Audio signal processing method and device, equipment, storage medium and product

Abstract

The invention discloses an audio signal processing method and device, equipment, a storage medium and a product, and belongs to the technical field of signal processing. According to the method, the input first signal can be acquired, and the audio signal type of the first signal is identified, so that the audio signal type of the audio signal included in the first signal is analyzed. The dynamic range control curves of the audio signals under different audio signal types can be the same or different, and are customized by fully considering the characteristics of the audio signal types in energy distribution and playing performance, so that a better dynamic range control effect can be achieved. For each related audio signal type included in the first signal, extracting a corresponding characteristic dynamic range control curve to obtain a target dynamic range control curve corresponding to the first signal, therefore, a good control effect can be achieved by performing dynamic range control synthesis on the first signal based on the target dynamic range control curve.

Description

Technical Field

[0001] The present application relates to the field of audio signal processing, and in particular to an audio signal processing method, apparatus, device, storage medium and product. Background Art

[0002] Dynamic Range Control (DRC) is a signal amplitude adjustment method that provides compression or amplification capabilities to make sounds softer or louder. DRC is widely used in audio signal processing. Dynamic range control maps the dynamic range of the input audio signal to a specified dynamic range. In some scenarios, the dynamic range after mapping is smaller than the dynamic range before mapping, hence the term dynamic range compression. DRC can control the overall dynamic range of the audio signal or divide it into several sub-bands for separate dynamic range control. However, the dynamic range control effect of current related technologies is unsatisfactory. Summary of the Invention

[0003] The embodiments of the present application provide an audio signal processing method, apparatus, device, storage medium, and product, which can significantly improve the dynamic range control effect.

[0004] According to one aspect of an embodiment of the present application, a method for processing an audio signal is provided, the method comprising:

[0005] Obtaining a first signal;

[0006] performing audio signal type recognition on the first signal to obtain an audio signal type recognition result;

[0007] obtaining a target dynamic range control curve corresponding to the first signal according to a dynamic range control curve corresponding to the audio signal type in the audio signal type identification result;

[0008] The first signal is processed according to the target dynamic range control curve to obtain a second signal.

[0009] According to one aspect of an embodiment of the present application, there is provided an audio signal processing apparatus, the apparatus comprising:

[0010] A signal acquisition module, configured to acquire a first signal;

[0011] a dynamic range control module, configured to perform audio signal type recognition on the first signal to obtain an audio signal type recognition result;

[0012] and obtaining a target dynamic range control curve corresponding to the first signal according to the dynamic range control curve corresponding to the audio signal type in the audio signal type identification result;

[0013] and is used to process the first signal according to the target dynamic range control curve to obtain a second signal.

[0014] According to one aspect of an embodiment of the present application, a computer device is provided, comprising a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set, or an instruction set, and the at least one instruction, at least one program, the code set, or the instruction set is loaded and executed by the processor to implement the above-mentioned audio signal processing method.

[0015] According to one aspect of an embodiment of the present application, a computer-readable storage medium is provided, wherein the storage medium stores at least one instruction, at least one program, a code set, or an instruction set, and the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by a processor to implement the above-mentioned audio signal processing method.

[0016] According to one aspect of an embodiment of the present application, a computer program product is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to implement the above-described audio signal processing method.

[0017] The technical solutions provided in the embodiments of the present application can bring the following beneficial effects:

[0018] An embodiment of the present application provides an audio signal processing method that can obtain an input audio signal (a first signal) and perform audio signal type identification on the first signal, thereby analyzing the audio signal type of the audio signal included in the first signal. The dynamic range control curves of audio signals under different audio signal types can be the same or different, and are specifically designed dynamic range control curves suitable for the audio signal type. This dynamic range control curve is customized with full consideration of the characteristics of the audio signal type itself in energy distribution and playback performance, and can therefore achieve a better dynamic range control effect for the audio signal of the corresponding audio signal type.

[0019] For each relevant audio signal type contained in the first signal, the corresponding characterized dynamic range control curve is extracted to obtain the target dynamic range control curve corresponding to the first signal. This makes it possible to achieve a good control effect by performing dynamic range control synthesis on the first signal based on the target dynamic range control curve, and avoid the problem of poor audio signal control for certain audio signal types.

[0020] The audio signal control method provided in the embodiment of the present application is different from the dynamic range control method in the related art. The dynamic range control method in the related art is a purely static control method based on the relevant parameters in the dynamic range control curve. This control method treats different audio types of signals indiscriminately and fails to take into account the sound range characteristics of different types of audio signals. This makes the dynamic range control result unable to express the sound range characteristics of different types of audio signals and the effect is poor. Correspondingly, the embodiment of the present application designs a dynamic range control curve with better sound effects for audio signals of different audio signal types, thereby effectively enhancing the sound quality and improving the effect of dynamic range control. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0022] Figure 1 This is a schematic diagram of a static characteristic curve provided by an embodiment of the present application;

[0023] Figure 2 is a schematic diagram of an application program operating environment provided by an embodiment of the present application;

[0024] Figure 3 is a flowchart of an audio signal processing method provided by an embodiment of the present application;

[0025] Figure 4 This is a flowchart of a method for identifying an audio signal type provided by an embodiment of the present application;

[0026] Figure 5 This is a flowchart of a method for identifying audio signal types using a deep learning model provided by an embodiment of the present application;

[0027] Figure 6 This is a schematic diagram of the results obtained by using multiple curves provided in one embodiment of the present application;

[0028] Figure 7 This is a flow chart of a method for processing a first signal based on a target dynamic range control curve provided by an embodiment of the present application;

[0029] Figure 8 This is an overall flow chart of an audio signal processing method provided by an embodiment of the present application;

[0030] Figure 9 This is a schematic diagram of a process of processing a first signal to obtain a second signal provided by an embodiment of the present application;

[0031] Figure 10 This is a schematic diagram of an application scenario provided by an embodiment of the present application;

[0032] Figure 11 is a block diagram of an audio signal processing device provided by one embodiment of the present application;

[0033] Figure 12 This is a structural block diagram of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0034] Before introducing the method embodiments provided in the present application, a brief introduction is first given to the relevant terms or nouns that may be involved in the method embodiments of the present application to facilitate understanding by those skilled in the art in the field of the present application.

[0035] Artificial Intelligence (AI) refers to the theories, methods, techniques, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, to perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that seeks to understand the essence of intelligence and produce new intelligent machines that can respond in a manner similar to human intelligence. AI also studies the design principles and implementation methods of various intelligent machines, enabling them to possess the capabilities of perception, reasoning, and decision-making.

[0036] Artificial intelligence (AI) technology is a comprehensive discipline encompassing a wide range of fields, encompassing both hardware and software technologies. Foundational AI technologies generally include sensors, specialized AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, speech processing, natural language processing, and machine learning / deep learning.

[0037] Machine learning (ML) is a multidisciplinary field that encompasses probability theory, statistics, approximation theory, convex analysis, and algorithmic complexity theory. It specifically studies how computers can simulate or implement human learning behaviors to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is at the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications span all areas of AI. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and self-learning.

[0038] Deep learning: The concept of deep learning originates from the study of artificial neural networks. A multilayer perceptron with multiple hidden layers is an example of a deep learning architecture. Deep learning discovers distributed feature representations of data by combining lower-level features to form more abstract higher-level representations of attribute categories or features.

[0039] Key technologies in speech technology include automatic speech recognition (ASR), text-to-speech (TTS), and voiceprint recognition. Enabling computers to hear, see, speak, and feel is the future direction of human-computer interaction, with speech becoming one of the most promising methods of human-computer interaction.

[0040] Cloud technology refers to a hosting technology that unifies hardware, software, and network resources within a wide area network (WAN) or local area network (LAN) to enable data computing, storage, processing, and sharing. Cloud technology is a general term for network, information technology, integration technology, management platform technology, and application technology, all based on the cloud computing business model. It can form a resource pool that can be used on demand with flexibility and convenience. Cloud computing technology will become a crucial support. Backend services for technical network systems, such as video websites, image websites, and more portals, require extensive computing and storage resources. With the rapid development and application of the internet industry, every item will likely have its own unique identification mark and will need to be transmitted to backend systems for logical processing. Data of varying levels will be processed separately, and data from all industries will require a strong system backend, which can only be achieved through cloud computing.

[0041] The Fourier transform is a method for expressing a function that satisfies certain conditions as a linear combination of trigonometric functions (sine and / or cosine functions) or their integrals. In different research fields, the Fourier transform has various variations, such as the continuous Fourier transform and the discrete Fourier transform. Fourier analysis was originally proposed as a tool for analytical analysis of thermal processes.

[0042] Sparse Fourier Transform: A simplified version of the discrete Fourier transform. Primarily used in digital signal processing, the Sparse Fourier Transform (SFT) is a simplified version of the discrete Fourier transform. Its core principle is the "sparse" nature of signals: that is, the frequency domain has only a small number of high-value points, while the majority of the remaining points approach zero. This algorithm is 10 to 100 times faster than the traditional Fast Fourier Transform (FFT) and promises revolutionary advances in data transmission.

[0043] Decibel is a unit of measurement that measures the ratio of two quantities of the same unit. It is mainly used to measure sound intensity and is usually expressed in dB.

[0044] Hertz is the SI unit of frequency, measuring the number of repetitions of a periodic fluctuation per second. Hertz is abbreviated as Hz. One vibration (or oscillation, or fluctuation) per second is one Hz, which can also be written as beats / second or cycles / second. Commonly used frequency units are Hertz (Hz) or Kilohertz (kHz).

[0045] Dynamic Range Control (DRC) is a method of signal amplitude adjustment that provides compression and / or amplification capabilities to make sounds softer or louder. DRC is widely used in audio signal processing. Dynamic range control maps the dynamic range of the input audio signal to a specified dynamic range. The mapped dynamic range is typically smaller than the unmapped dynamic range, hence the term dynamic range compression. The audio signal can be dynamically controlled overall or divided into several sub-bands for separate dynamic range control.

[0046] Dynamic range control has the function of automatically adjusting the dynamic range of the input signal. The dynamic range is the logarithm of the ratio of the maximum amplitude to the minimum amplitude of the signal, usually using dB as the unit. The main applications of dynamic range control include the following aspects:

[0047] 1. Dynamic Range Compressor: This reduces the volume of loud signals whose amplitude exceeds a given threshold. DRC processing can be used to protect hardware or increase overall loudness. A Dynamic Range Limiter is a type of dynamic range compressor that limits signals whose amplitude exceeds a given threshold.

[0048] 2. A dynamic range expander reduces the volume of quiet signals with amplitudes below a given threshold, making them appear quieter and preventing interference from small signals. A noise gate is a type of expander that limits signals with amplitudes below a given threshold to reduce the volume of noise.

[0049] The dynamic range control schemes in related technologies are mostly based on configuring different static characteristic curves for different sub-bands to control the audio signal playback effect. Figure 1, which shows a schematic diagram of a static characteristic curve provided by an embodiment of the present application. This static characteristic curve is a conventional dynamic range control curve, which describes the relationship between the dB value of the input signal as the independent variable and the target dB value of the output signal as the dependent variable. The shape of this curve is controlled by four key points. In other words, the shape of the static characteristic curve can be uniquely determined by the four key points, and the position of each of the four key points is described by the coordinate parameters of the corresponding key point. Figure 1 The key points in the equation are x0, x1, x2, and x3. These four points correspond to different target dB values ​​on the vertical axis. This creates a curve mapping the input signal's dB value to the target dB. This curve is known as the dynamic range control curve (static characteristic curve). The static characteristic curve is designed to suppress or attenuate small signals (noise), attenuate large signals, reduce loudness to protect audio hardware, or enhance or attenuate normal playback signals for speech enhancement.

[0050] As can be seen from the above, the solutions for implementing dynamic range control based on static characteristic curves in related technologies only consider the correspondence between the sub-bands of the signal and the static characteristic curves, without considering the components of the signal, that is, the type of audio signal included in the signal. The relevant parameters of the static characteristic curves corresponding to each sub-band do not change during playback, that is, different input signals are processed in the same manner regardless of the audio signal components present in the signal. However, the input signal may contain multiple types of audio signals, and the energy spectra of each type of audio signal vary greatly. If the differences in the energy spectra of different types of audio signals are not considered and the same processing methods are used for all of them, that is, the processing is performed using a unique correspondence between the sub-band and the static characteristic curve, the effect is not ideal.

[0051] Different types of audio signals have significantly different audio ranges and energy spectrum distributions. Some instruments, such as the piccolo, have virtually no energy below 600 Hz, but still have strong signal energy up to 4 kHz. Piano signals, on the other hand, have a very wide energy distribution. If the dynamic range control schemes used in related technologies were to apply the same dynamic range control to all input signals, the result would be that some audio signals would not be able to achieve their full potential, or the dynamic range control would not be effective enough to suppress noise.

[0052] An embodiment of the present application provides an audio signal processing method that can obtain an input audio signal (a first signal) and perform audio signal type identification on the first signal, thereby analyzing the audio signal type of the audio signal included in the first signal. The dynamic range control curves of audio signals under different audio signal types can be the same or different, and are specifically designed dynamic range control curves suitable for the audio signal type. This dynamic range control curve is customized with full consideration of the characteristics of the audio signal type itself in energy distribution and playback performance, and can therefore achieve a better dynamic range control effect for the audio signal of the corresponding audio signal type.

[0053] For each relevant audio signal type contained in the first signal, the corresponding characterized dynamic range control curve is extracted to obtain the target dynamic range control curve corresponding to the first signal. This makes it possible to achieve a good control effect by performing dynamic range control synthesis on the first signal based on the target dynamic range control curve, and avoid the problem of poor audio signal control for certain audio signal types.

[0054] The audio signal control method provided in the embodiment of the present application is different from the dynamic range control method in the related art. The dynamic range control method in the related art is a purely static control method based on the relevant parameters in the dynamic range control curve. This control method treats different audio types of signals indiscriminately and fails to take into account the sound range characteristics of different types of audio signals. This makes the dynamic range control result unable to express the sound range characteristics of different types of audio signals and the effect is poor. Correspondingly, the embodiment of the present application designs a dynamic range control curve with better sound effects for audio signals of different audio signal types, thereby effectively enhancing the sound quality and improving the effect of dynamic range control.

[0055] In order to make the objectives, technical solutions and advantages of this application clearer, the implementation methods of this application will be further described in detail below with reference to the accompanying drawings.

[0056] Please refer to Figure 2 , which shows a schematic diagram of an application program running environment provided by an embodiment of the present application. The application program running environment may include: a terminal 10 and a server 20.

[0057] The terminal 10 includes, but is not limited to, electronic devices such as mobile phones, computers, intelligent voice interaction devices, smart home appliances, car terminals, game consoles, e-book readers, multimedia playback devices, wearable devices, etc. The terminal 10 may be installed with a client of an application.

[0058] In an embodiment of the present application, the above-mentioned application can be any application that can provide audio signal processing services. Typically, the application is a video application. Of course, in addition to video applications, other types of applications can also provide related services that rely on audio signal processing services. For example, news applications, social applications, interactive entertainment applications, browser applications, shopping applications, content sharing applications, virtual reality (VR) applications, augmented reality (AR) applications, etc., are not limited to this in the embodiment of the present application. This embodiment of the present application is not limited to this. Optionally, a client of the above-mentioned application is running in the terminal 10.

[0059] The server 20 is used to provide background services for the client of the application in the terminal 10. For example, the server 20 can be the background server of the above-mentioned application. The server 20 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers. It can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms. Optionally, the server 20 provides background services for applications in multiple terminals 10 at the same time.

[0060] Optionally, the terminal 10 and the server 20 may communicate with each other via a network 30. The terminal 10 and the server 20 may be directly or indirectly connected via wired or wireless communication, which is not limited in this application.

[0061] Please refer to Figure 3 , which shows a flow chart of an audio signal processing method provided by an embodiment of the present application. The method can be applied to a computer device, which refers to an electronic device with data calculation and processing capabilities. For example, the execution subject of each step can be Figure 2 The server 20 in the application running environment shown. The method may include the following steps:

[0062] S301: Acquire a first signal.

[0063] The embodiment of the present application does not limit the source of the first signal, which can be Figure 2The first signal may be collected by a terminal in the system, or may be transmitted from other terminals. The first signal may include audio signals corresponding to one or more audio signal types. The embodiment of the present application does not limit the audio signal type. For example, the first signal may include audio signals corresponding to the following audio signal types: noise, male voice, female voice, children's voice, piano, flute, violin, percussion, etc.

[0064] S302: Perform audio signal type recognition on the first signal to obtain an audio signal type recognition result.

[0065] The embodiment of the present application does not limit the specific method of identifying the audio signal type of the first signal. For example, it can be implemented based on machine learning (ML) technology, deep learning technology or speech technology.

[0066] Please refer to Figure 4 , which shows a flow chart of a method for identifying an audio signal type provided by an embodiment of the present application, the method comprising:

[0067] S401. Perform Fourier transform on the first signal to obtain frequency domain sub-band amplitude information;

[0068] The input signal (first signal) is subjected to Fourier transform to obtain frequency domain subband amplitude information. The embodiment of the present application does not limit the Fourier transform, which can be Fourier transform, fast Fourier transform or sparse Fourier transform, etc.

[0069] S402. Perform feature extraction on the above frequency domain sub-band amplitude information to obtain frequency domain features;

[0070] The embodiments of the present application do not limit the method for extracting frequency domain features, which can be implemented through the neural network structure of the neural network in the related technology. For example, convolutional neural networks, pyramid neural networks, and neural networks with attention structures can all be used for feature extraction. The embodiments of the present application do not limit the neural network structure used.

[0071] S403. Fusing the above frequency domain features to obtain frequency domain fusion features;

[0072] The embodiments of the present application do not limit the fusion method. For example, one or more methods such as convolution, multiplication, addition, fitting, and splicing can be used to achieve the purpose of feature fusion.

[0073] S404. Perform audio signal type mapping on the frequency domain fusion features to obtain the audio signal type recognition result. The audio signal type recognition result includes at least one audio signal type and a probability corresponding to each of the audio signal types. The probability indicates the likelihood that the first signal includes an audio signal of the corresponding audio signal type.

[0074] For example, the embodiment of the present application can use a deep learning model to perform steps S402-S404. The deep learning model can be composed of a deep neural network, which may include a plurality of convolution units, gating units, and fully connected units. The convolution units, gating units, and fully connected units can be used to perform feature extraction operations, feature fusion control operations, and feature fusion operations respectively. The embodiment of the present application does not limit the specific structure of the deep neural network. It can be spliced ​​by convolution units, gating units, and fully connected units based on actual conditions, and the splicing method is not limited. The gating unit refers to a gated recurrent unit network (GRU). Since the three gates in the long short-term memory network contribute differently to improving its learning ability, omitting the gate with a small contribution and its corresponding weight can simplify the neural network structure and improve its learning efficiency. GRU is the structure corresponding to the algorithm proposed based on the above concept. Its corresponding recurrent unit only contains two gates: an update gate and a reset gate. The function of the reset gate is similar to the input gate of the long short-term memory network, and the update gate simultaneously realizes the functions of the forget gate and the output gate. The two gates of the GRU do not form a self-loop, but recurse directly between system states. Therefore, its update equation does not include internal state. There are many variants of the GRU, and you can choose the one according to your actual situation.

[0075] The recognition result of the deep learning model is the probability of each audio signal type. The higher the probability, the more likely it is that the audio signal corresponding to this audio signal type exists in the input signal (first signal). In actual applications, the input signal (first signal) is likely to be a mixed signal, that is, there are audio signals corresponding to multiple audio signal types at the same time. Therefore, the embodiment of the present application can make a judgment based on the confidence probability thresholds corresponding to different types. When the probability corresponding to a certain audio signal type is higher than its corresponding confidence probability threshold, it can be considered that the audio signal corresponding to the certain audio signal type does exist in the first signal, and the audio signal can be confirmed as the target audio signal. The dynamic range control curve corresponding to the target audio signal can affect the processing process of the first signal. On the contrary, when the probability corresponding to a certain audio signal type is lower than or equal to its corresponding confidence probability threshold, it can be considered that the audio signal corresponding to the certain audio signal type does not exist in the first signal, the audio signal is not the target audio signal, and its corresponding dynamic range control curve does not affect the processing process of the first signal.

[0076] In one embodiment, an audio signal type recognition result may be generated based on the audio signal type and probability of each target audio signal. In other words, the audio signal type recognition result may only include the target audio signal. Of course, in other embodiments, the audio signal type and corresponding probability of each of the target audio signal and the non-target audio signal may also be stored in the audio signal type recognition result. This embodiment of the present application is not limited to this.

[0077] In the embodiments of the present application, the confidence probability threshold can be set to different values ​​for different types, depending on the recognition accuracy of the deep learning model. The higher the recognition accuracy, the higher the confidence probability threshold value. Conversely, the lower the recognition accuracy, the lower the confidence probability threshold value can be appropriately set. The set value can be manually set based on actual conditions and application scenario requirements, and the embodiments of the present application do not limit this.

[0078] The present application embodiment does not limit the types of audio signals that the deep learning model can recognize. In an exemplary embodiment, the types of audio signals that the deep learning model can recognize may include the following: noise, male voice, female voice, children's voice, piano, flute, violin, and percussion. Please refer to Figure 5 , which shows a flow chart of a method for using a deep learning model to identify audio signal types in an embodiment of the present application.

[0079] Figure 5 SFFT stands for Semi-Fast Fourier Transform or Sparse Fourier Transform, Conv stands for Convolutional Neural Network layer, GRU stands for Gate-Related Neural Network layer, FC stands for Fully Connected Layer, and Sigmoid stands for Activation Layer. The output of this deep learning model contains the corresponding possibilities of various audio signal types, that is, the probability values ​​corresponding to various audio signal types.

[0080] The present application does not limit the training process of the deep learning model. In an exemplary embodiment, a Fourier transform can be performed on the sample signal to obtain the sample frequency domain sub-band amplitude information; then, based on a preset deep neural network, the following operations are performed on the sample frequency domain sub-band amplitude information:

[0081] Perform feature extraction on the sample frequency domain sub-band amplitude information to obtain the sample frequency domain features;

[0082] The above sample frequency domain features are fused to obtain sample frequency domain fusion features;

[0083] Performing audio signal type mapping on the sample frequency domain fusion features to obtain a sample audio signal type recognition result, wherein the sample audio signal type recognition result includes at least one audio signal type and a predicted probability corresponding to each of the audio signal types;

[0084] The aforementioned feature extraction, feature fusion, and audio signal type mapping have the same meaning as the previous related content and will not be repeated here.

[0085] Based on the difference between the known audio signal type component in the sample signal and the sample audio signal type recognition result, the parameters of the deep neural network are adjusted until the preset training stop condition is reached to obtain the deep learning model.

[0086] The present application does not limit the measurement method of the difference. For example, it can be measured based on the cross entropy loss corresponding to each audio signal type. In the embodiment of the present application, the parameters can be adjusted based on the gradient descent method. The gradient descent method is a method often used in the field of machine learning and deep learning to adjust network parameters. It is a method for first-order optimization adjustment of network parameters by gradient descent. In the embodiment of the present application, the gradient descent method can guide the above parameters to be adjusted in the direction of reducing losses. When the number of adjustments reaches a preset number threshold, or when the loss is less than a preset loss threshold, the parameter adjustment is stopped to obtain the above-mentioned deep learning model.

[0087] S303 : Obtain a target dynamic range control curve corresponding to the first signal according to the dynamic range control curve corresponding to the audio signal type in the audio signal type recognition result.

[0088] The obtaining of the target dynamic range control curve corresponding to the first signal based on the dynamic range control curve corresponding to the audio signal type in the audio signal type recognition result includes: obtaining, for each audio signal type in the audio signal type recognition result having a probability greater than a preset probability threshold, a dynamic range control curve corresponding to the audio signal type; and obtaining the target dynamic range control curve corresponding to the first signal based on the obtained dynamic range control curves corresponding to each of the audio signal types.

[0089] The preset probability threshold refers to the aforementioned confidence probability threshold corresponding to the audio signal type. The preset probability thresholds corresponding to different audio signal types may be the same or different.

[0090] In the embodiments of the present application, corresponding dynamic range control curves can be configured for audio signals of different frequency bands and different audio signal types. Of course, corresponding dynamic range control curves can also be configured for audio signals of all frequency bands and different audio signal types. In other words, the embodiments of the present application do not limit whether dynamic range control curves are differentiated based on frequency bands, and whether or not differentiation is performed does not affect the normal implementation of the present application. However, the embodiments of the present application can design corresponding dynamic range control curves for different audio signal types.

[0091] If no distinction is made, the target dynamic range control curve corresponding to the first signal can be directly obtained based on the dynamic range control curve corresponding to the audio signal type in the audio signal type identification result. If a distinction is made, frequency band division can be performed in advance, that is, at least two frequency bands and a mapping information set corresponding to each of the frequency bands are determined. The mapping information set includes at least two mapping information, each of which is used to find the dynamic range control curve for the corresponding audio signal type in the corresponding frequency band.

[0092] For example, in an embodiment of the present application, the frequency band range can be divided into three: Band 1, Band 2, and Band 3. Band 1, Band 2, and Band 3 can correspond to the low frequency band, mid frequency band, and high frequency band, respectively. Each audio signal type in Band 1 has a corresponding dynamic range control curve, each audio signal type in Band 2 has a corresponding dynamic range control curve, and each audio signal type in Band 3 has a corresponding dynamic range control curve. The dynamic range control curves corresponding to the same audio signal type under different frequency band conditions can be the same or different.

[0093] In one embodiment, obtaining the target dynamic range control curve corresponding to the first signal based on the dynamic range control curve corresponding to the audio signal type in the audio signal type identification result includes:

[0094] First, determining a target frequency band corresponding to the first signal;

[0095] That is, in the at least two frequency bands, the target frequency band that the first signal falls into is determined. Continuing with the above example, it is to find out whether the target frequency band corresponding to the first signal falls into frequency band 1, frequency band 2 or frequency band 3.

[0096] Second, for each target audio signal type in the audio signal type identification result, obtaining a dynamic range control curve corresponding to the target audio signal type under the target frequency band condition, where the target audio signal type is an audio signal type that meets a preset condition;

[0097] Obtaining the dynamic range control curve corresponding to the target audio signal type under the target frequency band includes: determining, in each of the mapping information sets, a target mapping information set corresponding to the target frequency band; and searching, within the target mapping information sets, for the dynamic range control curve corresponding to the target audio signal type under the target frequency band. For example, if the target frequency band corresponding to the first signal falls within frequency band 2, searching the mapping information set corresponding to frequency band 2 (the target mapping information set) for the dynamic range control curve corresponding to the target audio signal type.

[0098] This application does not limit the preset condition satisfied by the target audio signal type. The preset condition may refer to that the predicted probability corresponding to a certain audio signal type is greater than a confidence probability threshold corresponding to the audio signal type.

[0099] Third, the target dynamic range control curve is obtained according to the acquired dynamic range control curves.

[0100] The present application embodiment does not limit the obtaining method, and the obtaining methods of multiple curves in the related art can be applied to the present application embodiment. In an exemplary embodiment, the target dynamic range control curve can be obtained based on the outer contour formed by each of the dynamic range control curves obtained.

[0101] Please refer to Figure 6 , which shows a schematic diagram of multiple curve fitting results provided in the embodiments of the present application. Figure 6 The left figure includes three dynamic range control curves, that is, the dynamic range control curve corresponding to the audio signal type in the above audio signal type recognition result can be obtained. Figure 6 The three dynamic range control curves on the left have their shapes uniquely determined by four key points. Figure 6 The inflection points of the outer contour formed by the three dynamic range control curves in the left figure can be connected to obtain Figure 6 The target dynamic range control curve is shown in the figure on the right.

[0102] S304: Process the first signal according to the target dynamic range control curve to obtain a second signal.

[0103] The present application does not limit the method for processing the first signal. In one embodiment, original sound intensity information of the first signal can be obtained; target sound intensity information can be obtained based on the original sound intensity information and the target dynamic range control curve; a target adjustment gain can be determined based on the difference between the target sound intensity information and the original sound intensity information; a linear conversion can be performed on the target adjustment gain to obtain an adjustment parameter; and the adjustment parameter can be fused with the first signal to obtain the second signal.

[0104] Please refer to Figure 7 , which shows a flow chart of a method for processing a first signal based on a target dynamic range control curve in an exemplary embodiment. Input signal x[n], where n is the serial number of the input signal, in the embodiment of the present application, x[n] can refer to the first signal. Converted to dB unit through logarithm, that is, xdB=20log10(x), xdB is used as the input of the static characteristic unit. xdB represents the original sound intensity information. The static characteristic parameters stored in the static characteristic unit and the target dynamic range control curve both refer to the static characteristics on which the dynamic range control is based, and the only difference is the way the characteristics are expressed. What is output by the static characteristic unit is the target dB value, which represents the target sound intensity information. In fact, it is the process of obtaining the corresponding vertical coordinate when the horizontal coordinate xdB of the point on the target dynamic range control curve is known. The vertical coordinate is the target sound intensity information. By taking the difference vdB = mdB - xdB as the target adjustment gain, and performing a linear gain conversion v = pow(10, vdB / 20), we can obtain the adjustment parameter v. The final output y[n] = x[n] * v reflects the process of fusing the adjustment parameter with the first signal. This y[n] is reconstructed (e.g., through an inverse Fourier transform) to obtain the complete time domain signal, i.e., the second signal.

[0105] Please refer to Figure 8 , which shows the overall flow chart of the audio signal processing method of an embodiment of the present application. First, a signal to be played, i.e., a first signal, is obtained, and audio signal classification is performed on the first signal. Specifically, one or more audio signal types present in the first signal can be identified by making a confidence probability threshold judgment corresponding to each audio signal type. Each audio signal type has a corresponding dynamic range control parameter table, i.e., different audio signal types have different dynamic range control static characteristic curves, referred to as dynamic range control curves. Since audio signals of different audio signal types have different sound ranges and different sound characteristics, the parameters of the dynamic range control curve may also be different.

[0106] For example, for male and female voices, because male voices tend to be deeper and have lower fundamental frequencies, their energy is primarily concentrated in the low and mid-range frequencies, with weaker high-frequency signals. Therefore, the input value corresponding to position x0 of the dynamic range control curve for subbands below the fundamental frequency is larger, while the output value corresponding to position x3 is smaller. Furthermore, the input value corresponding to position x3 of the dynamic range control curve for the low and mid-range frequency subbands is larger to relax the upper output limit of the low and mid-range frequencies, while the output value corresponding to position x3 of the dynamic range control curve for the high and mid-range frequency subbands is smaller. Unlike male voices, female voices have higher fundamental frequencies and still have higher energy in the mid and high-range frequencies. Therefore, the input value corresponding to position x3 of the dynamic range control curve for the high-range subbands is larger, and the output value corresponding to position x3 is also larger. Static characteristic parameters for different types of dynamic range control are pre-stored on the terminal device or server.

[0107] A high-confidence category is extracted from the results obtained after audio signal classification. It can be considered that the high-confidence category corresponds to the aforementioned target audio signal type. The dynamic range control parameters corresponding to the high-confidence type are read and a union operation is performed to obtain a fusion result of one or more comprehensive dynamic range control static characteristic curves corresponding to the high confidence. Then, dynamic control is performed based on the fusion result to obtain a second signal.

[0108] Please refer to Figure 9 , which shows a schematic diagram of the process of processing the first signal to obtain the second signal in the embodiment of the present application. The embodiment of the present application proposes a dynamic range control method based on signal classification, which obtains the static characteristics based on which the dynamic range control of the first signal is performed by identifying the audio signal type corresponding to the audio signal present in the first signal, that is, performing signal type recognition. The static characteristics are reflected in Figure 9 In the static characteristic unit in , the static characteristic is the static characteristic possessed by the above-mentioned target dynamic range control curve. Based on the static characteristic unit, dynamic range control for the first signal can be performed. The specific control method has been described above and will not be repeated here. The embodiment of the present application does not limit the type of audio signal that can be identified in the signal type recognition link. For example, it can include noise, voice (male voice, female voice, children's voice, etc.), music (orchestral, piano, percussion, etc.). Different dynamic range control curves are pre-designed according to the sound range spectrum range and sound characteristics of different audio signal types. The corresponding parameters of the curves are saved in the terminal device or server. During playback, based on the type recognition result corresponding to the first signal, one or more groups of parameters corresponding to the dynamic range control curves are called and integrated in real time to obtain Figure 9The first signal is processed based on the static characteristics in the dynamic range control method in the related art, and the first signal is processed based on the static characteristics. The embodiment of the present application is different from the control scheme based on static parameters in the dynamic range control method in the related art. It makes up for the disadvantage that the dynamic range control method processes different audio types indiscriminately and cannot show the sound range characteristics of different types of audio signals. Instead, it designs dynamic range control parameters with better sound effects for different audio signal types, thereby obtaining an effective method for sound quality enhancement.

[0109] An embodiment of the present application provides an audio signal processing method that can obtain an input audio signal (a first signal) and perform audio signal type identification on the first signal, thereby analyzing the audio signal type of the audio signal included in the first signal. The dynamic range control curves of audio signals under different audio signal types can be the same or different, and are specifically designed dynamic range control curves suitable for the audio signal type. This dynamic range control curve is customized based on the characteristics of the audio signal type itself in energy distribution and playback performance, and can therefore achieve a better dynamic range control effect for the audio signal of the corresponding audio signal type.

[0110] For each relevant audio signal type contained in the first signal, the corresponding characterized dynamic range control curve is extracted to obtain the target dynamic range control curve corresponding to the first signal. This makes it possible to achieve a good control effect by performing dynamic range control synthesis on the first signal based on the target dynamic range control curve, and avoid the problem of poor audio signal control for certain audio signal types.

[0111] The audio signal control method provided in the embodiment of the present application is different from the dynamic range control method in the related art. The dynamic range control method in the related art is a purely static control method based on the relevant parameters in the dynamic range control curve. This control method treats different audio types of signals indiscriminately and fails to take into account the sound range characteristics of different types of audio signals. This makes the dynamic range control result unable to express the sound range characteristics of different types of audio signals and the effect is poor. Correspondingly, the embodiment of the present application designs a dynamic range control curve with better sound effects for signals of different audio types, thereby effectively enhancing the sound quality and improving the effect of dynamic range control.

[0112] The embodiment of the present application has broad application prospects. It can be deployed in audio hardware playback devices or embedded in software built-in algorithms, or it can be deployed in servers and used in different audio signal playback application scenarios, such as live broadcast, audio and video playback, broadcast, voice calls, etc. Please refer to Figure 10 , which shows a schematic diagram of an application scenario of an embodiment of the present application, Figure 10 The left figure shows a voice broadcast scenario applicable to the embodiment of the present application. Figure 10 The right picture shows the music playback scenario used in the embodiment of the present application.

[0113] The following are device embodiments of the present application, which can be used to implement the method embodiments of the present application. For details not disclosed in the device embodiments of the present application, please refer to the method embodiments of the present application.

[0114] Please refer to Figure 11 , which shows a block diagram of an audio signal processing device provided by one embodiment of the present application. The device has the function of implementing the above-mentioned audio signal processing method. The above-mentioned function can be implemented by hardware or by hardware executing corresponding software. The device can be a computer device or can be set in a computer device. The device can include:

[0115] A signal acquisition module 1101 is configured to acquire a first signal;

[0116] A dynamic range control module 1102 is configured to perform audio signal type recognition on the first signal to obtain an audio signal type recognition result;

[0117] and obtaining a target dynamic range control curve corresponding to the first signal according to the dynamic range control curve corresponding to the audio signal type in the audio signal type identification result;

[0118] And it is used to process the first signal according to the target dynamic range control curve to obtain a second signal.

[0119] In one embodiment, the dynamic range control module 1102 is configured to perform the following operations:

[0120] determining a target frequency band corresponding to the first signal;

[0121] For each target audio signal type in the audio signal type identification result, obtaining a dynamic range control curve corresponding to the target audio signal type under the target frequency band condition, where the target audio signal type is an audio signal type that meets a preset condition;

[0122] The target dynamic range control curve is obtained based on the acquired dynamic range control curves.

[0123] In one embodiment, the dynamic range control module 1102 is configured to perform the following operations:

[0124] The target dynamic range control curve is obtained according to the outer contours formed by the acquired dynamic range control curves.

[0125] In one embodiment, the dynamic range control module 1102 is configured to perform the following operations:

[0126] Performing Fourier transform on the first signal to obtain frequency domain sub-band amplitude information;

[0127] Perform feature extraction on the above frequency domain sub-band amplitude information to obtain frequency domain features;

[0128] The above frequency domain features are fused to obtain frequency domain fusion features;

[0129] Audio signal type mapping is performed on the above-mentioned frequency domain fusion features to obtain the above-mentioned audio signal type recognition result, where the above-mentioned audio signal type recognition result includes at least one audio signal type and the probability corresponding to each of the above-mentioned audio signal types, and the above-mentioned probability represents the possibility that the above-mentioned first signal includes the audio signal pointed to by the corresponding audio signal type.

[0130] In one embodiment, the dynamic range control module 1102 is configured to perform the following operations:

[0131] For each audio signal type in the audio signal type recognition result having a probability greater than a preset probability threshold, obtaining a dynamic range control curve corresponding to the audio signal type;

[0132] A target dynamic range control curve corresponding to the first signal is obtained according to the acquired dynamic range control curves corresponding to the respective audio signal types.

[0133] In one embodiment, the dynamic range control module 1102 is configured to perform the following operations:

[0134] Obtaining original sound intensity information of the first signal;

[0135] Obtaining target sound intensity information according to the original sound intensity information and the target dynamic range control curve;

[0136] determining a target adjustment gain according to a difference between the target sound intensity information and the original sound intensity information;

[0137] Performing linear conversion on the above target adjustment gain to obtain the adjustment parameter;

[0138] The adjustment parameters are combined with the first signal to obtain the second signal.

[0139] In one embodiment, the dynamic range control module 1102 is configured to perform the following operations:

[0140] Determining at least two frequency bands and a mapping information set corresponding to each of the frequency bands, wherein the mapping information set includes at least two mapping information, each of the mapping information is used to find a dynamic range control curve for a corresponding audio signal type in a corresponding frequency band;

[0141] Determining, among the at least two frequency bands, the target frequency band into which the first signal falls;

[0142] Determining, in each of the mapping information sets, a target mapping information set corresponding to the target frequency band;

[0143] In the target mapping information set, a dynamic range control curve corresponding to the target audio signal type under the target frequency band condition is searched.

[0144] It should be noted that the apparatus provided in the above embodiments, when implementing its functions, is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the apparatus and method embodiments provided in the above embodiments are based on the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0145] Please refer to Figure 12 , which shows a block diagram of a computer device provided by an embodiment of the present application. The computer device may be a server for executing the above-mentioned audio signal processing method. Specifically:

[0146] Computer device 1000 includes a central processing unit (CPU) 1001, a system memory 1004 including a random access memory (RAM) 1002 and a read-only memory (ROM) 1003, and a system bus 1005 connecting system memory 1004 and CPU 1001. Computer device 1000 also includes a basic input / output system (I / O system) 1006 for facilitating information transfer between various components within the computer, and a mass storage device 1007 for storing an operating system 1013, application programs 1014, and other program modules 1015.

[0147] The basic input / output system 1006 includes a display 1008 for displaying information and an input device 1009, such as a mouse and keyboard, for user input. Both the display 1008 and the input device 1009 are connected to the central processing unit 1001 via an input / output controller 1010 connected to the system bus 1005. The basic input / output system 1006 may also include an input / output controller 1010 for receiving and processing input from a variety of other devices, such as a keyboard, mouse, or electronic stylus. Similarly, the input / output controller 1010 also provides output to a display screen, printer, or other types of output devices.

[0148] The mass storage device 1007 is connected to the central processing unit 1001 via a mass storage controller (not shown) connected to the system bus 1005. The mass storage device 1007 and its associated computer-readable media provide non-volatile storage for the computer device 1000. In other words, the mass storage device 1007 may include a computer-readable medium (not shown) such as a hard disk or a CD-ROM (Compact Disc Read-Only Memory) drive.

[0149] Without loss of generality, computer-readable media may include computer storage media and communication media. Computer storage media include volatile and non-volatile, removable and non-removable media implemented by any method or technology for storing information such as computer-readable instructions, data structures, program modules or other data. Computer storage media include RAM, ROM, EPROM (Erasable Programmable Read Only Memory), EEPROM (Electrically Erasable Programmable Read Only Memory), flash memory or other solid-state storage technologies, CD-ROM, DVD (Digital Video Disc) or other optical storage, tape cassettes, magnetic tape, disk storage or other magnetic storage devices. Of course, those skilled in the art will appreciate that computer storage media are not limited to the above-mentioned ones. The above-mentioned system memory 1004 and mass storage device 1007 can be collectively referred to as memory.

[0150] According to various embodiments of the present application, the computer device 1000 may also be connected to a remote computer on a network such as the Internet for operation. That is, the computer device 1000 may be connected to the network 1012 via the network interface unit 1011 connected to the system bus 1005, or the network interface unit 1011 may be used to connect to other types of networks or remote computer systems (not shown).

[0151] The memory further includes a computer program, which is stored in the memory and configured to be executed by one or more processors to implement the audio signal processing method.

[0152] In an exemplary embodiment, a computer-readable storage medium is further provided, wherein the storage medium stores at least one instruction, at least one program, a code set, or an instruction set. When the at least one instruction, the at least one program, the code set, or the instruction set is executed by a processor, the audio signal processing method is implemented.

[0153] Specifically, the audio signal processing method includes:

[0154] Obtaining a first signal;

[0155] Performing audio signal type recognition on the first signal to obtain an audio signal type recognition result;

[0156] obtaining a target dynamic range control curve corresponding to the first signal according to the dynamic range control curve corresponding to the audio signal type in the audio signal type recognition result;

[0157] The first signal is processed according to the target dynamic range control curve to obtain a second signal.

[0158] In one embodiment, obtaining the target dynamic range control curve corresponding to the first signal based on the dynamic range control curve corresponding to the audio signal type in the audio signal type identification result includes:

[0159] determining a target frequency band corresponding to the first signal;

[0160] For each target audio signal type in the audio signal type identification result, obtaining a dynamic range control curve corresponding to the target audio signal type under the target frequency band condition, where the target audio signal type is an audio signal type that meets a preset condition;

[0161] The target dynamic range control curve is obtained based on the acquired dynamic range control curves.

[0162] In one embodiment, obtaining the target dynamic range control curve according to the acquired dynamic range control curves includes:

[0163] The target dynamic range control curve is obtained according to the outer contours formed by the acquired dynamic range control curves.

[0164] In one embodiment, performing audio signal type recognition on the first signal to obtain an audio signal type recognition result includes:

[0165] Performing Fourier transform on the first signal to obtain frequency domain sub-band amplitude information;

[0166] Perform feature extraction on the above frequency domain sub-band amplitude information to obtain frequency domain features;

[0167] The above frequency domain features are fused to obtain frequency domain fusion features;

[0168] Audio signal type mapping is performed on the above-mentioned frequency domain fusion features to obtain the above-mentioned audio signal type recognition result, where the above-mentioned audio signal type recognition result includes at least one audio signal type and the probability corresponding to each of the above-mentioned audio signal types, and the above-mentioned probability represents the possibility that the above-mentioned first signal includes the audio signal pointed to by the corresponding audio signal type.

[0169] In one embodiment, obtaining the target dynamic range control curve corresponding to the first signal based on the dynamic range control curve corresponding to the audio signal type in the audio signal type identification result includes:

[0170] For each audio signal type in the audio signal type recognition result having a probability greater than a preset probability threshold, obtaining a dynamic range control curve corresponding to the audio signal type;

[0171] A target dynamic range control curve corresponding to the first signal is obtained according to the acquired dynamic range control curves corresponding to the respective audio signal types.

[0172] In one embodiment, the processing of the first signal according to the target dynamic range control curve to obtain the second signal includes:

[0173] Obtaining original sound intensity information of the first signal;

[0174] Obtaining target sound intensity information according to the original sound intensity information and the target dynamic range control curve;

[0175] determining a target adjustment gain according to a difference between the target sound intensity information and the original sound intensity information;

[0176] Performing linear conversion on the above target adjustment gain to obtain the adjustment parameter;

[0177] The adjustment parameters are combined with the first signal to obtain the second signal.

[0178] In one embodiment, the method further includes:

[0179] Determining at least two frequency bands and a mapping information set corresponding to each of the frequency bands, wherein the mapping information set includes at least two mapping information, each of the mapping information is used to find a dynamic range control curve for a corresponding audio signal type in a corresponding frequency band;

[0180] The determining of the target frequency band corresponding to the first signal includes: determining, among the at least two frequency bands, the target frequency band into which the first signal falls;

[0181] The step of obtaining a dynamic range control curve corresponding to the target audio signal type under the target frequency band condition includes:

[0182] Determining, in each of the mapping information sets, a target mapping information set corresponding to the target frequency band;

[0183] In the target mapping information set, a dynamic range control curve corresponding to the target audio signal type under the target frequency band condition is searched.

[0184] Optionally, the computer-readable storage medium may include: ROM (Read Only Memory), RAM (Random Access Memory), SSD (Solid State Drives), or an optical disk, etc. Among them, the random access memory may include ReRAM (Resistance Random Access Memory) and DRAM (Dynamic Random Access Memory).

[0185] In an exemplary embodiment, a computer program product or computer program is further provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the above-described audio signal processing method.

[0186] It should be understood that the "multiple" mentioned in this article refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent three situations: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the previous and subsequent associated objects are in an "or" relationship. In addition, the step numbers described in this article only illustrate a possible execution sequence between the steps. In some other embodiments, the above steps may not be executed in the order of the numbers, such as two steps with different numbers are executed at the same time, or two steps with different numbers are executed in the opposite order to the diagram. The embodiments of the present application do not limit this.

[0187] In addition, in the specific implementation of this application, related data such as user information is involved. When the above embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions.

[0188] The above description is merely an exemplary embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. An audio signal processing method, characterized in that: The method comprises: Obtaining a first signal; Performing audio signal type recognition on the first signal to obtain an audio signal type recognition result; obtaining a target dynamic range control curve corresponding to the first signal according to a dynamic range control curve corresponding to the audio signal type in the audio signal type recognition result; The first signal is processed according to the target dynamic range control curve to obtain a second signal.

2. The method according to claim 1, characterized in that The obtaining, according to the dynamic range control curve corresponding to the audio signal type in the audio signal type recognition result, a target dynamic range control curve corresponding to the first signal, comprises: determining a target frequency band corresponding to the first signal; For each target audio signal type in the audio signal type identification result, obtaining a dynamic range control curve corresponding to the target audio signal type under the target frequency band condition, wherein the target audio signal type is an audio signal type that meets a preset condition; The target dynamic range control curve is obtained according to each of the acquired dynamic range control curves.

3. The method according to claim 2, characterized in that The step of obtaining the target dynamic range control curve according to each of the acquired dynamic range control curves includes: The target dynamic range control curve is obtained according to the outer contour formed by each of the acquired dynamic range control curves.

4. The method according to any one of claims 1 to 3, characterized in that The performing audio signal type identification on the first signal to obtain an audio signal type identification result includes: Performing Fourier transform on the first signal to obtain frequency domain sub-band amplitude information; Extracting features from the frequency domain sub-band amplitude information to obtain frequency domain features; Fusing the frequency domain features to obtain frequency domain fusion features; The frequency domain fusion feature is mapped to an audio signal type to obtain the audio signal type recognition result, wherein the audio signal type recognition result includes at least one audio signal type and a probability corresponding to each of the audio signal types, wherein the probability represents the possibility that the first signal includes an audio signal pointed to by the corresponding audio signal type.

5. The method according to claim 4, characterized in that The obtaining, according to the dynamic range control curve corresponding to the audio signal type in the audio signal type recognition result, a target dynamic range control curve corresponding to the first signal, comprises: For each audio signal type in the audio signal type recognition result with a probability greater than a preset probability threshold, obtaining a dynamic range control curve corresponding to the audio signal type; A target dynamic range control curve corresponding to the first signal is obtained according to the acquired dynamic range control curves corresponding to the respective audio signal types.

6. The method according to claim 5, characterized in that The step of processing the first signal according to the target dynamic range control curve to obtain a second signal includes: Acquiring original sound intensity information of the first signal; Obtaining target sound intensity information according to the original sound intensity information and the target dynamic range control curve; determining a target adjustment gain according to a difference between the target sound intensity information and the original sound intensity information; Performing linear conversion on the target adjustment gain to obtain an adjustment parameter; The adjustment parameter is integrated with the first signal to obtain the second signal.

7. The method according to claim 2, characterized in that The method further comprises: Determine at least two frequency bands and a mapping information set corresponding to each of the frequency bands, wherein the mapping information set includes at least two mapping information, and each of the mapping information is used to find a dynamic range control curve of a corresponding audio signal type in a corresponding frequency band; The determining a target frequency band corresponding to the first signal comprises: determining, among the at least two frequency bands, the target frequency band into which the first signal falls; The obtaining of a dynamic range control curve corresponding to the target audio signal type under the target frequency band condition comprises: In each of the mapping information sets, determining a target mapping information set corresponding to the target frequency band; In the target mapping information set, a dynamic range control curve corresponding to the target audio signal type under the target frequency band condition is searched.

8. An audio signal processing device, characterized in that: The device comprises: A signal acquisition module, used to acquire a first signal; A dynamic range control module, configured to perform audio signal type recognition on the first signal to obtain an audio signal type recognition result; and obtaining a target dynamic range control curve corresponding to the first signal according to the dynamic range control curve corresponding to the audio signal type in the audio signal type identification result; And used to process the first signal according to the target dynamic range control curve to obtain a second signal.

9. A computer device, characterized in that: The computer device includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement the audio signal processing method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The storage medium stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement the audio signal processing method according to any one of claims 1 to 7.

Images