Power amplifier distortion compensation system based on deep learning

Abstract

The invention relates to the technical field of audio amplification, and particularly discloses a power amplifier distortion compensation system based on deep learning, which comprises an input signal acquisition module, a deep learning distortion identification module, an electron tube simulation module, a distortion compensation module, a PWM (Pulse Width Modulation) generation and optimization module, a power amplification module and a feedback learning module, according to the system, a deep learning algorithm and a multi-band dynamic feature extraction method are combined, high-precision recognition of complex nonlinear distortion types such as harmonic distortion and dynamic distortion is achieved, compared with traditional spectral analysis or static filtering compensation, the system can dynamically adjust a compensation strategy, the influence of distortion on the tone quality is remarkably reduced, and the tone quality is improved. Through a nonlinear even harmonic enhancement algorithm, natural tone and auditory texture of an audio signal are highly restored, intelligent restoration of signal distortion is realized through a GAN framework, a generator dynamically generates a high-fidelity compensation signal, and meanwhile, a discriminator learns characteristics of different distortion types to ensure that the restored signal is close to an undistorted target.

Description

Technical Field

[0001] The present invention belongs to the technical field of audio amplification, and particularly relates to a power amplifier distortion compensation system based on deep learning. Background Art

[0002] In the field of audio amplification, a power amplifier is a key component for audio signal transmission, responsible for amplifying a low-level audio signal to a level sufficient to drive a speaker to produce sound. However, distortion inevitably occurs during the power amplification process, mainly due to the non-linear characteristics of the device, thermal effects, and physical limitations of circuit components. Distortion reduces the sound quality and affects the listening experience of the audience, especially in high-fidelity audio applications, where the distortion problem is particularly prominent.

[0003] Most of the existing power amplifier distortion compensation technologies are based on traditional spectrum analysis methods or static filtering techniques. These technologies identify and compensate specific distortion frequency components by analyzing the spectral characteristics of audio signals. However, this method has obvious limitations. First, they can often only handle single or simple distortion types, such as linear distortion or static harmonic distortion, and have limited compensation effects for complex non-linear distortions, such as dynamic distortion or intermodulation distortion. Second, traditional methods lack the ability of dynamic adjustment and cannot adjust the compensation strategy in real time according to the changes in audio signals, resulting in poor compensation effects when the dynamic range of audio signals is large or the distortion characteristics are complex.

[0004] In addition, although some advanced audio processing technologies, such as digital signal processing (DSP) algorithms, can simulate the non-linear characteristics of analog devices such as vacuum tubes to a certain extent, these algorithms usually rely on fixed parameter settings and lack the adaptive ability to different audio environments and power amplifier characteristics. Therefore, in practical applications, these algorithms are often difficult to achieve the ideal sound quality restoration effect.

[0005] In view of this, the inventor proposes a power amplifier distortion compensation system based on deep learning to solve the above problems. Summary of the Invention

[0006] The purpose of the present invention is to provide a power amplifier distortion compensation system based on deep learning to solve the problems raised in the above background art.

[0007] To achieve the above purpose, the present invention provides the following technical solutions:

[0008] A power amplifier distortion compensation system based on deep learning, comprising:

[0009] An input signal acquisition module, configured to acquire an analog audio signal and convert it into a digital signal through an analog-to-digital converter to obtain an audio digital signal;

[0010] A deep learning distortion recognition module, which is used to analyze the audio digital signal by using a deep learning model, identify the type of distortion generated by the power amplifier, and generate corresponding distortion compensation parameters;

[0011] A tube simulation module, which is used to add the even harmonic characteristics generated by simulating tubes to the audio digital signal to generate an enhanced signal;

[0012] A distortion compensation module, which is used to process the enhanced signal based on the distortion compensation parameters, eliminate the distortion and restore the original characteristics of the audio signal to generate a repaired signal;

[0013] A PWM generation and optimization module, which is used to convert the repaired signal into a pulse width modulation signal for power amplification;

[0014] A power amplification module, which is used to amplify the pulse width modulation signal and drive the speaker to emit sound to output an audio signal;

[0015] A feedback learning module, which is used to collect the audio signal output by the speaker, compare it with the audio digital signal, and generate feedback data for optimizing the distortion compensation ability of the deep learning model.

[0016] Preferably, the deep learning distortion recognition module uses a multi-band dynamic feature extraction algorithm to extract distortion features and adaptively generates distortion compensation parameters based on the frequency band and dynamic range.

[0017] Preferably, the formula of the multi-band dynamic feature extraction algorithm is:

[0018]

[0019] Where Dk: the degree of distortion of frequency band k;

[0020] Xk(t): the amplitude of the original signal in frequency band k;

[0021] The target distortion-free signal predicted by deep learning;

[0022] T: the size of the sampling time window;

[0023] α: a trade-off term used to control the contribution of dynamic changes to the degree of distortion;

[0024] Var(Xk(t)): the amount of dynamic change in frequency band k.

[0025] Preferably, the tube simulation module uses a digital signal processing algorithm to simulate the nonlinear characteristics of tubes in real time, and uses an even harmonic enhancement algorithm to dynamically generate even harmonics and enhance the naturalness of the signal quality.

[0026] Preferably, the formula of the even harmonic enhancement algorithm is:

[0027] Y(t) = X(t) + β·X 2 (t) + γ·X 4 (t)

[0028] where Y(t): the output enhanced audio signal;

[0029] X(t): the input audio signal;

[0030] β, γ: weight parameters for controlling the intensity of even harmonics, dynamically adjusted based on deep learning.

[0031] Preferably, the distortion compensation module combines a generative adversarial network to perform high-precision repair on the enhanced signal, dynamically compensating for harmonic distortion and noise distortion.

[0032] Preferably, the formula of the generative adversarial network is:

[0033] Generator objective function:

[0034] L G = -Ε[log(D(G(X)))] + λLrec

[0035] Discriminator objective function:

[0036] L D = -Ε[log(D(X))] - Ε[log(1 - D(G(X)))]

[0037] where G: the generator, used to repair the distorted signal;

[0038] D: the discriminator, used to distinguish the repaired signal from the distortion-free target signal;

[0039] X: the input distorted signal;

[0040] Lrec: the reconstruction loss, used to constrain the similarity between the repaired signal and the target signal;

[0041] λ: the weight parameter, controlling the influence of the reconstruction loss on the overall objective function.

[0042] Preferably, the PWM generation and optimization module optimizes the PWM modulation parameters based on deep reinforcement learning;

[0043] The formula of the deep reinforcement learning is:

[0044]

[0045] where Q(st, at): the value function of performing action at in state st;

[0046] η: Learning rate;

[0047] rt: Reward after executing an action;

[0048] γ: Discount factor, used to balance short - term and long - term rewards.

[0049] Preferably, the feedback learning module updates the parameters of the deep learning model in real - time through an online learning mechanism, adapts to different audio environments and power amplifier characteristics, and improves the compensation ability of the system.

[0050] Preferably, the formula for real - time update of the online learning mechanism is:

[0051]

[0052] where Li: Loss function of the i - th task;

[0053] ωi: Weight parameter, used to balance the losses of each task;

[0054] R(Θ): Regularization term, used to prevent overfitting;

[0055] μ: Regularization term weight;

[0056] Θ: Parameters of the deep learning model.

[0057] Compared with the prior art, the beneficial effects of the present invention are:

[0058] (1) The system of the present invention combines deep learning algorithms and multi - band dynamic feature extraction methods to achieve high - precision identification of complex non - linear distortion types such as harmonic distortion and dynamic distortion. Compared with traditional spectrum analysis or static filter compensation, this system can dynamically adjust the compensation strategy and significantly reduce the impact of distortion on sound quality.

[0059] (2) The present invention uses a vacuum tube simulation module. Through a non - linear even - harmonic enhancement algorithm, it highly restores the natural timbre and auditory texture of audio signals. Through the GAN framework, it realizes intelligent repair of signal distortion. The generator dynamically generates high - fidelity compensation signals, and at the same time, the discriminator learns the characteristics of different distortion types to ensure that the repaired signal approaches the distortion - free target. This mechanism has strong self - adaptability and generalization ability and can handle a variety of complex audio scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 It is a block diagram of a power amplifier distortion compensation system based on deep learning according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0061] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0062] Embodiment 1:

[0063] Please refer to Figure 1 As shown in the figure, a power amplifier distortion compensation system based on deep learning includes:

[0064] An input signal acquisition module, configured to acquire an analog audio signal and convert it into a digital signal through an analog-to-digital converter (ADC) to obtain an audio digital signal;

[0065] A deep learning distortion identification module, configured to analyze the audio digital signal using a deep learning model, identify the type of distortion generated by the power amplifier, and generate corresponding distortion compensation parameters;

[0066] A vacuum tube simulation module, configured to add the even harmonic characteristics generated by an analog vacuum tube to the audio digital signal to generate an enhanced signal;

[0067] A distortion compensation module, configured to process the enhanced signal based on the distortion compensation parameters, eliminate the distortion and restore the original characteristics of the audio signal, and generate a repaired signal;

[0068] A PWM generation and optimization module, configured to convert the repaired signal into a pulse width modulation signal for power amplification;

[0069] A power amplification module, configured to amplify the pulse width modulation signal and drive a speaker to emit sound, and output an audio signal;

[0070] A feedback learning module, configured to collect the audio signal output by the speaker, compare it with the audio digital signal, and generate feedback data for optimizing the distortion compensation ability of the deep learning model.

[0071] Specifically, the deep learning distortion identification module uses a multi-band dynamic feature extraction algorithm to extract distortion features, and adaptively generates distortion compensation parameters based on the frequency band and dynamic range;

[0072] The formula of the multi-band dynamic feature extraction algorithm is:

[0073]

[0074] Where Dk: the degree of distortion of frequency band k;

[0075] Xk(t): The amplitude of the original signal in frequency band k;

[0076] The target distortion - free signal predicted by deep learning;

[0077] T: The size of the sampling time window;

[0078] α: The trade - off term used to control the contribution of dynamic changes to the degree of distortion;

[0079] Var(Xk(t)): The amount of dynamic change in frequency band k;

[0080] By combining the dynamic changes in frequency bands and the deviation of distortion amplitude, this algorithm can accurately identify non - linear distortions (harmonic distortion, dynamic distortion). Compared with traditional spectrum analysis methods, this method can adapt to the complex structures of different audio signals and adjust the compensation strategy in real - time.

[0081] Specifically, the vacuum tube simulation module real - time simulates the non - linear characteristics of vacuum tubes through a digital signal processing (DSP) algorithm, and uses the even - harmonic enhancement algorithm to dynamically generate even harmonics and enhance the naturalness of the signal's sound quality;

[0082] The formula of the even - harmonic enhancement algorithm is:

[0083] Y(t) = X(t)+β·X 2 (t)+γ·X 4 (t)

[0084] Where Y(t): The output enhanced audio signal;

[0085] X(t): The input audio signal;

[0086] β,γ: The weight parameters for controlling the intensity of even harmonics, dynamically adjusted based on deep learning;

[0087] This algorithm uses non - linear transformation to generate even harmonics, enhancing the naturalness and expressiveness of the audio signal; compared with static vacuum tube simulation models, deep learning dynamically adjusts the parameters β and γ, which can optimize the harmonic distribution according to the frequency and amplitude characteristics of the signal.

[0088] Specifically, the distortion compensation module combines a generative adversarial network (GAN) to perform high - precision repair on the enhanced signal, dynamically compensating for harmonic distortion and noise distortion;

[0089] The formula of the generative adversarial network is:

[0090] Generator objective function:

[0091] L G =-Ε[log(D(G(X)))] + λLrec

[0092] Discriminator objective function:

[0093] L D = -Ε[log(D(X))] - Ε[log(1 - D(G(X)))]

[0094] where G: Generator, used to repair distorted signals;

[0095] D: Discriminator, used to distinguish the repaired signal from the undistorted target signal;

[0096] X: Input distorted signal;

[0097] Lrec: Reconstruction loss, used to constrain the similarity between the repaired signal and the target signal;

[0098] λ: Weight parameter, controlling the influence of the reconstruction loss on the overall objective function;

[0099] Through the GAN framework, the system can learn complex distortion characteristics and generate high-quality compensation signals. The reconstruction loss ensures the audio fidelity of the repaired signal and avoids distortion amplification caused by overcompensation;

[0100] The PWM generation and optimization module optimizes the PWM modulation parameters based on deep reinforcement learning to ensure the high fidelity and low distortion of the pulse width modulation signal;

[0101] The formula of the deep reinforcement learning is:

[0102]

[0103] where Q(st, at): Value function of performing action at in state st;

[0104] η: Learning rate;

[0105] rt: Reward after performing the action;

[0106] γ: Discount factor, used to balance short-term and long-term rewards;

[0107] Using reinforcement learning to optimize the PWM modulation parameters ensures the high-fidelity output of the modulation signal in complex audio situations. Compared with traditional static modulation strategies, this method can dynamically adapt to audio characteristics, improve system efficiency and reduce harmonic distortion.

[0108] Specifically, the feedback learning module adaptively updates the parameters of the deep learning model through an online learning mechanism to adapt to different audio environments and power amplifier characteristics, improving the compensation ability of the system;

[0109] The formula for the real-time update of the online learning mechanism is:

[0110]

[0111] where \(L_i\): the loss function of the \(i\)-th task (such as distortion classification, compensation parameter generation);

[0112] \(\omega_i\): the weight parameter used to balance the losses of each task;

[0113] \(R(\Theta)\): the regularization term used to prevent overfitting;

[0114] \(\mu\): the weight of the regularization term;

[0115] \(\Theta\): the parameters of the deep learning model;

[0116] By combining multi-task learning with an online learning mechanism, the system can simultaneously optimize multiple tasks such as distortion classification and compensation parameter generation, and use real-time feedback data to improve the model adaptability. Compared with single-task learning methods, this algorithm has stronger generalization ability.

[0117] As can be seen from the above, the system combines a deep learning algorithm and a multi-band dynamic feature extraction method to achieve high-precision recognition of complex non-linear distortion types such as harmonic distortion and dynamic distortion. Compared with traditional spectrum analysis or static filter compensation, this system can dynamically adjust the compensation strategy and significantly reduce the impact of distortion on the sound quality;

[0118] Using a vacuum tube simulation module, through a non-linear even harmonic enhancement algorithm, the natural timbre and auditory texture of the audio signal are highly restored. Through the GAN framework, intelligent repair of signal distortion is realized. The generator dynamically generates high-fidelity compensation signals, and at the same time, the discriminator learns the characteristics of different distortion types to ensure that the repaired signal is close to the distortion-free target. This mechanism has strong self-adaptability and generalization ability and can handle a variety of complex audio scenarios;

[0119] Reinforcement learning optimizes the PWM modulation parameters, so that the signal still maintains low harmonic distortion and high signal-to-noise ratio during high-power amplification. Through the real-time feedback learning module, the model parameters are dynamically updated during use to continuously adapt to environmental changes and user needs. This function ensures the stability and efficiency of the system during long-term operation; in addition, the system shows excellent effects in aspects such as harmonic distortion control, signal-to-noise ratio improvement, and dynamic range expansion, fully meeting various audio processing requirements from high-fidelity audio to live performances.

[0120] Embodiment 2:

[0121] This design is specifically applied to the power amplifier distortion compensation in a high-fidelity home audio system;

[0122] Furthermore, the system composition and parameter configuration:

[0123] Input signal acquisition module

[0124] Sampling rate: 96 kHz

[0125] Quantization precision: 24 bit

[0126] Input signal: Analog audio (20 Hz – 20 kHz)

[0127] Deep learning distortion recognition module

[0128] Using a convolutional neural network (CNN) model: 5 layers of convolution + 2 layers of fully connected

[0129] Training data: Including 500,000 pieces of audio data with different distortion types

[0130] Parameter configuration: Dynamically adjust the distortion compensation weight (α = 0.5)

[0131] Vacuum tube simulation module

[0132] Even harmonic parameters:

[0133] β = 0.03, γ = 0.002

[0134] Analog signal: Use the adaptive harmonic algorithm to generate a signal to enhance the timbre.

[0135] Distortion compensation module

[0136] GAN model structure: Generator (6 layers of convolution + ReLU), discriminator (4 layers of convolution + Sigmoid)

[0137] Reconstruction loss weight: λ = 10

[0138] Repair result: Harmonic distortion reduced to 2% of the original.

[0139] PWM generation and optimization module

[0140] PWM modulation frequency: 400 kHz

[0141] Reinforcement learning optimization: γ = 0.9, learning rate η = 0.01

[0142] Feedback learning module

[0143] Real-time feedback interval: Optimize the model parameters every 5 minutes

[0144] Task weights: Distortion classification (ω1 = 0.6), compensation parameter generation (ω2 = 0.4)

[0145] Test data and effects

[0146] Test environment: A 20-square-meter living room, the speaker frequency response range is 30 Hz – 22 kHz

[0147] Input signal: High-fidelity audio file (undistorted reference standard)

[0148] Output signal measurement: Using an audio analyzer (distortion, frequency response)

[0149] As shown in Table 1 below

[0150] Test Items Before Optimization After Optimization Total Harmonic Distortion (THD) 3.5％ 0.5％ Signal-to-Noise Ratio (SNR) 85 dB 96 dB Tone Reproduction Degree 80％ 98％ User Satisfaction Medium to High Extremely High

[0151] Table 1

[0152] As can be seen from the above, by implementing this system, the sound quality of high-fidelity audio equipment has been greatly improved, the sense of distortion has been significantly reduced, and at the same time, users' feedback on the performance of bass and mid-high frequencies is more natural.

[0153] Example 3:

[0154] Another aspect of this design is applied to power amplifier distortion compensation in live performance audio processing;

[0155] Among them, system composition and parameter configuration:

[0156] Input signal acquisition module

[0157] Sampling rate: 48 kHz

[0158] Quantization accuracy: 16 bit

[0159] Input signal: Live mixing audio signal (with a large dynamic range)

[0160] Deep learning distortion recognition module

[0161] Using a Transformer model

[0162] Training data: Including 100,000 common audio distortion samples of live performances

[0163] Parameter configuration: Dynamic range compression ratio α = 0.7

[0164] Vacuum tube simulation module

[0165] Even harmonic parameters:

[0166] β = 0.05, γ = 0.005

[0167] Simulate the distortion characteristics of real vacuum tubes and enhance the mid-frequency performance.

[0168] Distortion compensation module

[0169] GAN model structure: Generator (7-layer convolution + BatchNorm), discriminator (3-layer fully connected + Dropout) Reconstruction loss weight: λ = 5

[0170] Repair result: Dynamic distortion reduced to 1%.

[0171] PWM Generation and Optimization Module

[0172] PWM modulation frequency: 500 kHz

[0173] Reinforcement learning optimization: γ = 0.8, learning rate η = 0.02

[0174] Feedback Learning Module

[0175] Real-time feedback interval: Optimize model parameter task weights every 30 seconds: Distortion classification (ω1 = 0.5), compensation parameter generation (ω2 = 0.5)

[0176] Test data and effect test environment: Medium-sized performance venue for 500 people, maximum output power of the power amplifier is 1000 W

[0177] Input signal: Multi-track mixed input signal (drums, guitar, vocals, bass)

[0178] Output signal measurement: Use on-site audio test equipment to detect frequency response and distortion

[0179] As shown in Table 2 below

[0180] Test Items Before Optimization After Optimization Total Harmonic Distortion (THD) 7.0％ 1.0％ Signal-to-Noise Ratio (SNR) 75 dB 88 dB Sound Pressure Level Uniformity ±6 dB ±2 dB User Satisfaction Medium High

[0181] Table 2

[0182] As can be seen from the above, the system effectively solves the common dynamic distortion problem in on-site high-power amplification, while optimizing sound uniformity and dynamic performance, especially outstanding in the reverberation processing of vocals and musical instruments, greatly improving the auditory experience of the audience.

[0183] In the description of this specification, the description referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without conflict, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0184] In the attached drawings of the disclosed embodiments of the present invention, only the structures related to the disclosed embodiments are involved, and other structures can refer to the general design. Without conflict, the same embodiment and different embodiments of the present invention can be combined with each other.

[0185] Although embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A power amplifier distortion compensation system based on deep learning, characterized in that: include: An input signal acquisition module is used to acquire analog audio signals and convert them into digital signals through an analog-to-digital converter to obtain audio digital signals; A deep learning distortion recognition module, used to analyze the audio digital signal using a deep learning model, identify the type of distortion produced by the power amplifier, and generate corresponding distortion compensation parameters; The electron tube simulation module is used to add the even harmonic characteristics generated by the simulated electron tube to the audio digital signal to generate an enhanced signal; A distortion compensation module, used to process the enhanced signal based on the distortion compensation parameter, eliminate the distortion and restore the original characteristics of the audio signal, and generate a repair signal; A PWM generation and optimization module, used for converting the repair signal into a pulse width modulation signal for power amplification; A power amplifier module, used to amplify the pulse width modulation signal and drive the speaker to produce sound and output an audio signal; The feedback learning module is used to collect the audio signal output by the speaker, compare it with the audio digital signal, and generate feedback data for optimizing the distortion compensation capability of the deep learning model.

2. The power amplifier distortion compensation system based on deep learning according to claim 1, characterized in that: The deep learning distortion recognition module uses a multi-band dynamic feature extraction algorithm to extract distortion features, and adaptively generates distortion compensation parameters based on frequency bands and dynamic ranges.

3. The power amplifier distortion compensation system based on deep learning according to claim 2, characterized in that: The formula of the multi-band dynamic feature extraction algorithm is: Where Dk: distortion level of frequency band k; Xk(t): the amplitude of the original signal in frequency band k; Target undistorted signal predicted by deep learning; T: sampling time window size; α: a trade-off term used to control the contribution of dynamic changes to the degree of distortion; Var(Xk(t)): Dynamic change of frequency band k.

4. The power amplifier distortion compensation system based on deep learning according to claim 1, characterized in that: The electron tube simulation module simulates the nonlinear characteristics of the electron tube in real time through a digital signal processing algorithm, and adopts an even-order harmonic enhancement algorithm to dynamically generate even-order harmonics and enhance the naturalness of the sound quality of the signal.

5. The power amplifier distortion compensation system based on deep learning according to claim 4, characterized in that: The formula of the even harmonic enhancement algorithm is: Y(t)=X(t)+β·X 2 (t)+γ·X 4 (t) Where Y(t): output enhanced audio signal; X(t): input audio signal; β,γ: weight parameters that control the intensity of even harmonics, dynamically adjusted based on deep learning.

6. The power amplifier distortion compensation system based on deep learning according to claim 1, characterized in that: The distortion compensation module combines with the generative adversarial network to perform high-precision repair on the enhanced signal and dynamically compensate for harmonic distortion and noise distortion.

7. The power amplifier distortion compensation system based on deep learning according to claim 6, characterized in that: The formula for generating the adversarial network is: Generator objective function: L G =-Ε[log(D(G(X)))]+λLrec Discriminator objective function: L D =-Ε[log(D(X))]-Ε[log(1-D(G(X)))] Where G: generator, used to repair the distorted signal; D: Discriminator, used to distinguish the repaired signal from the undistorted target signal; X: input distorted signal; Lrec: reconstruction loss, used to constrain the similarity between the repaired signal and the target signal; λ: Weight parameter that controls the impact of reconstruction loss on the overall objective function.

8. The power amplifier distortion compensation system based on deep learning according to claim 1, characterized in that: The PWM generation and optimization module optimizes PWM modulation parameters based on deep reinforcement learning; The formula for deep reinforcement learning is: Where Q(st,at): the value function of executing action at in state st; η: learning rate; rt: reward after executing the action; γ: Discount factor used to balance short-term and long-term benefits.

9. The power amplifier distortion compensation system based on deep learning according to claim 1, characterized in that: The feedback learning module updates the parameters of the deep learning model in real time through an online learning mechanism, adapts to different audio environments and power amplifier characteristics, and improves the compensation capability of the system.

10. The power amplifier distortion compensation system based on deep learning according to claim 9, characterized in that: The formula for real-time updating of the online learning mechanism is: Where Li: the loss function of the i-th task; ωi: weight parameter used to balance the losses of each task; R(Θ): Regularization term, used to prevent overfitting; μ: regularization term weight; Θ: Parameters of the deep learning model.

Images