Automobile DSP power amplifier sound effect intelligent processing system

Abstract

The invention discloses an intelligent sound effect processing system for an automobile DSP (Digital Signal Processor) power amplifier, particularly relates to the field of optimization of a vehicle-mounted sound system, is used for solving the problem of insufficient sound effect consistency and intelligent adaptation in a dynamic sound field, and is characterized in that acoustic data are acquired in real time through a microphone array, and a three-dimensional acoustic geometric model is constructed by combining ray tracing and an image source method; sound wave propagation paths and dynamic changes are accurately captured; frequency bands of low-frequency, intermediate-frequency and high-frequency signals are divided, spectrum distribution differences and gain fluctuation centers of all seats are evaluated, and a sound field consistency adjustment coefficient is generated and used for guiding design and optimization of a directional beam forming filter, adjusting amplitudes and phases of all sound channels, forming directional sound beams and suppressing noise interference; therefore, the method can dynamically adapt to the acoustic environment change in the vehicle, improves the sound effect balance and high fidelity, and enhances the auditory experience and satisfaction of passengers.

Description

Technical Field

[0001] The present invention relates to the field of vehicle-mounted audio system optimization, and more specifically, to an automobile DSP power amplifier sound effect intelligent processing system. Background Art

[0002] The in-car sound field design directly affects the car's sound performance and is an important factor in improving the driving experience. However, as an acoustic environment with a limited and complex geometric structure, the sound field characteristics of the in-car space are dynamically affected by many factors, including seat layout, sound absorption and reflection characteristics of interior materials, and the number and position distribution of passengers. Dynamic conditions in the car, such as adjustments to seat positions, changes in passenger distribution, or changes in luggage loading conditions, will change the sound propagation path and reflection pattern, resulting in an unbalanced sound field distribution. In such a dynamic environment, traditional sound design techniques usually rely on static acoustic models, which assume that the sound field characteristics remain stable under fixed conditions and are difficult to cope with real-time changes in sound field characteristics in actual scenarios.

[0003] Existing DSP sound effect adjustment technology has significant limitations in adaptability to complex dynamic sound fields in the car. First, it is difficult for traditional methods to perceive in real time the changes in the sound field characteristics of the car space caused by seat adjustment or passenger movement, resulting in insufficient compensation for sound effects at different positions. For example, low-frequency enhancement may occur in the driver's seat area, and high-frequency loss may occur in the rear passenger area, causing consistency problems in sound effects at different positions. Secondly, the existing technology lacks the ability to intelligently adjust the dynamic changes in the sound field in the car, and cannot provide accurate compensation strategies based on the real-time sound field distribution. This results in significant differences in the perception of the same sound effect between the driver and the passenger, especially in high-end audio systems, which significantly weakens the user's sound experience. Therefore, in response to changes in the sound field characteristics in the car, it has become a key issue to be solved to propose an intelligent sound effect adjustment technology that can analyze the dynamic sound field distribution in real time, accurately generate adjustment parameters and respond quickly. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides an intelligent processing system for automobile DSP power amplifier sound effects, which uses a microphone array to collect acoustic data in real time, and combines a three-dimensional acoustic geometric model constructed by ray tracing and image source method to accurately capture the sound wave propagation path and its dynamic changes in space; on this basis, by dividing the frequency bands of low-frequency, medium-frequency and high-frequency audio signals, and evaluating the spectrum distribution differences and gain fluctuation center positions of each seat in different frequency bands, by measuring the distribution discreteness and gain offset of each frequency band, a sound field consistency adjustment coefficient is generated to guide the design and optimization of directional beamforming filters; based on the sound field consistency adjustment coefficient, the position of the main audio target sound source in the car is identified, the amplitude and phase of each channel are designed and adjusted, and a directional sound beam pointing to the target sound source is formed, while effectively suppressing noise interference in non-target directions; and then by dynamically adapting to changes in the acoustic environment in the car, the balance and high fidelity of the sound effects in the car are significantly improved, and the auditory experience and satisfaction of the passengers are enhanced, so as to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] The car DSP amplifier sound effect intelligent processing system includes: data acquisition module, 3D modeling module, frequency band analysis module and beam optimization module;

[0007] Data acquisition module: uses microphone array and acoustic sensor and transmits the collected data to the central processing unit;

[0008] 3D modeling module: Based on the data collected from multiple points, the central processing unit uses ray tracing and image source method to construct a 3D acoustic geometry model inside the car, and calculates the reflection path and propagation delay of sound waves on various surfaces inside the car;

[0009] Frequency band analysis module: divides the audio signal into low frequency, medium frequency and high frequency according to the predetermined frequency bands, measures the distribution dispersion and gain offset of each frequency band based on the spectrum distribution difference and gain fluctuation center position of each seat in the three-dimensional acoustic geometry model, generates the sound field consistency adjustment coefficient for evaluating the frequency band gain adjustment effect, and transmits the sound field consistency adjustment coefficient to the beam optimization;

[0010] Beam optimization module: Based on the sound field consistency adjustment coefficient results, the position of the main audio target sound source in the car is identified, and a directional beamforming filter is designed to form a directional sound beam pointing to the target sound source by adjusting the amplitude and phase of each channel, while suppressing noise interference in non-target directions.

[0011] In a preferred embodiment, the 3D modeling module includes the following contents:

[0012] First, all the collected point data are sorted by spatial position Pi =(x i ,y i ,z i ) to perform unified spatial coordinate calibration, and construct the interior space framework in a three-dimensional Cartesian coordinate system; then, the following specific methods are applied to complete the construction of the three-dimensional acoustic geometry model and the calculation of the reflection path:

[0013] According to the design parameters of the interior structure, a 3D geometric mesh model G containing all acoustic surfaces is created. j , where j represents the spatial index of each surface; each surface is labeled with specific attribute parameters A j =(r j ,c j ,θ j ), where r j is the reflection coefficient, c j is the absorption coefficient, θ j is the direction angle of the surface normal vector;

[0014] From each sound source point S k The sound wave path ray is emitted to the mesh surface, and the distance between the ray and each surface mesh G is calculated. j The intersection point Q k,j and the incident angle φ k,j ; The intersection calculation formula is: Q k,j =S k +t·d; where d is the ray direction vector and t is the path parameter, which is solved by the geometric constraint equation; record the path length L of each ray k,j =||S k -Q k,j ||; and propagation time T k,j =L k,j / υ, where υ is the speed of sound;

[0015] Intersection point Q based on first reflection k,j , generate a virtual mirror point M corresponding to the sound source point k,j , calculate the multiple reflection path from the mirror point to other surfaces; complete the multiple reflection path R through a recursive algorithm m,n The length and propagation delay of the path are calculated; the path recursive formula is: And accumulate the propagation time

[0016] The phase difference of all rays and reflected path superposition signals is calculated according to the propagation time T k,j and frequency f, through the phase formula: Δφ k,j =2πf·T k,j ; Calculate the phase offset of each path and accumulate it to obtain the phase superposition effect of the entire sound field;

[0017] The results are integrated to construct a three-dimensional acoustic geometry model M containing information on path length, propagation time, number of reflections, and phase change. acoustic , and record all its parameters in matrix form.

[0018] In a preferred embodiment, the frequency band analysis module includes the following contents:

[0019] Based on the original audio signal in the car, the audio signal is divided into three intervals: low frequency LF, medium frequency MF, and high frequency HF according to the preset boundary frequency;

[0020] For each large interval, several sub-bands are divided to describe the energy distribution within each frequency band; let BAND b,u represents the u-th subband in frequency band b, where b∈{LF,MF,HF} and u is the subband index;

[0021] Combined with the three-dimensional acoustic geometry model, the propagation attenuation coefficient and material absorption characteristics of each seat h in each sub-band are corrected to obtain the corrected sub-band energy E h,b,u ;

[0022] In order to measure the discrete degree of energy distribution of each seat in different frequency bands, a multi-layer threshold progressive coverage calculation method is first defined:

[0023] First, set the threshold sequence {ν for each seat and subband 1 ,ν 2 ,…,ν m}, where ν K Incremental and follows a geometric or logarithmic distribution to capture the change from low to high energy distribution;

[0024] For the threshold value ν K and the corrected subband energy E h,b,u , define a subband coverage determination function:

[0025] Then, the coverage judgment results of all sub-bands in the same frequency band are accumulated to obtain the coverage judgment result of seat h within the threshold value ν K The following subband covers and:

[0026] Using the multi-level fractal measurement method, we get the distribution dispersion FDI h,b : in, It is a nonlinear index used to magnify or reduce the difference in subband coverage as the threshold increases.

[0027] In a preferred embodiment, the frequency band analysis module further includes the following contents:

[0028] To evaluate gain offset, first define a gain relative scalar: Among them, G h,b is the current gain setting for seat h in band b, is a predefined ideal reference value, and tanh(·) is used to map the gain offset to the interval [0,1);

[0029] Then, to quantify the connectivity of the corresponding offsets between each subband, the gain offset connectivity is calculated: GDC h,b =exp(∑log(1+Z h,b )).

[0030] In a preferred embodiment, the frequency band analysis module further includes the following contents:

[0031] The distribution dispersion and gain offset connectivity are integrated in frequency band b to obtain the sound field deviation index at the corresponding seat frequency band level: h,b =(FDI h,b ) κ1 +ln(1+GDC h,b ); where κ1 is a nonlinear adjustment coefficient, which is used to control the contribution of distribution dispersion in this deviation index; finally, in order to comprehensively evaluate the sound field consistency of the seat in the low, medium and high frequency bands, the sound field consistency adjustment coefficient SCA is defined h : SCA h =∑ b∈{LF,MF,HF} Π h,b ; Arrange the calculation results of the sound field consistency adjustment coefficients of all seats according to the seat index h to form a one-dimensional sequence {SCA 1 ,SCA 2 ,…}.

[0032] In a preferred embodiment, the propagation attenuation coefficient and material absorption characteristics of each seat in each sub-band are corrected by the following method: first, based on the total length of the sound wave in the propagation path and the reflection angle of each path, the attenuation factor on the path is calculated, and the total propagation attenuation coefficient of all paths is obtained by integrating all paths; secondly, the absorption coefficient of each reflecting surface recorded in the three-dimensional model is queried, and the energy loss of each reflection point in the path is cumulatively calculated; finally, the initial energy value of each seat in the sub-band is multiplied by the propagation attenuation coefficient and the absorption loss correction coefficient in turn to obtain the corrected sub-band energy, which is used to describe the frequency band energy distribution under real propagation conditions.

[0033] In a preferred embodiment, the beam optimization module includes the following contents:

[0034] First, according to the sequence {SCA 1 ,SCA 2 ,…}, select the target seat index that exceeds the consistent threshold and has the largest value as hcrit , marking it as the priority optimization target seat; in the three-dimensional acoustic geometry model, the spatial coordinates of the target seat are recorded as The coordinate of the speaker in the car is C spk ; Define the target direction The target direction indicates the direction from the center of the speaker to the target seat;

[0035] In order to realize directional sound beam, the phase of each channel signal under different time delays is superimposed and controlled; suppose the array contains N independent channels, and their positions in the three-dimensional acoustic geometric model are A n ,n=1,2,…,N;

[0036] Define the phase shift function Used to perform phase correction on the nth channel in the frequency domain so that it is in the target direction Produce coherent superposition and try to achieve phase cancellation in other directions; the constraints are recorded as Where <·,·> represents the vector dot product, O represents the set of interference directions that are perpendicular or contrary to the target direction, Δ n Represents the combined correction of the nth channel in the phase dimension and small coordinate offset; this constraint ensures that the beam energy in the target direction is higher than that in other interference directions.

[0037] In a preferred embodiment, the beam optimization module further includes the following contents:

[0038] In addition to phase control, the amplitude of each channel output is distributed; define the amplitude function Where ζ is the amplification factor and τ represents the discrete time index; is the amplitude adjustment factor;

[0039] At the same time, in order to suppress the side lobe energy, a side lobe domain is defined within a certain angle range from the target direction vector. If an amplitude superposition trend higher than the standard is detected in the side lobe domain, the amplitude function of the corresponding channel is lowered;

[0040] In the time domain, the direction and amplitude of the main lobe are not constant. In order to adapt to the changes in the target position caused by the movement of passengers or the change of seat angles, dynamic beam width correction is performed;

[0041] The instantaneous value of the main lobe beam width is set to be the sum of the initially set basic beam width and the correction amount that changes with time; if the target position is detected to be displaced, the beam width range is automatically expanded or contracted, and the phase shift function is updated accordingly;

[0042] The dynamic beam width correction process monitors the energy distribution of the directivity diagram output by the array in the high frequency band and the medium frequency band. If the main lobe leaves the target area, the instantaneous value of the main lobe beam width and the amplification factor in the amplitude function are adjusted immediately, and the phase shift function is fine-tuned synchronously to redirect the main lobe peak.

[0043] Finally, a directional beamforming filter is defined and synthesized on all frequency bands respectively, and a filter coefficient group is generated separately for the frequency band b∈{LF, MF, HF} to obtain a multi-band directional beamforming filter. The multi-channel signal output by the multi-band directional beamforming filter forms a phase superposition and amplitude gain enhancement effect in the target direction, and presents phase cancellation and amplitude attenuation in the non-target direction, thereby improving the sound field consistency and listening performance of key seats.

[0044] The technical effects and advantages of the automobile DSP power amplifier sound effect intelligent processing system of the present invention are as follows:

[0045] The present invention utilizes a microphone array to collect acoustic data in real time, and combines a three-dimensional acoustic geometric model constructed by ray tracing and image source method, which can accurately capture the sound wave propagation path and its dynamic changes in space; on this basis, by dividing the frequency bands of low-frequency, medium-frequency and high-frequency audio signals, and evaluating the spectrum distribution differences and gain fluctuation center positions of each seat in different frequency bands, and by measuring the distribution discreteness and gain offset of each frequency band, a sound field consistency adjustment coefficient is generated to guide the design and optimization of directional beamforming filters; based on the sound field consistency adjustment coefficient, the position of the main audio target sound source in the car is identified, the amplitude and phase of each channel are designed and adjusted, and a directional sound beam pointing to the target sound source is formed, while effectively suppressing noise interference in non-target directions; and then by dynamically adapting to changes in the acoustic environment in the car, the balance and high fidelity of the sound effects in the car are significantly improved, and the auditory experience and satisfaction of the passengers are enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a structural schematic diagram of the automobile DSP power amplifier sound effect intelligent processing system of the present invention. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] Embodiment 1: Figure 1 The present invention provides an automobile DSP power amplifier sound effect intelligent processing system, comprising:

[0049] Data acquisition module: uses microphone array and acoustic sensor and transmits the collected data to the central processing unit;

[0050] 3D modeling module: Based on the data collected from multiple points, the central processing unit uses ray tracing and image source method to construct a 3D acoustic geometry model inside the car, and calculates the reflection path and propagation delay of sound waves on various surfaces inside the car;

[0051] Frequency band analysis module: divides the audio signal into low frequency, medium frequency and high frequency according to the predetermined frequency bands, measures the distribution dispersion and gain offset of each frequency band based on the spectrum distribution difference and gain fluctuation center position of each seat in the three-dimensional acoustic geometry model, generates the sound field consistency adjustment coefficient for evaluating the frequency band gain adjustment effect, and transmits the sound field consistency adjustment coefficient to the beam optimization;

[0052] Beam optimization module: Based on the sound field consistency adjustment coefficient results, the position of the main audio target sound source in the car is identified, and a directional beamforming filter is designed to form a directional sound beam pointing to the target sound source by adjusting the amplitude and phase of each channel, while suppressing noise interference in non-target directions.

[0053] The optimization of in-car sound effects depends on accurate sound field data collection, and the integrity and real-time performance of sound field data directly determine the accuracy of subsequent sound field modeling and optimization. In the complex acoustic environment of the car, the sound pressure level, spectrum characteristics and phase information of different seating areas show significant dynamic differences, and are significantly affected by vehicle speed, external environmental noise and in-car layout. In order to support the construction of a three-dimensional sound field model, it is necessary to design real-time acquisition technology in a targeted manner to ensure the acquisition and transmission of high-precision data.

[0054] The data acquisition module includes the following:

[0055] By arranging microphone arrays and acoustic sensors at key locations in the car, such as the front, middle, rear seats and doors, a multi-point distributed acquisition network is formed that covers the main acoustic areas of the entire car. Each acquisition point is sampled synchronously at equal intervals, and a unified clock reference is used to achieve time synchronization of multi-point acoustic signals through a timestamp calibration method to avoid data misalignment.

[0056] Each microphone and sensor unit adopts a wide-band design, covering the acquisition range of low-frequency to high-frequency acoustic signals, and eliminates external interference noise (such as wind noise and tire noise) through a built-in filter module. The data is digitized by a high-precision analog-to-digital converter (ADC) at a preset sampling rate of 192kHz or higher to ensure the complete recording of spectral characteristics. In order to improve the real-time and integrity of data transmission, a high-speed communication bus (such as CAN or Ethernet) is used to transmit the collected data to the central processing unit.

[0057] The acoustic data collected in real time include sound pressure level (used to characterize the intensity of the sound field), spectrum characteristics (used to describe the frequency distribution), and phase information (used to calculate the sound wave propagation path and interference). During the transmission process, the data stream is integrated through a distributed cache mechanism to support subsequent modeling calculations.

[0058] After completing the above processing, the collected multi-point acoustic data are stored in the central processing unit in a high-precision and high-time form. These data provide the necessary frequency, intensity and phase basic information for the construction of the 3D sound field model in the subsequent 3D modeling module, especially the dynamic changes of the acoustic characteristics of each region, which will be directly used as the input parameters for the calculation of reflection path and propagation delay in the 3D geometric model.

[0059] Through high-precision and comprehensive acoustic data collection, the accuracy of subsequent three-dimensional sound field modeling is ensured, thereby providing accurate original data support for the system to optimize the in-car sound effects.

[0060] In the data acquisition module, the sound pressure level, spectrum characteristics and phase information of each area in the car have been collected. The comprehensiveness and real-time nature of these data provide a basis for building an accurate three-dimensional sound field geometric model. The goal of this step is to use these acoustic data to establish a geometric model of the acoustic environment in the car by combining ray tracing and image source methods, and further calculate the propagation path, delay, reflection and interference characteristics of the sound waves, providing accurate geometric and physical basis for subsequent frequency band analysis and dynamic adjustment.

[0061] The 3D modeling module includes the following:

[0062] After the central processing unit receives the sound pressure level, spectrum characteristics and phase information transmitted by the data acquisition module, it first divides all the acquisition point data into spatial positions P i =(x i ,y i ,z i ) to perform unified spatial coordinate calibration, and construct the interior space framework in a three-dimensional Cartesian coordinate system. Subsequently, the following specific methods are applied to complete the construction of the three-dimensional acoustic geometry model and the calculation of the reflection path:

[0063] 1. Preliminary geometric framework construction:

[0064] According to the design parameters of the interior structure (such as the size and material properties of the doors, seats, and roof), a 3D geometric mesh model G containing all acoustic surfaces is created. j , where j represents the spatial index of each surface. Each surface is labeled with specific attribute parameters A j =(r j ,c j ,θ j ), where r j is the reflection coefficient, c jis the absorption coefficient, θ j is the direction angle of the surface normal vector.

[0065] 2. Ray tracing path calculation:

[0066] From each sound source point S k (such as the location of the sound source) to emit sound wave path rays to the mesh surface, and calculate the distance between the rays and each surface mesh G j The intersection point Q k,j and the incident angle φ k,j The intersection point calculation formula is: Q k,j =S k +t·d; where d is the ray direction vector and t is the path parameter, which is solved by the geometric constraint equation.

[0067] Record the path length L of each ray k,j =||S k -Q k,j ||; and propagation time T k,j =L k,j / υ, where υ is the speed of sound.

[0068] 3. Multiple reflections supplement of image source method:

[0069] Intersection point Q based on first reflection k,j , generate a virtual mirror point M corresponding to the sound source point k,j , calculate the multiple reflection path from the mirror point to other surfaces. The multiple reflection path R is completed by a recursive algorithm m,n The length and propagation delay of the path are calculated. The path recursion formula is: And accumulate the propagation time

[0070] 4. Interference and phase difference calculation:

[0071] The phase difference of all rays and reflected path superposition signals is calculated according to the propagation time T k,j and frequency f, through the phase formula: Δφ k,j =2πf·T k,j ; Calculate the phase offset of each path and accumulate it to obtain the phase superposition effect of the entire sound field.

[0072] 5. Output of 3D acoustic geometry model:

[0073] Based on the above results, a three-dimensional acoustic geometric model M containing information such as path length, propagation time, number of reflections, and phase change is constructed. acoustic , and record all its parameters in matrix form: M acoustic ={L k,j ,T k,j ,R m,n ,Δφk,j ,A j}; This model is used to fully reflect the characteristics of the sound field inside the car and provide input for the spectrum distribution difference and gain adjustment in the next step.

[0074] Through the above processing, the central processing unit successfully constructed a three-dimensional acoustic geometry model of the car, and accurately calculated the reflection path and propagation delay of the sound waves on various surfaces in the car. This detailed sound field distribution information will serve as the basis for the spectrum distribution difference and gain adjustment in the frequency band analysis module, ensuring that the dynamic adjustment of the frequency band gain can be based on the precise sound field characteristics, thereby achieving consistency in the frequency response at each seat in the car.

[0075] In the data acquisition module, the system collects the sound pressure level, spectrum characteristics and phase information of each seat through microphone arrays and acoustic sensors, and transmits them to the central processing unit after high-sampling rate analog-to-digital conversion; in the three-dimensional modeling module, the central processing unit builds a three-dimensional acoustic geometry model of the car based on the geometry data of the interior space and combines ray tracing and image source method, and records the sound wave propagation and reflection characteristics of each seat in detail. In order to achieve differentiated analysis and dynamic gain optimization of different seats in the car in the three frequency bands of low frequency, medium frequency and high frequency, this step will use the data of the three-dimensional acoustic geometry model to segment the audio signal and measure the distribution characteristics, and finally obtain the sound field consistency adjustment coefficient that can be used to evaluate and guide the frequency band gain adjustment.

[0076] The frequency band analysis module includes the following:

[0077] 1. Frequency band division and sub-band preprocessing:

[0078] Based on the original audio signal in the car collected by the data acquisition module, the audio signal is divided into three intervals: low frequency LF, medium frequency MF, and high frequency HF according to the preset boundary frequency.

[0079] For each large interval, we further divide it into several sub-bands to more finely characterize the energy distribution within each frequency band. b,u represents the u-th subband in frequency band b, where b∈{LF,MF,HF} and u is the subband index.

[0080] Combined with the 3D acoustic geometry model obtained by the 3D modeling module, the propagation attenuation coefficient and material absorption characteristics of each seat h in each sub-band are corrected to obtain the corrected sub-band energy E h,b,u At this time, E h,b,u Multiple reflections of the sound wave in the corresponding sub-band, path length, incident angle and surface properties have been comprehensively considered.

[0081] The propagation attenuation coefficient and material absorption characteristics of each seat in each sub-band are corrected by the following method: first, based on the total length of the sound wave on the propagation path and the reflection angle of each path, the attenuation factor on the path is calculated, and the total propagation attenuation coefficient of all paths is obtained by integrating all paths; secondly, the absorption coefficient of each reflecting surface recorded in the three-dimensional model is queried, and the energy loss of each reflection point in the path is cumulatively calculated; finally, the initial energy value of each seat in the sub-band is multiplied by the propagation attenuation coefficient and the absorption loss correction coefficient in turn to obtain the corrected sub-band energy, which is used to describe the frequency band energy distribution under real propagation conditions.

[0082] 2. Calculation of subband coverage with multi-layer threshold progression:

[0083] In order to measure the discrete degree of energy distribution of each seat in different frequency bands, a multi-layer threshold progressive coverage calculation method needs to be defined first.

[0084] First, set the threshold sequence {ν for each seat and subband 1 ,ν 2 ,…,ν m}, where ν K Incremental and follows a geometric or logarithmic distribution to capture the change from low to high energy distribution.

[0085] For the threshold value ν K and the corrected subband energy E h,b,u , define a subband coverage determination function:

[0086] This function determines whether the energy exceeds a certain threshold. Then the coverage judgment results of all sub-bands in the same frequency band are accumulated to obtain the coverage of seat h within the threshold ν. K The following subband covers and: Among them, the sub-band coverage and the decreasing trend with the increase of threshold can reflect the distribution pattern of the energy of the seat in this frequency band from low to high, laying the foundation for the subsequent quantification of the discrete degree of distribution.

[0087] 3. Multi-level fractal measurement of distribution dispersion:

[0088] In order to characterize the energy discrete characteristics of the seat in each frequency band, a multi-level fractal measurement method is used to obtain the distribution dispersion FDI h,b : in, It is a nonlinear index used to magnify or reduce the difference in subband coverage as the threshold increases; the larger the distribution dispersion value, the more layers and more unbalanced the energy distribution of seat h in frequency band b is. Since subband coverage depends on the energy threshold sequence and subband coverage, this fractal measure can characterize the dispersion from multiple angles at different energy levels.

[0089] 4. Nonlinear coupling analysis of gain offset connectivity:

[0090] In addition to dispersion, the gain offset also needs to be evaluated, that is, whether the gain distribution of seat h in frequency band b is improperly offset.

[0091] First define a gain relative scalar: Among them, G h,b is the current gain setting for seat h in band b, is a predefined ideal reference value (which can be set according to the target acoustic environment in the 3D modeling module or the average gain level in the vehicle), and tanh(·) is used to map the gain offset to the [0,1) interval.

[0092] Then, to further quantify the connectivity of the corresponding offsets between each subband, the gain offset connectivity is calculated: GDC h,b =exp(∑log(1+Z h,b )); The combination of logarithms and exponentials is chosen to avoid oversmoothing of linear summation or simple multiplication, which can increase significantly when there are large offsets between subbands, thereby highlighting the cumulative effect of improper gain.

[0093] 5. Fusion calculation of sound field consistency adjustment coefficient:

[0094] The distribution dispersion and gain offset connectivity are integrated in frequency band b to obtain the sound field deviation index at the corresponding seat frequency band level: h,b =(FDI h,b ) κ1 +ln(1+GDC h,b ), where κ1 is the nonlinear adjustment coefficient, which is used to control the contribution of distribution dispersion in this deviation index; and ln(1+GDC h,b ) ensures that the effect is quickly amplified when the gain excursion is too high.

[0095] Finally, in order to comprehensively evaluate the sound field consistency of the seats in the low, medium and high frequency bands, the sound field consistency adjustment coefficient SCA is defined h : SCA h =∑ b∈{LF,MF,HF} Π h,b ; When SCA h When it is high, it means that the seat h has high discreteness of energy distribution in multiple frequency bands and is connected to a serious gain offset, which needs to be optimized in subsequent steps; when the value is low, it means that the seat is already in a relatively balanced and reasonable sound field state.

[0096] The calculation results of the sound field consistency adjustment coefficients of all seats are arranged according to the seat index h to form a one-dimensional sequence {SCA 1 ,SCA 2,…} and passed to the beam optimization module.

[0097] Through the detailed calculation of gain offset and energy distribution of all seats, this step successfully constructed the sound field consistency adjustment coefficient. This result will serve as the basis for the next step of priority optimization to guide directional beamforming and dynamic gain adjustment to ensure that the sound field distribution in the car is more balanced and efficiently optimized.

[0098] In the frequency band analysis module, after frequency band division, discreteness analysis and gain offset measurement of each seat, the sound field consistency adjustment coefficient is obtained, which is used to characterize the comprehensive deviation degree of each seat in the three frequency bands of low frequency, medium frequency and high frequency. Based on the numerical distribution of this coefficient, the system can identify which seats have more significant sound field imbalance problems. Based on this result, the current step needs to integrate the three-dimensional acoustic geometry model (from the three-dimensional modeling module) to design and implement directional beamforming filters for the target seats with the most obvious imbalance or the desired optimization, so as to enhance the sound effect in the target direction and suppress the noise in the non-target direction.

[0099] The beam optimization module includes the following:

[0100] 1. Determination of target seat and direction vector:

[0101] First, according to the sequence {SCA 1 ,SCA 2 ,…}, select the target seat index that exceeds the consistent threshold and has the largest value as h crit , marking it as the priority optimization target seat.

[0102] In the three-dimensional acoustic geometry model, the spatial coordinates of the target seat are denoted as R hcrit , let the coordinate of the car speaker (or array center) be C spk . Define the target direction D hcrit =R hcrit -C spk ; The target direction indicates the direction from the center of the loudspeaker to the target seat; in the subsequent filter design, the directional gain and phase distribution need to be determined based on this direction vector.

[0103] 2. Phase shift and multi-angle interference suppression:

[0104] In order to realize directional sound beam, it is necessary to control the phase of each channel signal at different time delays. Assume that the array contains N independent channels, and their positions in the three-dimensional acoustic geometric model are A n ,n=1,2,…,N。

[0105] Define the phase shift function Used to perform phase correction on the nth channel in the frequency domain so that it is in the target direction Produce coherent superposition and try to achieve phase cancellation in other directions. The constraint can be written as Where <·,·> represents the vector dot product, O represents the set of interference directions that are perpendicular or contrary to the target direction, Δ n It represents the comprehensive correction of the nth channel in the phase dimension and the small coordinate offset. This constraint is intended to ensure that the beam energy in the target direction is higher than that in other interference directions.

[0106] 3. Amplitude allocation and side lobe constraints:

[0107] In addition to phase control, the amplitude of each channel output needs to be nonlinearly distributed. Define the amplitude function Where ζ is the amplification factor and τ represents the discrete time index. is the amplitude adjustment factor, which can be calculated by the following formula: in Represents the cosine value of the angle between the position vector of vocal channel nnn and the target direction vector, ranging from [-1,1]; It is the angle between the vocal tract position vector and the target direction vector, which is calculated by the three-dimensional acoustic geometry model; κ1 is the intensity coefficient of amplitude adjustment, which is used to control the influence of the sound field consistency adjustment coefficient on the amplitude factor.

[0108] At the same time, in order to suppress the side lobe energy, the side lobe region S is defined within a certain angle range from the target direction vector side If it is detected that there is an amplitude superposition trend higher than the standard in the side lobe domain, the amplitude function of the corresponding channel will be dynamically lowered to ensure that the main lobe energy is concentrated and the side lobe is suppressed.

[0109] 4. Dynamic beam width correction and phase fine-tuning:

[0110] In the time domain, the direction and amplitude of the main lobe are not constant. In order to adapt to the changes in target position caused by the movement of passengers in the car or the change of seat angle, dynamic beam width correction is required.

[0111] Let Θ(τ) represent the instantaneous value of the main lobe beamwidth, which is the sum of the initial basic beamwidth and the time-varying correction: in is the proportional constant for controlling the correction amplitude, It can be updated in real time based on in-vehicle motion monitoring (seat back angle, passenger displacement) and other information. If the target position is detected to be displaced, the beam width range is automatically expanded or contracted, and the phase shift function is updated accordingly.

[0112] This dynamic correction process can be achieved by monitoring the energy distribution of the directivity diagram output by the array in the high and medium frequency bands. If the main lobe leaves the target area or the side lobe rises too high, the instantaneous value of the main lobe beamwidth and the amplification factor in the amplitude function are adjusted in real time, and the phase shift function is fine-tuned synchronously to redirect the main lobe peak.

[0113] 5. The combined output of the beamforming filter:

[0114] Finally, a directional beamforming filter is defined and synthesized on all frequency bands, and a filter coefficient set is generated separately for the frequency band b∈{LF,MF,HF} The multi-band directional beamforming filter F is obtained by synthesis beam ={F LF ,F MF ,F HF}.

[0115] The multi-channel signal output by the multi-band directional beamforming filter forms a phase superposition and amplitude gain enhancement effect in the target direction, and presents phase cancellation and amplitude attenuation in the non-target direction, thereby improving the sound field consistency and listening performance of key seats.

[0116] Based on the sound field consistency adjustment coefficient output by the frequency band analysis module, this module gives priority to seats that are in an unbalanced or to-be-optimized state, and designs a multi-band directional beamforming filter by determining the target direction vector, phase shift function, nonlinear amplitude distribution, and dynamic beam width correction. The filter concentrates energy in the target direction and suppresses lateral and rearward noise, thereby significantly enhancing the sound effect of the severely unbalanced seat area and effectively improving its sound field.

[0117] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0118] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0119] It should be noted that, in this article, if there are relational terms such as first and second, etc., they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "including a..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0120] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. The car DSP amplifier sound effect intelligent processing system is characterized by: include: Data acquisition module, 3D modeling module, frequency band analysis module and beam optimization module; Data acquisition module: uses microphone array and acoustic sensor and transmits the collected data to the central processing unit; 3D modeling module: Based on the data collected from multiple points, the central processing unit uses ray tracing and image source method to construct a 3D acoustic geometry model inside the car, and calculates the reflection path and propagation delay of sound waves on various surfaces inside the car; Frequency band analysis module: divides the audio signal into low frequency, medium frequency and high frequency according to the predetermined frequency bands, measures the distribution dispersion and gain offset of each frequency band based on the spectrum distribution difference and gain fluctuation center position of each seat in the three-dimensional acoustic geometry model, generates the sound field consistency adjustment coefficient for evaluating the frequency band gain adjustment effect, and transmits the sound field consistency adjustment coefficient to the beam optimization; Beam optimization module: Based on the sound field consistency adjustment coefficient results, the position of the main audio target sound source in the car is identified, and a directional beamforming filter is designed to form a directional sound beam pointing to the target sound source by adjusting the amplitude and phase of each channel, while suppressing noise interference in non-target directions.

2. The automotive DSP power amplifier sound effect intelligent processing system according to claim 1 is characterized in that: The 3D modeling module includes the following: First, all the collected point data are sorted by spatial position P i =(x i ,y i ,z i ) to perform unified spatial coordinate calibration and construct the in-vehicle spatial framework with a three-dimensional Cartesian coordinate system; Then, the following specific methods are applied to complete the construction of the 3D acoustic geometry model and the calculation of the reflection path: According to the design parameters of the interior structure, a 3D geometric mesh model G containing all acoustic surfaces is created. j , where j represents the spatial index of each surface; each surface is labeled with specific attribute parameters A j =(r j ,c j ,θ j ), where r j is the reflection coefficient, c j is the absorption coefficient, θ j is the direction angle of the surface normal vector; From each sound source point S k The sound wave path ray is emitted to the mesh surface, and the distance between the ray and each surface mesh G is calculated. j The intersection point Q k,j and the incident angle φ k,j ; The intersection calculation formula is: Q k,j =S k +t·d; where d is the ray direction vector and t is the path parameter, which is solved by the geometric constraint equation; record the path length L of each ray k,j =||S k -Q k,j ||; and propagation time T k,j =L k,j / υ, where υ is the speed of sound; Intersection point Q based on first reflection k,j , generate a virtual mirror point M corresponding to the sound source point k,j , calculate the multiple reflection path from the mirror point to other surfaces; complete the multiple reflection path R through a recursive algorithm m,n The length and propagation delay of the path are calculated; the path recursive formula is: And accumulate the propagation time The phase difference of all rays and reflected path superposition signals is calculated according to the propagation time T k,j and frequency f, through the phase formula: Δv k,j =2πf·T k,j ; Calculate the phase offset of each path and accumulate it to obtain the phase superposition effect of the entire sound field; The results are integrated to construct a three-dimensional acoustic geometry model M containing information on path length, propagation time, number of reflections, and phase change. acoustic , and record all its parameters in matrix form.

3. The automotive DSP power amplifier sound effect intelligent processing system according to claim 2 is characterized in that: The frequency band analysis module includes the following: Based on the original audio signal in the car, the audio signal is divided into three intervals: low frequency LF, medium frequency MF, and high frequency HF according to the preset boundary frequency; For each large interval, several sub-bands are divided to describe the energy distribution within each frequency band; let BAND b,u represents the u-th subband in frequency band b, where b∈{LF,MF,HF} and u is the subband index; Combined with the three-dimensional acoustic geometry model, the propagation attenuation coefficient and material absorption characteristics of each seat h in each sub-band are corrected to obtain the corrected sub-band energy E h,b,u ; In order to measure the discrete degree of energy distribution of each seat in different frequency bands, a multi-layer threshold progressive coverage calculation method is first defined: First, a threshold sequence {ν1,ν2,…,ν m }, where ν K Incremental and follows a geometric or logarithmic distribution to capture the change from low to high energy distribution; For the threshold value ν K and the corrected subband energy E h,b,u , define a subband coverage determination function: Then, the coverage judgment results of all sub-bands in the same frequency band are accumulated to obtain the coverage judgment result of seat h within the threshold value ν K The following subband covers and: Using the multi-level fractal measurement method, we get the distribution dispersion FDI h,b : in, It is a nonlinear index used to amplify or reduce the difference in subband coverage as the threshold increases.

4. The automotive DSP power amplifier sound effect intelligent processing system according to claim 3 is characterized in that: The frequency band analysis module also includes the following: To evaluate gain offset, first define a gain relative scalar: Among them, G h,b is the current gain setting for seat h in band b, is a predefined ideal reference value, and tanh(·) is used to map the gain offset to the interval [0,1); Then, to quantify the connectivity of the corresponding offsets between each subband, the gain offset connectivity is calculated: GDC h,b =exp(∑log(1+Z h,b )).

5. The automotive DSP power amplifier sound effect intelligent processing system according to claim 4 is characterized in that: The frequency band analysis module also includes the following: The distribution dispersion and gain offset connectivity are integrated in frequency band b to obtain the sound field deviation index at the corresponding seat frequency band level: h,b =(FDI h,b ) κ1 +ln(1+GDC h,b ); where κ1 is a nonlinear adjustment coefficient, which is used to control the contribution of distribution dispersion in this deviation index; finally, in order to comprehensively evaluate the sound field consistency of the seat in the low, medium and high frequency bands, the sound field consistency adjustment coefficient SCA is defined h : SCA h =∑ b∈{LF,MF,HF} Π h,b ; Arrange the calculation results of the sound field consistency adjustment coefficients of all seats according to the seat index h to form a one-dimensional sequence {SCA1, SCA2, …}.

6. The automotive DSP power amplifier sound effect intelligent processing system according to claim 3 is characterized by: The propagation attenuation coefficient and material absorption characteristics of each seat in each sub-band are corrected by the following method: first, based on the total length of the sound wave on the propagation path and the reflection angle of each path, the attenuation factor on the path is calculated, and the total propagation attenuation coefficient of all paths is obtained by integrating all paths; secondly, the absorption coefficient of each reflecting surface recorded in the three-dimensional model is queried, and the energy loss of each reflection point in the path is cumulatively calculated; finally, the initial energy value of each seat in the sub-band is multiplied by the propagation attenuation coefficient and the absorption loss correction coefficient in turn to obtain the corrected sub-band energy, which is used to describe the frequency band energy distribution under real propagation conditions.

7. The automotive DSP power amplifier sound effect intelligent processing system according to claim 5 is characterized in that: The beam optimization module includes the following: First, according to the sequence {SCA1, SCA2, ...}, the target locus index with the largest value exceeding the consistent threshold is selected as h crit , marking it as the priority optimization target seat; in the three-dimensional acoustic geometry model, the spatial coordinates of the target seat are recorded as The coordinate of the speaker in the car is C spk ; Define the target direction The target direction indicates the direction from the center of the speaker to the target seat; In order to realize directional sound beam, the phase of each channel signal under different time delays is superimposed and controlled; suppose the array contains N independent channels, and their positions in the three-dimensional acoustic geometric model are A n ,n=1,2,…,N; Define the phase shift function Used to perform phase correction on the nth channel in the frequency domain so that it is in the target direction Produce coherent superposition and try to achieve phase cancellation in other directions; the constraints are recorded as Where <·,·> represents the vector dot product, O represents the set of interference directions that are perpendicular or contrary to the target direction, Δ n Represents the combined correction of the nth channel in the phase dimension and small coordinate offset; this constraint ensures that the beam energy in the target direction is higher than that in other interference directions.

8. The automotive DSP power amplifier sound effect intelligent processing system according to claim 7 is characterized in that: The beam optimization module also includes the following: In addition to phase control, the amplitude of each channel output is distributed; define the amplitude function in is the magnification factor, τ represents the discrete time index; is the amplitude adjustment factor; At the same time, in order to suppress the side lobe energy, a side lobe domain is defined within a certain angle range from the target direction vector. If an amplitude superposition trend higher than the standard is detected in the side lobe domain, the amplitude function of the corresponding channel is lowered; In the time domain, the direction and amplitude of the main lobe are not constant. In order to adapt to the changes in the target position caused by the movement of passengers or the change of seat angles, dynamic beam width correction is performed; The instantaneous value of the main lobe beam width is set to be the sum of the initially set basic beam width and the correction amount that changes with time; if the target position is detected to be displaced, the beam width range is automatically expanded or contracted, and the phase shift function is updated accordingly; The dynamic beam width correction process monitors the energy distribution of the directivity diagram output by the array in the high frequency band and the medium frequency band. If the main lobe leaves the target area, the instantaneous value of the main lobe beam width and the amplification factor in the amplitude function are adjusted immediately, and the phase shift function is fine-tuned synchronously to redirect the main lobe peak. Finally, a directional beamforming filter is defined and synthesized on all frequency bands, and a filter coefficient group is generated separately for the frequency band b∈{LF,MF,HF} to obtain a multi-band directional beamforming filter; The multi-channel signal output by the multi-band directional beamforming filter forms a phase superposition and amplitude gain enhancement effect in the target direction, and presents phase cancellation and amplitude attenuation in the non-target direction, thereby improving the sound field consistency and listening performance of key seats.

Images