Adaptive environment sound effect optimization method for vehicle-mounted intelligent sound box

Abstract

The invention discloses an adaptive environment sound effect optimization method for a vehicle-mounted intelligent sound box, and relates to the technical field of sound effect optimization. The method comprises the following steps: carrying out noise analysis and identification on background audio in a vehicle to obtain noise audio characteristics; playing audio content is obtained, audio analysis and recognition are carried out, and playing audio features are obtained; performing feature influence analysis by using the noise audio features and the playing audio features to determine audio influence information; performing feature optimization compensation on the playing audio features to obtain compensation control parameters; constructing a user playing feature library according to the playing adjustment record; and carrying out constraint adjustment on the compensation control parameters by taking the playing characteristics as constraint conditions, determining an audio optimization adjustment scheme, and carrying out sound effect optimization control on the audio content played by the vehicle-mounted sound box. The technical problem that in the prior art, audio playing in a vehicle-mounted environment is interfered by noise in a vehicle, and consequently the sound effect quality is reduced is solved, and the technical effect of improving the sound effect quality is achieved by conducting sound effect optimization on the audio content.

Description

Technical Field

[0001] The present invention relates to the technical field of sound effect optimization, and in particular to a method for optimizing the sound effect of an in-vehicle intelligent speaker in an adaptive manner. Background Art

[0002] With the rapid development of automobile intelligence, in-car entertainment systems have become an indispensable part of the modern driving experience. As a core component, in-car smart speakers not only play music and provide navigation prompts, but also become an important tool for improving driving pleasure and riding comfort. However, in real-life car use scenarios, the interior environment is often filled with various noises, such as the low-frequency roar of the engine, wind noise at high speeds, road noise caused by friction between tires and the road surface, and traffic noise from the outside world. These noises not only interfere with the driver's attention, but also seriously weaken the sound quality of the in-car smart speakers, making the originally beautiful music unclear and difficult to achieve the expected auditory enjoyment. Summary of the invention

[0003] The present application provides a method for optimizing the adaptive ambient sound effects of an in-vehicle smart speaker, which solves the technical problem in the prior art that the audio playback in the in-vehicle environment is disturbed by the noise inside the car, resulting in a decrease in the sound quality.

[0004] In view of the above problems, the present application provides a method for optimizing the adaptive ambient sound effects of an in-vehicle smart speaker.

[0005] The present application provides a method for optimizing the adaptive ambient sound effects of an in-vehicle smart speaker, the method comprising: The in-car sound is collected through the in-car monitoring equipment, and the collected in-car background audio is subjected to noise analysis and identification to obtain the noise audio characteristics; the in-car smart speaker is connected to obtain the playback audio content, and the playback audio content is subjected to audio analysis and identification to obtain the playback audio characteristics; the noise audio characteristics and the playback audio characteristics are used to perform feature impact analysis to determine the audio impact information, and the audio impact information includes the impact characteristics and the impact description parameters; the playback audio characteristics are subjected to feature optimization compensation according to the impact characteristics and the impact description parameters to obtain the compensation control parameters; the user information is obtained, and a user playback feature library is constructed according to the playback adjustment record of the user information; the playback characteristics in the user playback feature library are used as constraints to perform constraint adjustment on the compensation control parameters to determine the audio optimization adjustment scheme, and the audio optimization adjustment scheme is used to perform sound effect optimization control on the audio content played by the in-car speaker.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: The in-car sound is collected by the in-car monitoring equipment, and the collected in-car background audio is subjected to noise analysis and identification to obtain the noise audio features. Further, the in-car smart speaker is connected to obtain the playback audio content, and the audio content is subjected to audio analysis and identification to obtain the playback audio features. Then, the noise audio features and the playback audio features are used to perform feature impact analysis to determine the audio impact information, and the audio impact information includes the impact features and the impact description parameters. Then, the playback audio features are feature optimized and compensated according to the impact features and the impact description parameters to obtain the compensation control parameters. By obtaining user information, a user playback feature library is constructed according to the playback adjustment records of the user information. Finally, the playback features in the user playback feature library are used as constraints to adjust the compensation control parameters, and the audio optimization adjustment scheme is determined. The audio optimization adjustment scheme is used to optimize the sound effect of the audio content played by the in-car speaker. The technical problem that the audio playback in the in-car environment is interfered by the in-car noise in the prior art, resulting in a decrease in the sound quality, is solved, and the technical effect of improving the sound quality is achieved by optimizing the sound effect of the audio content. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, 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 invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0008] Figure 1 A schematic diagram of a method for optimizing the adaptive ambient sound effects of an in-vehicle smart speaker provided in an embodiment of the present application; Figure 2 A flowchart of noise analysis and identification in a method for optimizing the adaptive environmental sound effects of an in-vehicle smart speaker provided in an embodiment of the present application. DETAILED DESCRIPTION

[0009] The present application solves the technical problem in the prior art that the audio playback in the vehicle environment is disturbed by the noise inside the vehicle, resulting in a decrease in sound quality, by providing a method for optimizing the adaptive ambient sound effects of an in-vehicle smart speaker.

[0010] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0011] It should be noted that the terms "including" and "having" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or are inherent to these processes, methods, products or devices.

[0012] Examples, such as Figure 1 As shown, the embodiment of the present application provides a method for optimizing the adaptive environmental sound effects of a vehicle-mounted smart speaker, wherein the method includes: The in-car sound is collected through in-car monitoring equipment, and the collected in-car background audio is subjected to noise analysis and identification to obtain noise audio features.

[0013] The in-car sounds, including passenger conversations, road noise, engine noise, wind noise, tire noise, etc., are collected through in-car monitoring equipment (such as microphone arrays, sound sensors, etc.), and the collected in-car background audio is subjected to noise analysis and identification. Audio analysis technology (such as spectrum analysis, time-frequency analysis, machine learning algorithms, etc.) can be used to analyze the in-car background audio, identify and separate the background noise, including the vehicle itself (such as the engine, air-conditioning system), the external environment (such as traffic noise, wind noise) or other non-target sound sources in the car (such as passengers' chatting), thereby extracting the noise audio features, including the noise's spectral distribution, energy distribution, duration, etc.

[0014] Furthermore, if Figure 2 As shown, the noise analysis and recognition of the collected in-vehicle background audio includes: Construct a training data set, including audio data with labels, wherein the labels are music labels and voice labels; use the training data set to train the machine learning model to converge, and obtain a noise discrimination model, wherein the noise discrimination model is used to output a noise audio recognition result after removing the labeled audio recognition; use the noise discrimination model to perform noise recognition on the collected in-vehicle background audio to obtain noise audio.

[0015] Preferably, audio data, including music data and vocal data, is extracted from a database, and each type of audio data is labeled, such as a music label and a vocal label, and the labeled audio data is classified and organized according to the labels to form a training data set; according to the audio classification characteristics, a machine learning model is selected, such as a convolutional neural network, a support vector machine, etc., and the selected machine learning model is trained using the training data set. During the training process, the parameters of the model (such as learning rate, batch size, training rounds, etc.) are continuously adjusted until the performance of the model reaches a stable and acceptable level. The model is considered to have converged, that is, the training is completed, and a noise discrimination model is obtained. The noise discrimination model is used to identify and eliminate the labeled audio in the input audio data and then output the noise audio recognition result; by inputting the collected background audio in the car into the noise discrimination model, the noise discrimination model will automatically analyze the audio, identify and eliminate the labeled audio, and output the recognition result of the noise audio.

[0016] Furthermore, obtaining the noise audio feature includes: The noise audio is subjected to wavelet transform to convert the time domain signal into the frequency domain signal to obtain a spectrum diagram of the noise; frequency interval segmentation is performed based on the spectrum diagram of the noise; and frequency components and intensity features of each partition are extracted according to the segmented spectrum intervals to obtain the noise audio features.

[0017] Preferably, the noise audio is subjected to a wavelet transform. Specifically, the noise audio signal is subjected to a multi-level wavelet decomposition to obtain wavelet coefficients at different scales (frequencies), the wavelet coefficients at each level are reconstructed to obtain time domain signals of different frequency bands, and the reconstructed signal of each wavelet decomposition level is subjected to a fast Fourier transform (FFT) or a short-time Fourier transform (STFT), and a graph showing how the amplitude or power of the spectrum changes over time, i.e., a spectrogram, is calculated and plotted; the spectrogram is divided into different frequency intervals according to the spectral characteristics of the noise so that the frequency components in each interval can be analyzed separately; for each frequency interval, the average frequency, center frequency, bandwidth and other parameters therein are calculated, and at the same time, the total energy or average energy in each interval is calculated as an intensity feature, thereby obtaining the noise audio feature.

[0018] Connect to the in-vehicle smart speaker to obtain the playback audio content, perform audio analysis and recognition on the playback audio content, and obtain the playback audio features.

[0019] By connecting to the in-vehicle smart speaker, the audio content being played can be obtained, and audio analysis and recognition can be performed on the audio content being played. For example, the time domain audio signal can be converted into a frequency domain representation through Fourier transform (such as FFT) to analyze the spectral characteristics of the audio, and then the audio features being played can be extracted from the audio signal, such as pitch, rhythm, timbre, energy distribution, etc.

[0020] Furthermore, performing audio analysis and recognition on the played audio content to obtain the played audio features includes: Obtain a spectrum diagram of the played audio content, identify the frequency intensity and frequency components of the frequency domain signal of the played audio, and obtain standard playback audio features; obtain attribute information in the played audio, and set attribute audio constraints according to the attribute information; set an audio amplification interval based on the attribute audio constraints, and perform interval amplification on the spectrum diagram of the played audio content; identify the frequency components and intensity features based on the interval-amplified spectrum diagram, mark the constrained audio features, fit them to the standard playback audio features, and obtain the playback audio features.

[0021] Preferably, a fast Fourier transform (FFT) is used to convert the time domain signal of the played audio into a frequency domain signal, thereby generating a spectrogram of the played audio content, which displays the energy distribution of the audio signal at different frequencies; by analyzing the amplitude of each frequency point in the spectrogram, it is determined which frequencies have higher energy (i.e., intensity); at the same time, the peaks in the spectrogram are identified, which correspond to the main frequency components in the audio, such as the fundamental frequency and harmonics, to obtain standard playback audio features, which include the main frequency components and their corresponding intensities; the attribute information in the played audio may include metadata of the audio (such as song name, artist, genre, etc.), and specific constraints on audio processing are set based on the attribute information. For example, if the audio is classical music, you may want to keep its original pitch and timbre; if it is pop music, you may pay more attention to the rhythm and bass; analyze the spectrum graph, combine the attribute audio constraints, determine which frequency intervals need to be amplified or adjusted, and set the audio amplification intervals. These intervals may be based on specific parts of the audio (such as the climax), specific frequency ranges (such as the bass or treble area), or based on the needs of dynamic range compression; within the selected audio amplification interval, adjust the amplitude of the corresponding frequency points in the spectrum graph to achieve the audio interval amplification effect; re-analyze the spectrum graph after interval amplification, identify new frequency components and intensity features, and mark those audio features affected by the attribute audio constraints, and merge or adjust the marked constrained audio features with the original standard playback audio features to ensure that the final playback audio features meet both the original characteristics of the audio and the specific attribute audio constraints.

[0022] The noise audio feature and the playback audio feature are used to perform feature impact analysis to determine audio impact information, where the audio impact information includes impact features and impact description parameters.

[0023] Compare the noise audio features with the playback audio features, including comparing the key features of the two, such as frequency components, intensity distribution, and spectrum shapes. Through comparative analysis, it is possible to identify which playback audio features are significantly affected by the noise. These impact features may include energy attenuation in a specific frequency band, distortion of frequency components, distortion of spectrum shapes, etc.; for the identified impact features, the impact degree of noise on them can be further quantified by calculating the change amount, change rate, or relative error of the impact features before and after the presence of noise. Based on the quantification results, the descriptive parameters of the audio impact information are determined, including the type of impact (such as energy attenuation, frequency distortion, etc.), the range of impact (such as a specific frequency band), and the degree of impact (such as attenuation, distortion rate, etc.). For example, in the frequency band from XX Hz to YY Hz, the energy is attenuated by ZZ dB, and at a specific frequency point, the frequency component is offset by AA Hz, etc.; the identified impact features and their descriptive parameters are integrated into complete audio impact information.

[0024] Furthermore, performing feature impact analysis using the noise audio feature and the playback audio feature to determine audio impact information includes: Comparing the frequency components in the noise audio feature and the playback audio feature to determine an overlapping frequency region, where the overlapping frequency region is a region where the noise has a masking effect on the playback audio; quantifying the overlapping frequency band strength of the overlapping frequency region, and calculating the weakening influence degree, wherein the weakening influence degree is the ratio of the overlapping frequency band strength to the playing audio intensity feature; Based on the attenuation impact, audio feature screening and positioning are performed to determine the impact characteristics and impact description parameters, and the audio impact information is obtained, wherein the impact characteristics are the frequency range in which the attenuation impact reaches a preset threshold, including high frequency, medium frequency, and low frequency, and the impact description parameters are the specific frequency range, frequency intensity, and attenuation impact of the affected frequency band.

[0025] Preferably, a detailed comparative analysis is performed on the frequency components in the noise audio features and the playback audio features, and the frequency spectra of the two can be compared to find the overlapping parts in the frequency distribution, so as to determine the overlapping frequency areas. These areas are the areas where the noise and the playback audio interfere with each other in frequency, and are also the areas where the noise may have a masking effect on the playback audio; in the overlapping frequency area, the intensity of the noise audio and the playback audio are quantified respectively (usually a measure of energy or amplitude), and the overlapping frequency band intensity is obtained by weighted calculation, and then the weakening influence is obtained by calculating the ratio of the overlapping frequency band intensity to the intensity characteristic of the playback audio in the area. The weakening influence is an indicator to measure the degree of influence of noise on the playback audio in the overlapping frequency area; according to the weakening influence The calculation results are used to screen out those frequency ranges whose attenuation effects reach or exceed the preset thresholds. These ranges are the areas where noise has a significant impact on the played audio; the specific frequency ranges, frequency strengths and attenuation effects of these affected frequency bands are further analyzed to determine the impact characteristics and impact description parameters. The impact characteristics refer to the frequency ranges whose attenuation effects reach the preset thresholds, usually including different frequency bands such as high frequency, medium frequency and low frequency. The impact description parameters are the specific frequency ranges, frequency strengths and attenuation effects of the affected frequency bands; the impact characteristics and impact description parameters are integrated to obtain the audio impact information. The audio impact information contains all important information about the impact of noise on the played audio, such as the affected frequency bands, frequency strengths, attenuation effects, etc.

[0026] Feature optimization compensation is performed on the playback audio feature according to the impact feature and the impact description parameter to obtain a compensation control parameter.

[0027] After obtaining the impact characteristics and impact description parameters of noise on audio playback, it is necessary to optimize and compensate the audio playback characteristics to reduce or eliminate the negative impact of noise. Specifically, based on the impact characteristics and impact description parameters, a targeted compensation strategy is designed, including enhancing the intensity of the audio playback in the affected frequency band, adjusting the frequency response curve, applying dynamic range compression or expansion, etc. After adjustment and optimization, a set of compensation control parameters is finally obtained. The compensation control parameters define the specific operations used for optimization compensation, such as gain adjustment amount, frequency response curve parameters, dynamic range compression ratio, etc., to ensure that the audio playback can still maintain good sound quality when disturbed by noise.

[0028] Furthermore, performing feature optimization compensation on the playback audio feature according to the impact feature and the impact description parameter to obtain the compensation control parameter includes: Perform standard quantization conversion of sound effect parameters according to the playback audio characteristics to determine the sound effect adjustment parameter baseline; perform difference measurement between the impact characteristics and the impact description parameters and the sound effect adjustment parameter baseline; take the parameter quantization difference as the target, perform feature compensation based on the sound effect adjustment parameter baseline, and obtain compensation control parameters for each sound effect parameter.

[0029] Furthermore, the sound effect parameters include: volume adjustment, equalizer adjustment, audio range adjustment, audio gain control, and spatial stereo adjustment.

[0030] Preferably, the sound effect parameters include volume adjustment, equalizer adjustment, audio range adjustment, audio gain control, and spatial stereo adjustment; according to the characteristics of the played audio (such as frequency response, dynamic range, timbre, etc.), the sound effect parameters are converted into standardized quantized values, that is, using industry standards or custom quantization standards, the continuous changes of the sound effect parameters are mapped to discrete quantization levels, and after standard quantization conversion, a sound effect adjustment parameter baseline is obtained, which represents the default sound effect settings in the absence of noise influence or other specific adjustment requirements; according to the impact characteristics and impact description parameters (such as affected frequency bands, weakening impact, etc.), the difference measurement is performed with the sound effect adjustment parameter baseline, For each sound effect parameter, the difference between its actual value and the baseline value is calculated, and the calculated parameter quantization difference is used as the target. These differences represent the sound effect deviations that need to be eliminated or reduced through feature compensation; based on the parameter quantization difference, analyze which sound effect parameters need to be adjusted and the direction and degree of adjustment; based on the sound effect adjustment parameter baseline, perform feature compensation on the sound effect parameters that need to be adjusted, including increasing or decreasing the gain of a certain frequency band of the equalizer, adjusting the threshold and ratio of the compressor, changing the depth and time of the reverberation, etc.; after feature compensation, the compensation control parameters of each sound effect parameter are obtained, which represent the specific values ​​of the sound effect adjustment required to eliminate or reduce the impact of noise. Through the above steps, the compensation control parameters of the sound effect parameters can be accurately determined according to the characteristics of the played audio and the impact characteristics of the noise, thereby optimizing the audio quality.

[0031] User information is obtained, and a user playback feature library is constructed according to the playback adjustment records of the user information.

[0032] By obtaining user information such as age, gender, region, etc., different user accounts can be set up, and data can be collected from the playback history based on different user accounts, including playback time, playback content (such as music type, movies, audiobooks, etc.), playback volume, equalizer settings, etc.; through statistical analysis, common patterns and trends of users in playback adjustments, that is, user preference settings or experience data for playback sound effects, for example, a user may always lower the volume and enhance the bass at night, or prefer specific equalizer settings when listening to classical music, and then build a user playback feature library, which includes the playback volume, equalizer settings, sound effect modes, etc. preferred by different users.

[0033] The playback features in the user playback feature library are used as constraints to constrain the compensation control parameters and determine an audio optimization adjustment scheme, which is used to optimize the sound effects of the audio content played by the vehicle speaker.

[0034] In the process of applying the playback features in the user playback feature library to audio optimization and adjustment, these playback features can be used as constraints to guide the adjustment of compensation control parameters, so as to develop a sound optimization adjustment plan for the audio content played by the car speaker. Specifically, an optimization algorithm, such as a particle swarm algorithm, a genetic algorithm, etc., can be used to take each control parameter as an optimization parameter or particle, and the original parameter adjustment deviation currently set for the playback audio content is minimized, and the optimization direction is searched to be closest to satisfying the user's preference characteristics. Through continuous optimization iteration until convergence, such as reaching the maximum number of iterations, etc., a sound optimization adjustment plan is finally obtained, and then the sound optimization adjustment plan is used to optimize the audio content played by the car speaker.

[0035] Further, including: Using the constrained audio features to perform constraint matching on the audio optimization adjustment scheme, and when the matching result satisfies the constraint condition, determining the audio optimization adjustment scheme; When the matching result does not satisfy the constraint condition, an adjustment target relationship is established based on the constraint condition and the constrained audio feature, and constrained audio adjustment is performed on the audio optimization adjustment scheme according to the adjustment target relationship.

[0036] Preferably, each parameter in the audio optimization adjustment scheme (such as volume, equalizer setting, sound effect mode, etc.) is compared one by one with the corresponding constraint conditions in the constrained audio features, wherein the constraint conditions refer to the value range of each parameter; according to the comparison results, it is determined whether the audio optimization adjustment scheme fully satisfies all the constraint conditions; if the matching result shows that the audio optimization adjustment scheme satisfies all the constraint conditions, then the scheme can be determined as the final scheme and used to optimize the sound effect of the audio content played by the vehicle speaker; if the matching result does not meet the constraint conditions, then based on the constraint conditions and the constrained audio features, the target parameters for adjustment and the expected effect are determined, and then the adjustment target relationship is established, and the adjustment target relationship refers to the constraint conditions and the current deviation value; according to the adjustment target relationship, the direction and amplitude to be adjusted can be determined, so as to make targeted adjustments to the relevant parameters in the audio optimization adjustment scheme, and during the adjustment process, the stability of other parameters that have met the constraint conditions is maintained to avoid introducing new mismatch problems, and repeated testing and adjustment are performed until all parameters meet the constraint conditions. Through the above process, it can be ensured that the audio optimization adjustment scheme not only meets the personalized needs of users, but also presents the best sound effect in the vehicle environment.

[0037] In summary, the embodiments of the present application have at least the following technical effects: The in-car sound is collected by the in-car monitoring equipment, and the collected in-car background audio is subjected to noise analysis and identification to obtain the noise audio features. Further, the in-car smart speaker is connected to obtain the playback audio content, and the audio content is subjected to audio analysis and identification to obtain the playback audio features. Then, the noise audio features and the playback audio features are used to perform feature impact analysis to determine the audio impact information, and the audio impact information includes the impact features and the impact description parameters. Then, the playback audio features are feature optimized and compensated according to the impact features and the impact description parameters to obtain the compensation control parameters. By obtaining user information, a user playback feature library is constructed according to the playback adjustment records of the user information. Finally, the playback features in the user playback feature library are used as constraints to adjust the compensation control parameters, and the audio optimization adjustment scheme is determined. The audio optimization adjustment scheme is used to optimize the sound effect of the audio content played by the in-car speaker. The technical problem that the audio playback in the in-car environment is interfered by the in-car noise in the prior art, resulting in a decrease in the sound quality, is solved, and the technical effect of improving the sound quality is achieved by optimizing the sound effect of the audio content.

[0038] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0039] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0040] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A method for optimizing the adaptive environmental sound effects of an in-vehicle intelligent speaker, characterized in that: The method for optimizing the adaptive environmental sound effects of the vehicle-mounted intelligent speaker comprises: The in-car sound is collected by in-car monitoring equipment, and the collected in-car background audio is analyzed and identified to obtain noise audio features; Connecting to the in-vehicle smart speaker to obtain the played audio content, performing audio analysis and recognition on the played audio content, and obtaining the played audio features; Performing feature impact analysis using the noise audio feature and the playback audio feature to determine audio impact information, where the audio impact information includes impact features and impact description parameters; Perform feature optimization compensation on the playback audio feature according to the impact feature and the impact description parameter to obtain a compensation control parameter; Acquire user information, and build a user playback feature library according to the playback adjustment record of the user information; The playback features in the user playback feature library are used as constraints to constrain the compensation control parameters and determine an audio optimization adjustment scheme, which is used to optimize the sound effects of the audio content played by the vehicle speaker.

2. The method for optimizing the adaptive environmental sound effect of an in-vehicle intelligent speaker according to claim 1, characterized in that: The noise analysis and recognition of the collected in-vehicle background audio includes: Construct a training data set, including audio data with labels, such as music labels and vocal labels; The training data set is used to train the machine learning model to converge, and a noise discrimination model is obtained, wherein the noise discrimination model is used to output a noise audio recognition result after removing the label audio recognition; The noise discrimination model is used to perform noise recognition on the collected in-vehicle background audio to obtain noise audio.

3. The method for optimizing the adaptive environmental sound effect of an in-vehicle intelligent speaker according to claim 2, characterized in that: The step of obtaining the noise audio feature comprises: Performing wavelet transform on the noise audio to convert the time domain signal into a frequency domain signal to obtain a spectrum diagram of the noise; Performing frequency interval segmentation based on the frequency spectrum of the noise; The frequency components and intensity features of each partition are extracted according to the segmented spectrum intervals to obtain the noise audio features.

4. The method for optimizing the adaptive environmental sound effect of an in-vehicle intelligent speaker according to claim 1, characterized in that: Performing audio analysis and recognition on the played audio content to obtain played audio features includes: Obtaining a spectrum diagram of the played audio content, identifying the frequency intensity and frequency components of the frequency domain signal of the played audio, and obtaining standard played audio features; Acquire attribute information in the played audio, and set attribute audio constraints according to the attribute information; Setting an audio amplification interval based on the attribute audio constraint, and performing interval amplification on the frequency spectrum of the played audio content; The frequency components and intensity features are identified according to the interval-enlarged spectrum diagram, the constrained audio features are marked, and the standard playback audio features are fitted to obtain the playback audio features.

5. The method for optimizing the adaptive environmental sound effect of an in-vehicle intelligent speaker according to claim 2, characterized in that: Performing feature impact analysis using the noise audio feature and the playback audio feature to determine audio impact information includes: Comparing the frequency components in the noise audio feature and the playback audio feature to determine an overlapping frequency region, where the overlapping frequency region is a region where the noise has a masking effect on the playback audio; quantifying the overlapping frequency band strength of the overlapping frequency region, and calculating the weakening influence degree, wherein the weakening influence degree is the ratio of the overlapping frequency band strength to the playing audio intensity feature; Based on the attenuation impact, audio feature screening and positioning are performed to determine the impact characteristics and impact description parameters, and the audio impact information is obtained, wherein the impact characteristics are the frequency range in which the attenuation impact reaches a preset threshold, including high frequency, medium frequency, and low frequency, and the impact description parameters are the specific frequency range, frequency intensity, and attenuation impact of the affected frequency band.

6. The method for optimizing the adaptive environmental sound effect of an in-vehicle intelligent speaker according to claim 5, characterized in that: Performing feature optimization compensation on the playback audio feature according to the impact feature and the impact description parameter to obtain a compensation control parameter includes: Performing standard quantization conversion of sound effect parameters according to the playback audio characteristics to determine a sound effect adjustment parameter baseline; Performing difference measurement based on the impact characteristics and impact description parameters and the sound effect adjustment parameter baseline; The parameter quantization difference is taken as a target, feature compensation is performed based on the sound effect adjustment parameter baseline, and compensation control parameters of each sound effect parameter are obtained.

7. The method for optimizing the adaptive environmental sound effect of an in-vehicle intelligent speaker according to claim 6, characterized in that: The sound effect parameters include: volume adjustment, equalizer adjustment, audio range adjustment, audio gain control, and spatial stereo adjustment.

8. The method for optimizing the adaptive environmental sound effects of an in-vehicle intelligent speaker according to claim 4, characterized in that: include: Using the constrained audio features to perform constraint matching on the audio optimization adjustment scheme, and when the matching result satisfies the constraint condition, determining the audio optimization adjustment scheme; When the matching result does not satisfy the constraint condition, an adjustment target relationship is established based on the constraint condition and the constrained audio feature, and constrained audio adjustment is performed on the audio optimization adjustment scheme according to the adjustment target relationship.

Images