to be processed, and inputting the reverberation parameters into a Moorer reverberation model to obtain a first array set; further, determining a reverberation delay array participating in reverberation processing according to a first number of reverberation delay arrays in the first array set to obtain a second array set, wherein the number of reverberation delay arrays in the second array set is a second number; and determining a comb filter participating in reverberation processing in the Moorer reverberation model according to the second number to obtain a target comb filter, and using the target comb filter to process the second array set to obtain a target reverberation signal. That is to say, the present invention obtains reverberation parameters corresponding to the audio signal to be processed by inputting the audio signal to be processed into a trained learning model, and inputs the reverberation parameters into the Moorer reverberation model to obtain a first array set; further, the number of comb filters participating in reverberation processing in the Moorer reverberation model can be adaptively adjusted according to the first number of reverberation delay arrays in the first array set. By adaptively adjusting the number of target comb filters, unnecessary processing volume can be reduced, the real-time performance of the Moorer reverberation model can be improved, and the reverberation delay can be reduced. At the same time, the resource occupation of the vehicle computer is relatively small.
[0015] In an optional implementation, determining the reverberation delay array participating in the reverberation processing according to the first quantity and the first array set to obtain the second array set includes: judging whether the first quantity is greater than a preset first threshold; when the first quantity is greater than the first threshold, screening out some reverberation delay arrays in the first array set to obtain the second array set; when the first quantity is less than or equal to the first threshold, using the first array set as the second array set.
[0016] This can reduce the number of comb filters involved in reverberation processing in the Moorer reverberation model, reduce the computational complexity to a certain extent, and improve the real-time performance of the Moorer reverberation model.
[0017] In an optional embodiment, filtering out some reverberation delay arrays in the first array set to obtain the second array set includes: eliminating the last reverberation delay array from multiple reverberation delay arrays to obtain a third array set, the number of reverberation delay arrays in the third array set being a third number; determining whether the third number is greater than a preset first threshold; when the third number is greater than the first threshold, filtering out a second number of reverberation delay arrays from the third array set according to the delay time of the reverberation delay arrays; and when the third number is equal to the first threshold, using the third array set as the second array set.
[0018] This can make the selection of the reverberation delay array more representative, thereby making the obtained target reverberation signal effect better.
[0019] In an optional implementation, determining the comb filters participating in reverberation processing in the Moorer reverberation model according to the second quantity to obtain the target comb filters includes: selecting the second number of comb filters in the Moorer reverberation model to obtain the target comb filters.
[0020] This achieves adaptive adjustment of the target comb filter quantity.
[0021] In an optional embodiment, the second array set is processed using a target comb filter to obtain a target reverberation signal, including: inputting each reverberation delay array in the second array set into different target number comb filters to obtain multiple sampling points; averaging the multiple sampling points to obtain an average sampling point; and obtaining the target reverberation signal based on the average sampling point.
[0022] Therefore, the target reverberation signal can be obtained by using the determined target comb filter.
[0023] In an optional implementation, before inputting the audio signal to be processed into the trained learning model, the method further includes: converting the audio signal to be processed into the frequency domain to obtain a frequency domain signal, and extracting a mid-frequency band portion of the frequency domain signal.
[0024] This is because, according to the characteristics of human voice signals, the audio signal to be processed is relatively rich in mid- and low-frequency parts during singing, so only the mid- and low-frequency parts are extracted. For example, the extracted frequency band range is below 3000hz.
[0025] In an optional embodiment, obtaining a target reverberation signal according to the average sampling point includes: inputting the average sampling point into an all-pass filter of a Moorer reverberation model for processing to obtain an initial reverberation signal; multiplying the initial reverberation signal by a preset wet sound gain to obtain a wet reverberation signal; multiplying the audio signal to be processed by a preset dry sound gain to obtain a dry reverberation signal; and adding the wet reverberation signal to the dry reverberation signal to obtain a target reverberation signal.
[0026] This can improve the echo density and spectral flatness.
[0027] In a second aspect, the present invention further provides a reverberation device, including an acquisition module, a reverberation parameter determination module, a reverberation delay array determination module, a reverberation delay array processing module, a target comb filter determination module and a post-processing module; the acquisition module is used to acquire an audio signal to be processed; the reverberation parameter determination module is used to input the audio signal to be processed into a trained learning model to obtain reverberation parameters corresponding to the audio signal to be processed; the reverberation delay array determination module is used to input the reverberation parameters into a preset Moorer reverberation model to obtain a first array set, wherein the number of reverberation delay arrays in the first array set is a first number; the reverberation delay array processing module is used to determine the reverberation delay arrays participating in the reverberation processing according to the first number and the first array set to obtain a second array set, wherein the number of reverberation delay arrays in the second array set is a second number, and the second number is less than or equal to the first number; the target comb filter determination module is used to determine the comb filters participating in the reverberation processing in the Moorer reverberation model according to the second number to obtain the target comb filter; the post-processing module is used to process the second array set using the target comb filter to obtain the target reverberation signal.
[0028] In a third aspect, the present invention further provides a vehicle computer, comprising a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the reverberation method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.
[0029] In a fourth aspect, the present invention further provides a vehicle, comprising the vehicle computer of the third aspect.
[0030] In a fifth aspect, the present invention further provides a computer program product, comprising computer instructions, wherein the computer instructions are used to enable a computer to execute the reverberation method of the first aspect or any corresponding embodiment thereof.
[0031] The reverberation method, device, vehicle computer, storage medium and program product provided by the present invention have the following beneficial effects: by inputting the audio signal to be processed into a trained learning model, the reverberation parameters corresponding to the audio signal to be processed are obtained, and the reverberation parameters are input into the Moorer reverberation model to obtain a first array set; further, the number of comb filters participating in the reverberation processing in the Moorer reverberation model can be adaptively adjusted according to the first number of the reverberation delay array in the first array set, and the number of target comb filters can be adaptively adjusted to reduce unnecessary processing, improve the real-time performance of the Moorer reverberation model, reduce reverberation delay, and at the same time, the resource occupation of the vehicle computer is relatively small. In the related art, since the reverberation algorithm has a large processing volume and the general vehicle computer cannot meet the operating requirements, it needs to be processed on the cloud server; and the reverberation method of the present invention can reduce unnecessary processing volume, so that the reverberation method can be completed on the vehicle computer side, reducing the dependence of the reverberation method on the cloud server. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0033] Figure 1 is a flow chart of a reverberation method according to an embodiment of the present invention;
[0034] Figure 2 is a flow chart of another reverberation method according to an embodiment of the present invention;
[0035] Figure 3 is a flow chart of an example of a reverberation method according to an embodiment of the present invention;
[0036] Figure 4 is a schematic diagram of the structure of a learning model according to an embodiment of the present invention;
[0037] Figure 5 is the waveform of the audio signal to be processed;
[0038] Figure 6 is the waveform of the target reverberation signal;
[0039] Figure 7 is a structural block diagram of a reverberation device according to an embodiment of the present invention;
[0040] Figure 8 It is a schematic diagram of the hardware structure of the vehicle computer according to an embodiment of the present invention. DETAILED DESCRIPTION
[0041] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.
[0042] According to an embodiment of the present invention, a reverberation method embodiment is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0043] In this embodiment, a reverberation method is provided, which can be used in a vehicle computer. Figure 1 is a flow chart of a reverberation method according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:
[0044] Step S101: obtaining an audio signal to be processed.
[0045] Step S102: inputting the audio signal to be processed into a trained learning model to obtain reverberation parameters corresponding to the audio signal to be processed.
[0046] Specifically, according to the requirements of the algorithm in the Moorer reverberation model, a suitable machine learning model, such as a neural network, a support vector machine (SVM) or a decision tree, can be selected to construct the input and output layers of the model to ensure that the model can accept audio features and output adaptive reverberation parameters. For example, the reverberation parameters can be pre-delay, reverberation time, reflection coefficient, etc.
[0047] Step S103: inputting the reverberation parameters into a preset Moorer reverberation model to obtain a first array set, wherein the number of reverberation delay arrays in the first array set is a first number.
[0048] That is, the reverberation parameters are input into a preset Moorer reverberation model to obtain a plurality of reverberation delay arrays, each reverberation data group includes parameter values corresponding to the above reverberation parameters, and the plurality of reverberation delay arrays constitute a first array set.
[0049] Step S104: determining the reverberation delay arrays participating in the reverberation processing according to the first number and the first array set to obtain a second array set, wherein the number of reverberation delay arrays in the second array set is a second number, which is less than or equal to the first number.
[0050] Step S105: determining the comb filter involved in the reverberation processing in the Moorer reverberation model according to the second quantity, and obtaining a target comb filter.
[0051] Step S106: Process the second array set using a target comb filter to obtain a target reverberation signal.
[0052] The reverberation method provided in this embodiment inputs the audio signal to be processed into a trained learning model to obtain reverberation parameters corresponding to the audio signal to be processed, and inputs the reverberation parameters into a Moorer reverberation model to obtain a first array set; further, the reverberation delay arrays participating in the reverberation processing are determined according to a first number of reverberation delay arrays in the first array set to obtain a second array set, wherein the number of reverberation delay arrays in the second array set is a second number; and the comb filter participating in the reverberation processing in the Moorer reverberation model can be determined according to the second number to obtain a target comb filter, and the second array set is processed using the target comb filter to obtain a target reverberation signal. That is to say, in this embodiment, the reverberation parameters corresponding to the audio signal to be processed are obtained by inputting the audio signal to be processed into a trained learning model, and the reverberation parameters are input into the Moorer reverberation model to obtain a first array set; further, the number of comb filters participating in reverberation processing in the Moorer reverberation model can be adaptively adjusted according to the first number of reverberation delay arrays in the first array set. By adaptively adjusting the number of target comb filters, unnecessary processing volume can be reduced, the real-time performance of the Moorer reverberation model can be improved, and the reverberation delay can be reduced. At the same time, the resource occupation of the vehicle computer is relatively small.
[0053] In this embodiment, a reverberation method is provided, which can be used in a vehicle computer. Figure 2 is a flow chart of another reverberation method according to an embodiment of the present invention, Figure 3 is a flow chart of an example of a reverberation method according to an embodiment of the present invention. Figure 2 and Figure 3 As shown, the process includes the following steps:
[0054] Step S201: Acquire an audio signal to be processed.
[0055] Step S202: inputting the audio signal to be processed into a trained learning model to obtain reverberation parameters corresponding to the audio signal to be processed.
[0056] In this embodiment, the learning module is used to extract relevant features from the input audio signal to be processed, such as spectral characteristics, energy distribution, time structure, etc. These features are used to evaluate the current audio environment and user needs in order to adjust the reverberation parameters. In fact, the reverberation effect is affected by many factors, such as room size, shape, material, and sound source location. In this embodiment, by inputting the audio signal to be processed into the trained learning model, the reverberation parameters can be automatically adjusted according to the characteristics of the input signal to obtain a better sound effect.
[0057] Figure 4 Schematic diagram of the structure of the learning model according to an embodiment of the present invention. Figure 4 As shown, the learning model includes a CNN model, where the first layer includes a 2D convolution consisting of 15 convolution kernels, an activation function and a pooling layer; the second layer includes a 2D convolution consisting of 15 convolution kernels, an activation function and a pooling layer; the third layer includes a 2D convolution consisting of 15 convolution kernels, an activation function and a pooling layer; the fourth and fifth layers are LSTM models; the sixth layer is a fully connected layer, which is used to generate corresponding output information after reducing the dimensionality of the multi-dimensional vector.
[0058] Specifically, the learning model can be obtained by the following method. Collect a set of representative audio samples that can cover different sound sources, environmental conditions and reverberation effects, annotate the audio samples, and provide the desired reverberation effect or parameters as the training target. According to the evaluation results, adjust and optimize the model, such as increasing the number of network layers, changing the activation function, etc. By adjusting the parameters and weights of the model, the model can gradually learn the mapping relationship from audio features to reverberation parameters. Supervised learning methods can be used to train by minimizing the error between the predicted reverberation parameters and the actual reverberation parameters. In addition, in order to avoid overfitting and improve the generalization ability of the model, optimization algorithms (such as gradient descent, Adam, etc.) and regularization techniques (such as L1, L2 regularization) can be introduced. The training process of the model is optimized by adjusting hyperparameters such as learning rate and batch size.
[0059] In addition, it is necessary to select a suitable adaptive control algorithm in the learning model, such as the least mean square error (LMS) algorithm, the recursive least squares (RLS) algorithm, the neural network, etc., to calculate and adjust the reverberation parameters in real time according to the extracted features and the output error of the algorithm; it is necessary to design the rules and methods for parameter updating to ensure the convergence and stability of the algorithm. According to the calculation results of the adaptive control algorithm, the reverberation parameters are gradually adjusted to reduce the output error and optimize the sound effect.
[0060] After the learning model is trained, it can be tested in actual application scenarios to verify the performance and stability of the model, and make necessary adjustments and optimizations based on feedback. The performance and stability of the algorithm can be tested in different audio environments and conditions, and subjective and objective evaluation methods can be used to verify whether the reverberation effect meets user needs. It should be noted that the training process may require a lot of computing resources and time, especially when dealing with large-scale audio data sets and complex machine learning models. In addition, the quality of the training results is also affected by many factors such as data set quality, model selection, and hyperparameter adjustment. Therefore, sufficient experiments and debugging are required during the training process to obtain satisfactory performance.
[0061] Step S203: inputting the reverberation parameters into a preset Moorer reverberation model to obtain a first array set, wherein the number of reverberation delay arrays in the first array set is a first number.
[0062] Step S204: determining the reverberation delay arrays participating in the reverberation processing according to the first number and the first array set to obtain a second array set, wherein the number of reverberation delay arrays in the second array set is a second number, which is less than or equal to the first number.
[0063] In an optional implementation, determining the reverberation delay array involved in the reverberation processing according to the first quantity and the first array set to obtain the second array set includes the following steps:
[0064] Step S2041: determine whether the first number is greater than a preset first threshold; when the first number is greater than the first threshold, proceed to step S2042; otherwise, proceed to step S2043. For example, the first threshold may be 5.
[0065] Step S2042: Filter out some reverberation delay arrays from the first array set to obtain a second array set.
[0066] Specifically, selecting a part of the reverberation delay arrays from the first array set to obtain the second array set includes the following steps a1 to a4.
[0067] Step a1: Eliminate the last reverberation delay array from a plurality of reverberation delay arrays to obtain a third array set, wherein the number of reverberation delay arrays in the third array set is a third number.
[0068] Step a2: Determine whether the third number is greater than a preset first threshold; when the third number is greater than the first threshold, proceed to step a3; when the third number is equal to the first threshold, proceed to step a4.
[0069] Step a3: uniformly selecting a second number of reverberation delay arrays from the third array set according to the delay times of the reverberation delay arrays.
[0070] For example, reverberation delay arrays with even order may be selected from the third array set to obtain a second number of reverberation delay arrays.
[0071] Step a4: Use the third array set as the second array set.
[0072] Step S2043: Use the first array set as the second array set.
[0073] Step S205: Determine the comb filter (Comb) involved in the reverberation processing in the Moorer reverberation model according to the second quantity, and obtain a target comb filter.
[0074] Specifically, a second number of comb filters are selected in the Moorer reverberation model to obtain a target comb filter.
[0075] Step S206: Process the second array set using a target comb filter to obtain a target reverberation signal.
[0076] In an optional implementation, processing the second array set by using a target comb filter to obtain a target reverberation signal includes the following steps:
[0077] Step S2061: input each reverberation delay array in the second array set into different target number comb filters to obtain multiple sampling points.
[0078] Step S2062: Calculate the average value of multiple sampling points to obtain an average sampling point.
[0079] Step S2063: Obtain a target reverberation signal according to the average sampling point.
[0080] Specifically, obtaining the target reverberation signal according to the average sampling point includes: inputting the average sampling point into an all-pass filter of a Moorer reverberation model for processing to obtain an initial reverberation signal; multiplying the initial reverberation signal by a preset wet sound gain to obtain a wet reverberation signal; multiplying the audio signal to be processed by a preset dry sound gain to obtain a dry reverberation signal; and adding the wet reverberation signal to the dry reverberation signal to obtain the target reverberation signal.
[0081] This is because the all-pass filter has a good group delay, so it can increase the density of the echo. For example, the all-pass filter is designed as follows: y[n] = -gx[n] + x[nm] + gy[nm]; where x is a vector representing a set of vectors from 0 to n-1, n represents the index of a subscript of this vector, m represents the delay, x(n) represents the current sampling point, x(nm) represents the reflected signal delayed by m sampling points, and y(nm) represents the reflected signal delayed by m output sampling points.
[0082] Specifically, the wet sound gain and the dry sound gain are both between 0 and 1, and the dry sound gain: the wet sound gain = 0.7:0.9.
[0083] For example, Figure 5 is the waveform of the audio signal to be processed. Figure 6 It is the waveform of the target reverberation signal.
[0084] The present embodiment provides a method of inputting a to-be-processed audio signal into a trained learning model to obtain reverberation parameters corresponding to the to-be-processed audio signal, and inputting the reverberation parameters into a Moorer reverberation model to obtain a first array set; further, determining the reverberation delay arrays involved in the reverberation processing according to the first number of reverberation delay arrays in the first array set to obtain a second array set, wherein the number of reverberation delay arrays in the second array set is a second number; and determining the comb filters involved in the reverberation processing in the Moorer reverberation model according to the second number to obtain a target comb filter, and using the target comb filter to process the second array set to obtain a target reverberation signal. That is, the present embodiment inputs the to-be-processed audio signal into a trained learning model to obtain reverberation parameters corresponding to the to-be-processed audio signal, and inputs the reverberation parameters into a Moorer reverberation model to obtain a first array set; further, the number of comb filters involved in the reverberation processing in the Moorer reverberation model can be adaptively adjusted according to the first number of reverberation delay arrays in the first array set to improve the real-time performance of the Moorer reverberation.
[0085] In this embodiment, a reverberation device is also provided, which is used to implement the above embodiments and preferred implementations, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.
[0086] This embodiment provides a reverberation device, such as Figure 7 As shown, including:
[0087] The acquisition module 701 is used to acquire the audio signal to be processed.
[0088] The reverberation parameter determination module 702 is used to input the audio signal to be processed into a trained learning model to obtain the reverberation parameter corresponding to the audio signal to be processed.
[0089] The reverberation delay array determination module 703 is used to input the reverberation parameters into a preset Moorer reverberation model to obtain a first array set, wherein the number of reverberation delay arrays in the first array set is a first number.
[0090] The reverberation delay array processing module 704 is used to determine the reverberation delay array participating in the reverberation processing according to the first number and the first array set, and obtain a second array set, wherein the number of reverberation delay arrays in the second array set is a second number, and the second number is less than or equal to the first number.
[0091] The target comb filter determination module 705 is used to determine the comb filter involved in the reverberation processing in the Moorer reverberation model according to the second quantity to obtain the target comb filter.
[0092] The post-processing module 706 is used to process the second array set using a target comb filter to obtain a target reverberation signal.
[0093] In an optional implementation, the reverberation delay array processing module 704 includes a judgment unit, a first processing unit, and a second processing unit. The judgment unit is used to judge whether the first number is greater than a preset first threshold; the first processing unit is used to select part of the reverberation delay arrays in the first array set to obtain a second array set when the first number is greater than the first threshold; the second processing unit is used to use the first array set as the second array set when the first number is less than or equal to the first threshold.
[0094] In an optional implementation, the first processing unit is specifically used to: eliminate the last reverberation delay array from multiple reverberation delay arrays to obtain a third array set, the number of reverberation delay arrays in the third array set being a third number; determine whether the third number is greater than a preset first threshold; when the third number is greater than the first threshold, filter out a second number of reverberation delay arrays from the third array set according to the delay time of the reverberation delay arrays; when the third number is equal to the first threshold, use the third array set as the second array set.
[0095] In an optional implementation, the target comb filter determination module 705 is specifically configured to: select a second number of comb filters in the Moorer reverberation model to obtain a target comb filter.
[0096] In an optional implementation, the post-processing module 706 is specifically used to: input each reverberation delay array in the second array set into different target number comb filters to obtain multiple sampling points; calculate the average value of the multiple sampling points to obtain an average sampling point; and obtain a target reverberation signal based on the average sampling point.
[0097] In an optional embodiment, the reverberation device further includes a pre-processing module. Before the audio signal to be processed is input into the trained learning model, the pre-processing module is used to convert the audio signal to be processed into the frequency domain to obtain a frequency domain signal and extract the mid-frequency band part of the frequency domain signal.
[0098] In an optional embodiment, obtaining a target reverberation signal according to the average sampling point includes: inputting the average sampling point into an all-pass filter of a Moorer reverberation model for processing to obtain an initial reverberation signal; multiplying the initial reverberation signal by a preset wet sound gain to obtain a wet reverberation signal; multiplying the audio signal to be processed by a preset dry sound gain to obtain a dry reverberation signal; and adding the wet reverberation signal to the dry reverberation signal to obtain a target reverberation signal.
[0099] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.
[0100] The reverberation device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.
[0101] The embodiment of the present invention further provides a vehicle computer having the above Figure 8 The reverberation device shown.
[0102] An embodiment of the present invention further provides a vehicle, comprising the above-mentioned vehicle computer.
[0103] See also Figure 8 , Figure 8 is a structural schematic diagram of a vehicle computer provided by an optional embodiment of the present invention, such as Figure 8As shown, the vehicle computer includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. The various components are connected to each other using different buses for communication, and can be installed on a common motherboard or installed in other ways as needed. The processor can process instructions executed in the vehicle computer, including instructions stored in or on the memory to display graphical information of the GUI on an external input / output device (such as a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple vehicle computers can be connected, and each device provides part of the necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 8 A processor 10 is taken as an example.
[0104] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.
[0105] The memory 20 stores instructions executable by at least one processor 10, so that at least one processor 10 executes the method shown in the above embodiment.
[0106] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and applications required for at least one function; the data storage area may store data created according to the use of the vehicle computer, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the vehicle computer via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.
[0107] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.
[0108] The vehicle computer also includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 may be connected via a bus or other means. Figure 8The example of connecting through bus is taken in the following.
[0109] The input device 30 can receive input digital or character information, and generate key signal input related to the user settings and function control of the vehicle machine, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator rod, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 may include a display device, an auxiliary lighting device (e.g., an LED) and a tactile feedback device (e.g., a vibration motor), etc. The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display and a plasma display. In some optional embodiments, the display device may be a touch screen.
[0110] The vehicle computer further includes a communication interface 30 for the vehicle computer to communicate with other devices or a communication network.
[0111] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.
[0112] A part of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the existence of the computer program instruction in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc., and accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium accessible to the computer.
[0113] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.
Claims
1. A reverberation method, characterized in that: include: Obtaining the audio signal to be processed; Inputting the audio signal to be processed into a trained learning model to obtain a reverberation parameter corresponding to the audio signal to be processed; Inputting the reverberation parameters into a preset Moorer reverberation model to obtain a first array set, wherein the number of reverberation delay arrays in the first array set is a first number; Determine a reverberation delay array participating in reverberation processing according to the first number and the first array set, to obtain a second array set, wherein the number of the reverberation delay arrays in the second array set is a second number, and the second number is less than or equal to the first number; Determine a comb filter involved in reverberation processing in the Moorer reverberation model according to the second quantity, and obtain a target comb filter; The second array set is processed by using the target comb filter to obtain a target reverberation signal.
2. The method according to claim 1, characterized in that The step of determining a reverberation delay array involved in reverberation processing according to the first number and the first array set to obtain a second array set includes: Determining whether the first number is greater than a preset first threshold; When the first number is greater than the first threshold, selecting part of the reverberation delay arrays from the first array set to obtain the second array set; When the first number is less than or equal to the first threshold, the first array set is used as the second array set.
3. The method according to claim 2, characterized in that The step of selecting part of the reverberation delay arrays from the first array set to obtain the second array set includes: Eliminate the last reverberation delay array from the multiple reverberation delay arrays to obtain a third array set, wherein the number of the reverberation delay arrays in the third array set is a third number; Determining whether the third number is greater than a preset first threshold; When the third number is greater than the first threshold, uniformly screening out the second number of reverberation delay arrays from the third array set according to the delay times of the reverberation delay arrays; When the third number is equal to the first threshold, the third array set is used as the second array set.
4. The method according to claim 1, characterized in that: Determining the comb filter involved in reverberation processing in the Moorer reverberation model according to the second quantity to obtain a target comb filter includes: The second number of comb filters is selected in the Moorer reverberation model to obtain the target comb filter.
5. The method according to claim 1, characterized in that The using the target comb filter to process the second array set to obtain a target reverberation signal comprises: Input each of the reverberation delay arrays in the second array set into different target number comb filters to obtain a plurality of sampling points; Calculate the average value of the plurality of sampling points to obtain an average sampling point; A target reverberation signal is obtained according to the average sampling point.
6. The method according to claim 1, characterized in that Before inputting the audio signal to be processed into the trained learning model, the method further includes: The audio signal to be processed is converted into the frequency domain to obtain a frequency domain signal, and a mid-frequency band portion of the frequency domain signal is extracted.
7. The method according to claim 5, characterized in that The obtaining of the target reverberation signal according to the average sampling point comprises: Inputting the average sampling point into the all-pass filter of the Moorer reverberation model for processing to obtain an initial reverberation signal; Multiplying the initial reverberation signal by a preset wet sound gain to obtain a wet reverberation signal; Multiplying the audio signal to be processed by a preset dry sound gain to obtain a dry reverberation signal; The wet reverberation signal and the dry reverberation signal are added to obtain the target reverberation signal.
8. A reverberation device, characterized in that: include: An acquisition module, used for acquiring an audio signal to be processed; A reverberation parameter determination module, used for inputting the audio signal to be processed into a trained learning model to obtain a reverberation parameter corresponding to the audio signal to be processed; a reverberation delay array determination module, configured to input the reverberation parameters into a preset Moorer reverberation model to obtain a first array set, wherein the number of reverberation delay arrays in the first array set is a first number; a reverberation delay array processing module, configured to determine a reverberation delay array participating in reverberation processing according to the first number and the first array set, to obtain a second array set, wherein the number of the reverberation delay arrays in the second array set is a second number, and the second number is less than or equal to the first number; a target comb filter determination module, configured to determine a comb filter involved in reverberation processing in the Moorer reverberation model according to the second quantity, to obtain a target comb filter; A post-processing module is used to process the second array set using the target comb filter to obtain a target reverberation signal.
9. A vehicle computer, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the reverberation method according to any one of claims 1 to 6 by executing the computer instructions.
10. A vehicle, characterized in that: Including the vehicle computer as described in claim 9.
11. A computer program product, characterized in that The method comprises computer instructions, wherein the computer instructions are used to cause a computer to execute the reverberation method according to any one of claims 1 to 6.