Self-adaptive audio signal processing method and system for inhibiting physical displacement of Bluetooth sound box

Abstract

The invention discloses a self-adaptive audio signal processing method and system for suppressing physical displacement of a Bluetooth sound box. The method monitors system volume gain in real time, and triggers control logic when it exceeds a displacement critical threshold calibrated by hardware coherence parameters. The system utilizes a band-pass filter to separate a target frequency band component corresponding to the physical resonance frequency of the box body, performs energy time domain integration on the component in a time sliding window, and calculates a displacement potential energy accumulation value. And when the accumulated value exceeds a static friction force breakthrough threshold value combined with the voice coil heat accumulation model, generating a dynamic attenuation coefficient to perform smooth suppression on the amplitude of the frequency band. According to the invention, machine-by-machine calibration is carried out on algorithm parameters by using loudspeaker impedance and inductance data actually measured by a precise LCR digital bridge in a production stage, and the suppressed low-frequency hearing feeling is compensated through higher harmonics in combination with a psychoacoustic principle, so that the problem of automatic sliding of the lightweight portable sound box during large dynamic playing is effectively solved.

Description

Technical Field

[0001] This invention relates to the field of acoustic structure and audio signal processing technology for consumer electronics, and in particular to an adaptive audio signal processing method and system for suppressing physical displacement of a Bluetooth speaker. Background Technology

[0002] In the engineering practice of portable Bluetooth speakers, there is a significant dynamic contradiction between miniaturizing the physical size and maximizing low-frequency acoustic performance. To achieve deep bass extension, speakers are typically equipped with long-stroke speaker units and high-power amplifiers, but this results in a huge reaction force momentum generated by the speaker vibration system during high dynamic operation. In actual testing and usage scenarios, when the speaker is placed on a smooth surface such as glass, polished marble, or lacquered wooden tabletops, and music tracks containing test signals with high crest factors and sustained high-level low-frequency components are played at maximum volume, the horizontal thrust on the bottom of the speaker can easily overcome the maximum static friction between the feet and the contact surface. This forced vibration manifests as micro-vibration in the short term, but after several tens of minutes of continuous playback, it often accumulates into a macroscopic displacement of more than one centimeter, and in severe cases, it can even cause the device to fall off the table and be damaged.

[0003] Existing solutions typically employ fixed high-pass filters to cut off low frequencies or use global limiters. While these can suppress displacement, they severely sacrifice bass quality, resulting in a flat and weak sound. Furthermore, in large-scale industrial production, the DC impedance and inductance parameters of speaker drivers objectively exhibit statistical dispersion. Different batches of drivers, driven by the same voltage, produce inconsistent mechanical thrust. Algorithm parameters fixed based on standard prototypes often fail to adapt to specific machines that require slightly lower impedance or higher compliance, leading to higher driving force and causing displacement faults in some mass-produced products at the user end. Simultaneously, as playback time increases, the voice coil temperature rises, causing thermal drift in the resistor. The compliance of the suspension system also increases due to material softening, making the speaker stable when cold but causing displacement when hot due to decreased damping. Static algorithms cannot dynamically compensate for this. Summary of the Invention

[0004] The purpose of this invention is to provide an adaptive audio signal processing method and system for suppressing the physical displacement of a Bluetooth speaker, so as to solve the problems pointed out in the background art.

[0005] In a first aspect, the present invention provides an adaptive audio signal processing method for suppressing physical displacement of a Bluetooth speaker. The method is applied to a Bluetooth speaker system including a digital signal processor. The method includes acquiring a digital audio signal to be played in real time and monitoring the current system volume gain of the Bluetooth speaker.

[0006] The feature is that when the system volume gain exceeds a preset displacement critical threshold, anti-displacement control logic is triggered, and the anti-displacement control logic includes the following steps:

[0007] A target resonant frequency band component is separated from the digital audio signal using a bandpass filter. The target resonant frequency band component corresponds to the physical resonant frequency range of the Bluetooth speaker's enclosure.

[0008] Set a time sliding window, and perform time-domain integration calculation on the energy of the target resonant frequency band component within the time sliding window to obtain the current cumulative displacement potential energy value;

[0009] The accumulated value of displacement potential energy is compared with the preset static friction force breakthrough threshold in real time.

[0010] If the accumulated displacement potential energy exceeds the static friction force threshold, a dynamic attenuation coefficient is generated. The amplitude of the component in the digital audio signal located in the target resonant frequency band is smoothed and suppressed using the dynamic attenuation coefficient, so that the suppressed accumulated displacement potential energy is maintained below the static friction force threshold.

[0011] Optionally, the frequency range of the target resonant frequency band component is pre-calibrated and configured in the digital signal processor in the following manner:

[0012] The total harmonic distortion curve generated when the Bluetooth speaker is scanned across the entire frequency band by an audio analyzer;

[0013] Identify the frequency corresponding to the peak of the total harmonic distortion curve in the low-frequency region;

[0014] Using the frequency as the center and combining it with a preset bandwidth factor, the upper and lower cutoff frequencies of the target resonant frequency band components are defined.

[0015] Optionally, the step of smoothing and suppressing the amplitude using the dynamic attenuation coefficient employs an asymmetric time control strategy:

[0016] A rapid start-up time at the millisecond level is set so that when the accumulated displacement potential energy exceeds the limit, the dynamic attenuation coefficient can take effect immediately to cut off the accumulation of kinetic energy.

[0017] By setting a slow release time on the order of hundreds of milliseconds, the gain of the target resonant frequency band component is slowly restored after the displacement risk is eliminated, thus avoiding a sudden, fluctuating breathing effect in the audio listening experience.

[0018] Optionally, the method further includes a look-ahead buffering step:

[0019] The digital audio signal is stored in a first-in-first-out buffer of a preset length;

[0020] The time-domain integration calculation is performed on the audio data at the beginning of the buffer, while the smoothing suppression is performed on the audio data at the end of the buffer.

[0021] The time difference of the buffer is used to predict the trend of the cumulative displacement potential energy in advance, and the dynamic attenuation coefficient is applied in advance before the actual physical displacement occurs.

[0022] Optionally, the length of the time sliding window is configured to match the sustain period of common low-frequency instruments to distinguish between transient percussion signals and steady-state continuous low-frequency signals, ensuring that suppression is triggered only for audio components with sustained drive.

[0023] Optionally, the method further includes a psychoacoustic low-frequency compensation step:

[0024] While suppressing the amplitude of the target resonant frequency band component, the time-domain fundamental frequency signal data of the target resonant frequency band component is extracted;

[0025] The second and third harmonic components of the fundamental frequency are generated using a nonlinear harmonic generator.

[0026] The compensation gain is calculated based on the magnitude of the dynamic attenuation coefficient, and the second and third harmonic components are superimposed on the digital audio signal to reconstruct the suppressed low-frequency loudness in hearing by utilizing the fundamental frequency loss effect of the human ear.

[0027] Optionally, the center frequency of the target resonant frequency band component and the displacement critical threshold are calibrated based on hardware consistency parameters;

[0028] The calibration process includes: retrieving the impedance and inductance characteristic values ​​of the speaker unit stored in the non-volatile memory of the Bluetooth speaker, wherein the impedance and inductance characteristic values ​​are statistical data obtained by measuring the speaker unit based on an LCR digital bridge;

[0029] Calculate the deviation ratio between the impedance characteristic value and the standard design value;

[0030] The default resonant center frequency is corrected for frequency deviation based on the deviation ratio, and the displacement critical threshold is compensated for sensitivity to adapt to the physical vibration characteristics of the current hardware batch.

[0031] Optionally, the static friction force exceeding the threshold has adaptive thermal attenuation characteristics to compensate for thermal drift caused by prolonged playback:

[0032] The digital signal processor calculates the high-power continuous playback duration of the Bluetooth speaker in real time.

[0033] A thermal accumulation model of the speaker voice coil is established, and the voice coil temperature and compliance change of the suspension system are estimated based on the continuous playback duration.

[0034] As the voice coil temperature increases, the static friction force exceeding the threshold is automatically reduced according to the preset thermal compensation curve to prevent displacement loss of control due to the decrease in the damping performance of the loudspeaker.

[0035] Optionally, the thermal accumulation model is also used to dynamically adjust the quality factor Q value of the bandpass filter:

[0036] When the continuous playback duration is short and the cumulative displacement potential energy is low, a high quality factor Q value is maintained to minimize the impact on the sound quality of non-resonant frequency bands.

[0037] When the continuous playback duration exceeds the preset heat saturation time point, the quality factor Q value is automatically reduced to broaden the filter bandwidth, thereby covering the resonant frequency range that is detached due to temperature changes.

[0038] In a second aspect, the present invention provides an adaptive audio signal processing system for suppressing the physical displacement of a Bluetooth speaker, comprising:

[0039] The signal acquisition module is used to acquire the digital audio signal to be played in real time and monitor the system volume gain;

[0040] The logic trigger module is used to activate the subsequent processing unit when the system volume gain exceeds the displacement critical threshold.

[0041] The resonance extraction module is used to separate the target resonance frequency band component through a bandpass filter;

[0042] The potential energy analysis module is used to calculate the cumulative displacement potential energy of the target resonant frequency band component within a time sliding window;

[0043] The threshold decision module is used to compare the accumulated displacement potential energy value with the static friction force exceeding the threshold.

[0044] The suppression execution module is used to generate a dynamic attenuation coefficient and suppress the signal amplitude when the cumulative displacement potential energy exceeds the limit;

[0045] The system also includes a non-volatile memory storing hardware consistency parameters generated based on LCR bridge test data, which are used to calibrate the filtering parameters of the resonance extraction module.

[0046] The present invention has achieved the following beneficial effects:

[0047] This invention effectively solves the problem of anti-displacement algorithm failure caused by the discrete parameters of individual loudspeakers in mass production by introducing a hardware consistency calibration mechanism based on physically measured data into the signal processing link. During the production phase, the system utilizes the impedance and inductance characteristic values ​​obtained from a full inspection of the loudspeakers using a precision LCR digital bridge to precisely correct the resonant center frequency and judgment threshold on a machine-by-machine basis. This mechanism ensures that each piece of equipment leaving the factory can automatically adjust its control strategy according to its own physical hardware characteristics, eliminating performance fluctuations caused by component tolerances and guaranteeing the consistency and stability of batch products.

[0048] This invention constructs a displacement potential energy analysis model based on energy time-domain integration, achieving accurate differentiation between transient percussion and sustained actuated low frequencies. The system sets a monitoring frequency band based on the cabinet resonance frequency determined by an audio analyzer scan. Suppression is only triggered when the energy accumulation in the target frequency band is about to exceed the static friction critical point, while maintaining signal pass-through during normal music playback. Combined with psychoacoustic low-frequency compensation technology, the algorithm physically reduces the fundamental frequency energy causing displacement while simultaneously using a nonlinear harmonic generator to generate second and third harmonics, reconstructing low-frequency loudness at the auditory level. This achieves a dynamic balance between physical anti-shake and high-fidelity sound quality.

[0049] Furthermore, this invention integrates a speaker voice coil thermal accumulation model, endowing the system with adaptive capabilities to cope with prolonged high-power playback. Addressing the increased resistivity caused by voice coil heating and thermal softening of the suspension system during continuous playback, the system can adjust the static friction force to exceed the threshold and correct the filter quality factor in real time based on the estimated temperature. This dynamic thermal compensation mechanism effectively prevents late-stage displacement of the speaker due to the degradation of physical damping performance after prolonged operation, ensuring the mechanical stability of the equipment under all time periods and operating conditions.

[0050] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0051] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0052] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0053] Figure 1 This is a flowchart of an adaptive audio signal processing method for suppressing the physical displacement of a Bluetooth speaker, as described in an embodiment of the present invention.

[0054] Figure 2 This is a schematic diagram of an adaptive audio signal processing system for suppressing the physical displacement of a Bluetooth speaker, as described in an embodiment of the present invention. Detailed Implementation

[0055] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0056] In the field of modern consumer electronics, especially for portable Bluetooth speakers, industrial design tends towards extreme miniaturization and lightweighting, while acoustic performance requirements are constantly evolving towards greater dynamic range and lower bass response. This physical constraint between physical volume and acoustic energy leads to a significant engineering contradiction: when a speaker plays high-energy-density low-frequency signals, the mechanical momentum generated by the speaker's vibration system is easily coupled to the lightweight enclosure, generating a reaction force sufficient to overcome the static friction of the base. When playing test tracks containing sustained high-level low-frequency components, such as "Burning" or "Illegal Bass," on a smooth surface (e.g., glass, polished marble, or lacquered tabletop), existing Bluetooth speaker products are prone to unexpected physical displacement. This displacement may only manifest as slight vibrations in a short period, but under prolonged continuous playback (e.g., more than 30 minutes), the cumulative displacement often exceeds 1 cm, and in severe cases, it can even lead to the device being dropped and damaged. To address this problem, embodiments of the present invention provide an adaptive audio signal processing method and system for suppressing physical displacement of Bluetooth speakers.

[0057] Reference Figure 1 As shown, this embodiment of the invention discloses an adaptive audio signal processing method for suppressing the physical displacement of a Bluetooth speaker, such as... Figure 2 As shown, the method operates in a Bluetooth speaker embedded system that includes a high-performance digital signal processor (DSP), non-volatile memory, and a Class D digital power amplifier. The specific steps include the following:

[0058] Step S11: During the system power-on initialization phase, read the hardware consistency parameters based on LCR digital bridge calibration stored in the non-volatile memory, and perform initialization calibration on the displacement critical threshold and resonance extraction filter parameters of the anti-displacement control logic based on the hardware consistency parameters.

[0059] In this embodiment, in existing large-scale industrial manufacturing, due to limitations in raw material characteristics (such as the purity of the voice coil copper wire and the consistency of the magnetic flux density of the magnet) and assembly process tolerances (such as glue weight and assembly gap), the key Thiele / Small parameters of the speaker unit objectively exhibit a statistically discrete distribution. Conventional audio algorithms are usually fixed based on the parameters of a standard prototype (Golden Sample) determined during the R&D phase; however, this strategy cannot cope with the challenges posed by hardware discreteness. For example, if the DC resistance of a speaker in a certain machine is slightly lower than the standard value, its voice coil current will increase significantly under the same driving voltage, according to the Lorentz force formula (…). ,in This refers to the electromagnetic driving force acting on the speaker's voice coil. The magnetic flux density in the magnetic circuit gap, The effective conductor length for the voice coil to cut magnetic field lines. (The real-time current intensity flowing through the voice coil) will generate a mechanical thrust that exceeds design expectations, causing the machine to become a displacement failure in the production batch.

[0060] To address the interference of hardware inconsistencies on the accuracy of the anti-displacement algorithm, this invention introduces a full-inspection calibration mechanism based on precision instruments. The hardware consistency parameters are not factory default values, but rather measured data pre-stored in specific protected sectors of the Bluetooth speaker's non-volatile memory (such as SPI Flash or EEPROM). This data is obtained through precise scanning and measurement of each speaker unit to be assembled using a high-precision LCR digital bridge (preferably the Tonghui TH2811D model) during the individual incoming quality control (IQC) or in-circuit testing (ICT) stages on the production line.

[0061] Specifically, the measurement process strictly follows the Kelvin four-terminal sensing method. The TH2811D digital bridge uses a dedicated test fixture to connect the positive and negative terminals of the speaker to minimize measurement errors introduced by the resistance of the test leads and contact resistance. The test conditions are set as follows: at a standard frequency of 1kHz and a constant level of 1Vrms, the nominal impedance characteristic value of the speaker (denoted as...) is measured. To accurately characterize the DC resistance characteristics of the voice coil; at low frequencies of 100Hz or lower and a level of 1Vrms, the low-frequency inductance characteristic value of the loudspeaker (denoted as ) is measured. ), to characterize the inductive reactance of the voice coil under low-frequency, high-dynamic conditions.

[0062] In step S11, when the Bluetooth speaker system powers on, the digital signal processor first reads the stored data via the internal bus. and And compare it with the standard design parameters (standard design impedance value) preset in the firmware. and standard design inductance value Compare according to the formula. Calculate the hardware deviation ratios for impedance and inductance respectively (where...) To calculate the obtained deviation ratio, Represents the measured feature value read. or , Represents the corresponding standard design value or ).

[0063] The digital signal processor dynamically compensates for the system's preset displacement critical threshold based on the impedance deviation ratio. If the measured impedance... Below the standard value (For example, 3% lower) means that at the same volume gain, the machine will generate a larger drive current and mechanical thrust. To compensate for this physical difference, the digital signal processor automatically generates a negative threshold correction coefficient, proportionally reducing the displacement critical threshold, making the anti-displacement control logic more sensitive on this machine, thus intervening in control earlier.

[0064] Simultaneously, the digital signal processor corrects the center frequency of the bandpass filter in subsequent steps based on the inductance deviation ratio. Physical acoustic principles indicate that the speaker voice coil inductance... Changes in frequency are often accompanied by slight variations in the equivalent mass or magnetic circuit characteristics of the vibrating system, which can lead to changes in the free-field resonant frequency of the loudspeaker driver. Drift occurs. The digital signal processor uses a preset mapping algorithm, based on... The offset is used to fine-tune the center frequency parameters of the bandpass filter (for example, if the inductance is too large, the center frequency is corrected towards the lower frequency direction) to ensure that the filter always accurately covers the actual physical resonance point of the machine, so as to achieve precise control of one machine and one policy.

[0065] Specifically, parameter calibration is performed using the following linear mapping formula:

[0066] The standard design impedance value is set to The measured impedance value is The preset displacement critical threshold (gain value) is Corrected displacement critical threshold The calculation is as follows:

[0067] ;

[0068] in, The default displacement critical threshold preset for the system is represented by a linear voltage gain ratio; This is the calibrated execution threshold. This is the impedance sensitivity coefficient, which takes a positive value. When the measured impedance... Less than the standard value When the result calculated within the parentheses is less than 1, the threshold is lowered. .

[0069] To ensure that the energy clamping control in subsequent step S14 is also compatible with the current hardware characteristics, the system simultaneously calibrates the static friction force exceedance threshold (energy threshold) based on the impedance deviation ratio. The calibrated reference energy threshold... The calculation is as follows:

[0070] ;

[0071] in, This represents the nominal energy threshold determined based on a standard prototype. (Coefficient) This is used to compensate for the physical property that causes the current to increase due to the decrease in impedance, which in turn causes the thrust to increase in a square relationship. This ensures that speakers with different impedances have a consistent mechanical thrust limit when reaching this threshold.

[0072] Similarly, for the calibration of the center frequency of the bandpass filter, the following formula is used to compensate for the resonance drift caused by the inductor deviation:

[0073] ;

[0074] In the formula, This indicates the default bandpass filter center frequency (Hz) measured based on a standard prototype. This indicates the actual execution center frequency (Hz) for the bandpass filter after current hardware calibration. The nominal inductance value (mH) of a standard loudspeaker driver. The measured inductance value (mH) is shown.

[0075] In this formula, Defined as the resonant frequency drift coefficient, its physical significance lies in characterizing inductance. Changes in the equivalent mass of the vibration system Considering the influence of resonant frequency, the recommended value in engineering is approximately 0.5 (based on...). Estimate, of which (This refers to the equivalent mechanical mass of the loudspeaker vibration system). Meanwhile, the impedance sensitivity coefficient in the aforementioned formula... The recommended value range is 0.5 to 1.0, which is used to adjust the system's sensitivity to impedance deviation.

[0076] Step S12: Acquire the digital audio signal to be played in real time, and perform high-frequency monitoring of the current system volume gain of the Bluetooth speaker throughout the day to determine whether to activate the anti-displacement control logic.

[0077] After initial calibration is complete, the system enters the real-time signal processing loop. The digital signal processor receives PCM format digital audio streams from the front-end Bluetooth SoC or USB interface in real time via a high-speed I2S (Inter-IC Sound) or TDM (Time Division Multiplexing) audio bus. To ensure high dynamic range and low noise floor in signal processing, the input audio data is converted into 32-bit or 64-bit floating-point format internally by the DSP.

[0078] Simultaneously, the system establishes a global gain monitoring thread to collect the current digital volume, EQ processing gain, and the fixed gain of the backend analog power amplifier in real time, and calculates the current total voltage gain of the system through multiplication. The system has a preset displacement critical threshold. This threshold is set based on rigorous R&D verification testing: on a standard smooth glass test bench, a specific displacement test signal is played, and a laser displacement sensor monitors the state of the enclosure, recording the critical output power point at which the enclosure begins to produce micro-slippage, and setting the gain value corresponding to this power point as the baseline threshold. It is worth noting that this baseline threshold has been corrected in step S11 according to hardware consistency parameters.

[0079] The digital signal processor (DSP) compares the real-time monitored total system voltage gain with the corrected displacement threshold. If the current total voltage gain is lower than the displacement threshold, the system determines that the current driving energy is insufficient to overcome the static friction at the bottom of the enclosure, and the risk of physical displacement is extremely low. At this time, to ensure optimal audio signal integrity, the anti-displacement control logic is in bypass mode, and the audio signal is output directly. If the current total voltage gain exceeds the displacement threshold, the system determines that it is currently in a high-energy output state, with a potential risk of physical displacement, and immediately activates subsequent in-depth analysis and control logic.

[0080] Step S13: Use a bandpass filter to separate the target resonant frequency band component from the digital audio signal, and set a time sliding window. Perform time-domain integration calculation on the energy of the target resonant frequency band component within the time sliding window to obtain the current displacement potential energy accumulation value.

[0081] The system uses a discrete-time moving average energy algorithm to calculate the displacement potential energy accumulation value (denoted as ). Calculation formula:

[0082] ;

[0083] The current sampling time The calculated cumulative value of displacement potential energy; Index for the current discrete sampling time; The length of the time sliding window (number of sampling points) is configured to correspond to a physical time of 50ms to 100ms to cover the typical sustain period of low-frequency instruments; The backtracking index variable within the sliding window (range of values) arrive ); For corresponding time The target resonant frequency band signal amplitude, i.e., the historical audio data after separation by the bandpass filter; These are the time-weighted window function coefficients, used to perform weighted statistics on the energy within the window.

[0084] In order to obtain accurate energy statistics characteristics, It can be configured to one of the following two modes according to system requirements, and the normalization condition must be met. :

[0085] 1. Rectangular window:

[0086] Suitable for simple moving average calculations. .

[0087] 2. Hanning Window:

[0088] Suitable for scenarios requiring smoother energy statistics and reduced spectral leakage. The calculation results need to be normalized to ensure that the sum of the coefficients is 1.

[0089] The cumulative value The model physically quantifies the effective mechanical driving work obtained by the loudspeaker vibration system within the current time window. By integrating (summing) the square of the amplitude, the model can effectively distinguish between transient percussion signals (short-lived energy) and steady-state low-frequency signals (continuous energy), ensuring that suppression is triggered only for audio components with continuous driving force.

[0090] This step aims to precisely isolate the causal components causing physical displacement from the full-band audio signal. Physical dynamics analysis shows that, as a rigid body system, the Bluetooth speaker's movement on the horizontal plane is not triggered by all low-frequency signals, but rather induced by intense mechanical resonance at a specific frequency. At this specific frequency, the mechanical impedance of the speaker's vibration system is minimized, and the amplitude reaches its maximum, thus transmitting the maximum reaction force to the enclosure.

[0091] The frequency range of the target resonant frequency band components was determined through rigorous acoustic modal analysis. During the R&D phase, the Audio Precision APX525 audio analyzer, in conjunction with the AMP50-D power drive module, was used to perform a stepped frequency sweep of 20Hz to 20kHz on the entire Bluetooth speaker. By analyzing the total harmonic distortion (THD) and noise-frequency sweep curves and the impedance modulus-frequency curves generated by the APX525, a significant distortion peak and impedance peak were identified in the low-frequency region (typically between 60Hz and 100Hz). The frequency corresponding to this peak (e.g., 82Hz) is the physical actuation sensitive frequency of this model. Based on this data, a high-order IIR band-pass filter is configured internally in the digital signal processor. Its center frequency is set to the aforementioned sensitive frequency (and superimposed with the calibration bias from step S11). Its quality factor (Q value) is calculated based on a preset bandwidth factor, which defines the filter's -3dB cutoff frequency range (i.e., passband width) based on the center frequency. The specific formula for converting the Q value is as follows:

[0092] Let the preset bandwidth factor be... (Unit: octaves), then the quality factor Q is calculated using the standard second-order filter bandwidth conversion formula:

[0093] ;

[0094] For example, when the bandwidth is set to 1 / 3 octave ( When ), the calculated result This calculation ensures that the filter's passband width is precisely matched to the physically determined resonant frequency band of the enclosure.

[0095] Furthermore, to distinguish between transient percussion signals and steady-state continuous low-frequency signals (the latter being the primary cause of displacement), this invention introduces a displacement potential energy accumulation calculation model. The digital signal processor sets a time-sliding window in memory, the window length of which is configured to match the typical sustain period of a low-frequency instrument (e.g., set to 50ms to 100ms). Within this window, the sample energy (amplitude squared) of the filtered target resonant frequency band component is integrated in the time domain. This integral value physically simulates the tendency of the enclosure to accumulate mechanical kinetic energy during the current time period, attempting to break free from the constraints of static friction.

[0096] Furthermore, to address the inherent latency issues in digital signal processing, this step incorporates a look-ahead mechanism. The digital audio signal is stored in a first-in, first-out (FIFO) buffer of a preset length (e.g., 5ms). The energy integration analysis is performed on the data at the head of the buffer (i.e., the data about to be played), while subsequent control operations are performed on the data at the tail of the buffer (i.e., the data that needs to be played now). Utilizing this 5ms time difference, the system can anticipate upcoming large dynamic impacts, thereby calculating control parameters before physical vibrations occur, achieving a proactive delayed response.

[0097] Step S14: The accumulated displacement potential energy value is compared with a preset static friction force breakthrough threshold in real time. If the accumulated displacement potential energy value exceeds the static friction force breakthrough threshold, a dynamic attenuation coefficient is generated, and the threshold is adaptively adjusted using a thermodynamic model. Let the static friction force breakthrough threshold be denoted as... The physical dimensions of this threshold are... To maintain consistency, both represent energy (the integral of the square of the amplitude over time). Their initial values... The critical sliding power of the chamber is determined by the integral of the measured critical sliding power in the laboratory and dynamically adjusted according to the thermal model.

[0098] The calculated cumulative displacement potential energy is fed into the logic decision unit and compared with the static friction threshold. This static friction threshold is a dynamic variable with adaptive thermal attenuation characteristics. This design fully considers the changes in the thermophysical properties of the speaker under prolonged high-power operation.

[0099] During rigorous aging tests, such as continuous playback at maximum volume for 30 minutes, the speaker's voice coil temperature rises significantly. Physical characteristics indicate two things: first, the resistivity of the voice coil copper wire increases with temperature (thermal power compression effect), causing a change in the speaker's electromagnetic damping coefficient (Qes); second, the speaker's centering support (spider) becomes more compliant (softer) after being heated and subjected to continuous mechanical movement. This means the enclosure's mechanical control over the vibration system decreases, resulting in a larger vibration amplitude under the same driving force, leading to a sharp increase in the risk of displacement.

[0100] To this end, a loudspeaker voice coil heat accumulation model based on the law of conservation of energy runs inside the digital signal processor.

[0101] Specifically, the thermal accumulation model uses a first-order IIR (Infinite Impulse Response) digital filter to simulate the thermal inertia characteristics of the voice coil. The system calculates the instantaneous power of the input signal in real time. The voice coil temperature rise was estimated based on the difference equation. :

[0102] ;

[0103] in, The voice coil temperature estimated at the previous moment. The cooling coefficient (with a value slightly less than 1) is used to simulate heat dissipation characteristics. This represents the temperature rise coefficient. Additionally, initial conditions for the iteration are set. ,in The ambient temperature detected by the system or the default room temperature setting (e.g., 25°C). ), representing the initial temperature of the voice coil when the speaker is cold-started.

[0104] The specific calculation logic for the above parameters is as follows:

[0105] Instantaneous power Calculations must be based on measured impedance: ;in, represent The instantaneous value of the equivalent physical voltage applied across the speaker at any given time. Its calculation method is as follows: , Normalized digital audio sample point values ​​(range -1.0 to +1.0). The peak output voltage (Volts) of the power amplifier at maximum digital input is determined by the amplifier's supply voltage and gain stage.

[0106] Cooling coefficient From the thermal time constant of the voice coil and sampling rate Decide: In the formula, The thermal time constant of the loudspeaker voice coil, measured in seconds (s), is determined by the specific heat capacity and mass of the voice coil. The system's audio sampling rate (Hz);

[0107] Temperature rise coefficient Due to the thermal resistance of the voice coil Decide: . The thermal resistance of the voice coil to the environment, expressed in degrees Celsius per watt (°C). ), used to convert input electrical power into steady-state temperature rise value.

[0108] With estimated temperature When the value increases, the system implements a dual compensation strategy:

[0109] First, dynamically reduce static friction beyond the threshold to prevent thermal softening of the suspension system (stiffness coefficient). Displacement loss of control due to descent;

[0110] Second, dynamically adjust the quality factor Q of the bandpass filter. When When the preset thermal saturation point is exceeded, the system automatically reduces the Q value with a linear slope (e.g., from 2.0 to 1.2) to broaden the filter's passband range. This adjustment is to compensate for the DC resistance increase caused by the rise in voice coil temperature. This increases, which in turn causes the speaker's total electrical quality factor to increase. Changes and resonant frequencies The thermal drift phenomenon occurs, ensuring that the resonance point after drift remains within the monitoring range.

[0111] Furthermore, as the estimated temperature increases, the system automatically reduces the static friction threshold according to a preset thermal compensation curve. For example, when the estimated temperature reaches 60°C, the threshold is automatically lowered by a certain percentage. This dynamic adjustment mechanism makes the anti-displacement algorithm more sensitive under high-temperature conditions, proactively intervening to suppress displacement and significantly reducing the risk of uncontrolled displacement caused by thermal decay of the speaker's mechanical performance.

[0112] Specifically, with the estimated temperature As the static friction force increases, the system calculates the current threshold for exceeding the threshold in real time based on the following linear thermal decay equation. :

[0113] ;

[0114] in, The reference energy threshold after hardware calibration in step S11; Thermal protection start temperature (e.g.) ); Thermosensitive attenuation coefficient (unit: ), used to characterize the drift rate of suspension system compliance as a function of temperature.

[0115] If the real-time comparison results show that the cumulative displacement potential energy exceeds the static friction threshold (after thermal compensation and hardware calibration), the digital signal processor will immediately calculate the target dynamic attenuation coefficient (denoted as...). The purpose is to limit displacement potential energy within a safe range by suppressing the signal amplitude.

[0116] The system generates the following logic based on the current energy excess ratio. :

[0117] when At this time, the system determines that the current situation is within a safe range and no suppression is required:

[0118] ;

[0119] when At that time, the system calculates the amount of attenuation that needs to be applied:

[0120] ;

[0121] in, This represents the current discrete sampling time; This is the cumulative value of displacement potential energy (measured energy) calculated at the current moment. The threshold (safe energy limit) for static friction force allowed at the current moment.

[0122] The formula introduces the square root operation ( The reason is that, and All are in the dimension of energy, and their magnitude is proportional to the square of the signal amplitude. To obtain the linear attenuation coefficient acting on the voltage / digital signal amplitude, the energy ratio must be square-rooted.

[0123] To prevent the audible popping or "pumping effect" caused by abrupt gain changes, the system does not directly use the abrupt changes. Instead, it employs an asymmetric time control strategy to generate the smoothed final dynamic decay coefficient. .

[0124] Smoothing algorithm formula:

[0125] ;

[0126] Smoothing coefficient Dynamic switching logic:

[0127] The system compares the current target value with the actual value at the previous time step. This determines whether the system is in a startup or shutdown state:

[0128] when When signal suppression is needed immediately, a larger smoothing coefficient should be used. To achieve a rapid response (cut off kinetic energy):

[0129] ;

[0130] in, For rapid start-up time constant (typical value) ), This represents the system audio sampling rate.

[0131] when When the risk is eliminated and the gain is restored, a smaller smoothing coefficient is used. This causes the gain to recover slowly.

[0132] ;

[0133] Slow release time constant (typical value) ).

[0134] After calculating the smoothed coefficients, they are applied to the target resonant frequency band signal:

[0135] .

[0136] in, Refers to the sampling time Utilizing dynamic attenuation coefficient The suppressed target resonance frequency band signal amplitude obtained after weighting the original resonance signal; Refers to the current sampling time The original amplitude of the target resonant frequency band signal separated from the original digital audio signal by a bandpass filter.

[0137] In addition to suppressing the signal amplitude, the system also monitors whether the estimated voice coil temperature has reached thermal saturation, thereby dynamically adjusting the quality factor (Q value) of the bandpass filter to cover the resonant frequency that has changed due to thermal drift.

[0138] Judgment and calculation formulas:

[0139] The system compares and estimates the temperature in real time. With the preset temperature rise threshold :

[0140] like Keep the default quality factor .

[0141] like Calculate the new quality factor by decreasing the Q value with a linear slope. :

[0142] ;

[0143] in, The current voice coil temperature is estimated in real time by the thermal accumulation model; The thermal saturation onset temperature rise threshold is set to 50%~60% of the maximum withstand temperature of the voice coil (e.g., ); The default quality factor (high Q value, narrow band) determined during system initialization; Thermal attenuation slope coefficient (unit) ), used to define the rate at which the Q value decreases with increasing temperature; Used as a protective limit to ensure The response should be no less than 0.707 (Butterworth response) to prevent excessive flattening of the filter from causing unintended damage to non-resonant frequency bands.

[0144] Step S15: Use the dynamic attenuation coefficient to smooth and suppress the amplitude of the component in the target resonant frequency band of the digital audio signal, and simultaneously perform psychoacoustic low-frequency compensation to reconstruct auditory loudness.

[0145] Digital signal processors utilize generated dynamic attenuation coefficients to multiply suppress signal amplitude only within the target resonant frequency band. However, significant suppression of the resonant frequency band (usually the bass fundamental frequency) inevitably leads to a loss of perceived bass power. To compensate for this problem, this invention introduces psychoacoustic low-frequency compensation (Virtual Bass) technology.

[0146] Specifically, while suppressing the target resonant frequency band component, a parallel harmonic generator module extracts the time-domain fundamental frequency signal data of this component. Subsequently, a nonlinear mathematical model (such as Chebyshev polynomials) is used to generate the second and third harmonics of this fundamental frequency in real time. Since the frequencies of these higher harmonics (e.g., 164Hz, 246Hz) are far from the physical resonance sensitive zone of the enclosure (82Hz), even with high energy, they will not cause severe mechanical resonance or displacement of the enclosure. However, based on the missing fundamental phenomenon of the human auditory system, when the human ear receives these harmonic sequences, the cerebral cortex automatically reconstructs the missing fundamental pitch perception. The digital signal processor calculates the amount of harmonic gain that needs to be compensated based on the current dynamic attenuation coefficient: the deeper the suppression, the more harmonics are injected. These generated harmonic components are superimposed back into the main audio stream, allowing the user to still subjectively perceive deep and powerful bass, achieving an effective balance between physical displacement prevention and high-fidelity sound.

[0147] In this embodiment, the nonlinear harmonic generator uses a Chebyshev polynomial of the first kind to generate harmonics. Let the fundamental frequency signal extracted and normalized from the target resonant frequency band components be... The generated second harmonic and third harmonic They respectively satisfy:

[0148] ;

[0149] .

[0150] in, This represents the sampling point number at the current moment. It should be noted that, due to the even harmonic generation formula (... When the input signal is zero or non-zero, a DC bias component is generated. In order to prevent the DC output from damaging the speaker voice coil, before superimposing the generated harmonic components back into the main signal, step S15 also includes passing the harmonic signal through a digital high-pass filter (HPF) with a cutoff frequency of less than 20 Hz or performing a DC blocking algorithm to filter out the non-audio DC components generated by mathematical operations.

[0151] The system calculates the compensation gain based on the dynamic attenuation coefficient and adds the harmonic components to the original signal proportionally to fill the loudness gap after the fundamental frequency is suppressed.

[0152] The compensation gain With dynamic attenuation coefficient They are negatively correlated, and the calculation formula is:

[0153] ;

[0154] In the formula, An intensity factor (user-adjustable or preset, typically 0.5 to 1.0) is injected into the virtual bass to control the mixing ratio of harmonic components. Final output signal. for:

[0155] ;

[0156] in, This is a direct-through signal in the non-resonant frequency band. Its calculation method is as follows: (Using the method of subtracting the bandpass component from the full-frequency signal), or using a high-pass filter with a cutoff frequency matched to the bandpass filter to filter the original signal. The resulting audio information is processed to retain all audio information except for the target resonant frequency band.

[0157] It should also be noted that although this invention uses a Bluetooth speaker as an example, it is equally applicable to other portable audio playback devices such as Wi-Fi speakers and smart screen speakers.

[0158] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. An adaptive audio signal processing method for suppressing physical displacement of a Bluetooth speaker, characterized in that, The method is applied to a Bluetooth speaker system containing a digital signal processor, and the method includes acquiring the digital audio signal to be played in real time and monitoring the current system volume gain of the Bluetooth speaker. The feature is that when the system volume gain exceeds a preset displacement critical threshold, anti-displacement control logic is triggered, and the anti-displacement control logic includes the following steps: A target resonant frequency band component is separated from the digital audio signal using a bandpass filter. The target resonant frequency band component corresponds to the physical resonant frequency range of the Bluetooth speaker's enclosure. Set a time sliding window, and perform time-domain integration calculation on the energy of the target resonant frequency band component within the time sliding window to obtain the current cumulative displacement potential energy value; The accumulated value of displacement potential energy is compared with the preset static friction force breakthrough threshold in real time. If the accumulated displacement potential energy exceeds the static friction force threshold, a dynamic attenuation coefficient is generated. The amplitude of the component in the digital audio signal located in the target resonant frequency band is smoothed and suppressed using the dynamic attenuation coefficient, so that the suppressed accumulated displacement potential energy is maintained below the static friction force threshold.

2. The adaptive audio signal processing method for suppressing physical displacement of a Bluetooth speaker according to claim 1, characterized in that, The frequency range of the target resonant frequency band component is pre-calibrated and configured in the digital signal processor in the following manner: The total harmonic distortion curve generated when the Bluetooth speaker is scanned across the entire frequency band by an audio analyzer; Identify the frequency corresponding to the peak of the total harmonic distortion curve in the low-frequency region; Using the frequency as the center and combining it with a preset bandwidth factor, the upper and lower cutoff frequencies of the target resonant frequency band components are defined.

3. The adaptive audio signal processing method for suppressing physical displacement of a Bluetooth speaker according to claim 1, characterized in that, The step of using the dynamic attenuation coefficient to smoothly suppress the amplitude employs an asymmetric time control strategy: A rapid start-up time at the millisecond level is set so that when the accumulated displacement potential energy exceeds the limit, the dynamic attenuation coefficient can take effect immediately to cut off the accumulation of kinetic energy. By setting a slow release time on the order of hundreds of milliseconds, the gain of the target resonant frequency band component is slowly restored after the displacement risk is eliminated, thus avoiding a sudden, fluctuating breathing effect in the audio listening experience.

4. The adaptive audio signal processing method for suppressing physical displacement of a Bluetooth speaker according to claim 1, characterized in that, The method also includes a look-ahead buffering step: The digital audio signal is stored in a first-in-first-out buffer of a preset length; The time-domain integration calculation is performed on the audio data at the beginning of the buffer, while the smoothing suppression is performed on the audio data at the end of the buffer. The time difference of the buffer is used to predict the trend of the cumulative displacement potential energy in advance, and the dynamic attenuation coefficient is applied in advance before the actual physical displacement occurs.

5. The adaptive audio signal processing method for suppressing physical displacement of a Bluetooth speaker according to claim 1, characterized in that, The length of the time sliding window is configured to match the sustain period of common low-frequency instruments to distinguish between transient percussion signals and steady-state continuous low-frequency signals, ensuring that suppression is triggered only for audio components with sustained driving force.

6. The adaptive audio signal processing method for suppressing physical displacement of a Bluetooth speaker according to claim 1, characterized in that, The method also includes a psychoacoustic low-frequency compensation step: While suppressing the amplitude of the target resonant frequency band component, the time-domain fundamental frequency signal data of the target resonant frequency band component is extracted; The second and third harmonic components of the fundamental frequency are generated using a nonlinear harmonic generator. The compensation gain is calculated based on the magnitude of the dynamic attenuation coefficient, and the second and third harmonic components are superimposed on the digital audio signal to reconstruct the suppressed low-frequency loudness in hearing by utilizing the fundamental frequency loss effect of the human ear.

7. The adaptive audio signal processing method for suppressing physical displacement of a Bluetooth speaker according to claim 1, characterized in that, The center frequency of the target resonant frequency band component and the critical displacement threshold are calibrated based on hardware consistency parameters; The calibration process includes: retrieving the impedance and inductance characteristic values ​​of the speaker unit stored in the non-volatile memory of the Bluetooth speaker, wherein the impedance and inductance characteristic values ​​are statistical data obtained by measuring the speaker unit based on an LCR digital bridge; Calculate the deviation ratio between the impedance characteristic value and the standard design value; The default resonant center frequency is corrected for frequency deviation based on the deviation ratio, and the displacement critical threshold is compensated for sensitivity to adapt to the physical vibration characteristics of the current hardware batch.

8. The adaptive audio signal processing method for suppressing physical displacement of a Bluetooth speaker according to claim 1, characterized in that, The static friction force exceeding the threshold has adaptive thermal attenuation characteristics to compensate for thermal drift caused by prolonged playback: The digital signal processor calculates the high-power continuous playback duration of the Bluetooth speaker in real time. A thermal accumulation model of the speaker voice coil is established, and the voice coil temperature and compliance change of the suspension system are estimated based on the continuous playback duration. As the voice coil temperature increases, the static friction force exceeding the threshold is automatically reduced according to the preset thermal compensation curve to prevent displacement loss of control due to the decrease in the damping performance of the loudspeaker.

9. The adaptive audio signal processing method for suppressing physical displacement of a Bluetooth speaker according to claim 8, characterized in that, The thermal accumulation model is also used to dynamically adjust the quality factor Q value of the bandpass filter: When the continuous playback duration is short and the cumulative displacement potential energy is low, a high quality factor Q value is maintained to minimize the impact on the sound quality of non-resonant frequency bands. When the continuous playback duration exceeds the preset heat saturation time point, the quality factor Q value is automatically reduced to broaden the filter bandwidth, thereby covering the resonant frequency range that is detached due to temperature changes.

10. An adaptive audio signal processing system for suppressing physical displacement of a Bluetooth speaker, characterized in that, include: The signal acquisition module is used to acquire the digital audio signal to be played in real time and monitor the system volume gain; The logic trigger module is used to activate the subsequent processing unit when the system volume gain exceeds the displacement critical threshold. The resonance extraction module is used to separate the target resonance frequency band component through a bandpass filter; The potential energy analysis module is used to calculate the cumulative displacement potential energy of the target resonant frequency band component within a time sliding window; The threshold decision module is used to compare the accumulated displacement potential energy value with the static friction force exceeding the threshold. The suppression execution module is used to generate a dynamic attenuation coefficient and suppress the signal amplitude when the cumulative displacement potential energy exceeds the limit; The system also includes a non-volatile memory storing hardware consistency parameters generated based on LCR bridge test data, which are used to calibrate the filtering parameters of the resonance extraction module.

Images