A noise reduction method, device and medium for a Bluetooth headset noise reduction system

Abstract

This invention discloses a noise reduction method, device, and medium for a Bluetooth headset noise reduction system, relating to the field of acoustic noise reduction technology. The method includes: collecting speaker drive voltage, speaker coil current, and residual sound from the in-ear microphone during a noise reduction update cycle to determine the target noise reduction frequency band; determining the speaker electroacoustic load change value based on voltage and current, and prohibiting the writing of the current noise reduction adjustment amount when the residual sound increases and the load change reaches a boundary; when the residual sound increases but the load change does not reach a boundary, determining the speaker residual vibration time change value based on the coil current decay process, and generating an anti-noise adjustment scheme; and generating real-time control parameters based on the direction of residual sound change before and after adjustment. This invention can identify wearing disturbances or changes in ear canal load, reduce erroneous updates of anti-noise parameters, and effectively adjust through residual vibration time compensation output time and closed-loop verification, thereby improving noise reduction stability, matching accuracy, and overall reliability.

Description

Technical Field

[0001] This invention relates to the field of acoustic noise reduction technology, and in particular to a noise reduction method, device and medium for a Bluetooth headset noise reduction system. Background Technology

[0002] Bluetooth headset active noise cancellation systems typically use feedforward microphones, feedback microphones, or in-ear error microphones to collect ambient noise and residual sound in the ear canal. A digital signal processor then generates an anti-noise signal according to a preset update cycle to cancel out low-frequency noise entering the ear canal. In practical applications, the system generally adjusts the anti-noise amplitude, phase, or filtering parameters in real time based on the residual sound amplitude or frequency domain energy detected by the in-ear microphone, thereby maintaining the headset's noise cancellation effect in different noise environments.

[0003] The above methods typically do not take into account the electroacoustic load state reflected by the speaker drive voltage and coil current during parameter updates. Therefore, when the wearing posture of the headphones changes or the sealing state of the ear canal changes, the increase in residual sound may be misjudged as insufficient noise reduction. At the same time, conventional adjustments focus more on amplitude compensation and pay insufficient attention to the change in residual vibration time reflected by the attenuation of the speaker coil current, making it difficult to correct the effective time of anti-noise action in a timely manner. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a noise reduction method for a Bluetooth headset noise reduction system to solve the problems of existing technologies, such as difficulty in distinguishing changes in ear canal load from actual noise reduction and insufficient accuracy of anti-noise output compensation.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a noise reduction method for a Bluetooth headset noise reduction system, comprising: collecting the speaker driving voltage, speaker coil current, and in-ear microphone residual sound of the Bluetooth headset during a noise reduction update cycle; determining a target noise reduction frequency band based on the in-ear microphone residual sound; within the target noise reduction frequency band, determining the speaker electroacoustic load change value based on the speaker driving voltage and speaker coil current; when the in-ear microphone residual sound reaches the residual sound rise boundary and the speaker electroacoustic load change value reaches the load change boundary, prohibiting the writing of noise reduction adjustment amount formed by the current in-ear microphone residual sound; when the in-ear microphone residual sound reaches the residual sound rise boundary and the speaker electroacoustic load change value does not reach the load change boundary, determining the speaker residual vibration time change value based on the speaker coil current attenuation process, and generating an anti-noise adjustment scheme based on the speaker residual vibration time change value; after executing the anti-noise adjustment scheme, generating real-time anti-noise control parameters for the target noise reduction frequency band based on the direction of change of the in-ear microphone residual sound before and after adjustment.

[0007] As a preferred embodiment of the noise reduction method of the Bluetooth headset noise reduction system of the present invention, the step of determining the target noise reduction frequency band includes: reading the playback audio sequence after Bluetooth audio decoding, and determining the predictable playback sound component formed when the playback audio sequence is transmitted to the in-ear microphone according to the playback sound subtraction filter coefficient table; subtracting the predictable playback sound component from the sound collected by the in-ear microphone to obtain the residual sound of the in-ear microphone; performing a short-time Fourier transform on the residual sound of the in-ear microphone to obtain the frequency domain amplitude of the residual sound in the in-ear microphone within the current noise reduction update cycle; and determining the target noise reduction frequency band from multiple candidate noise reduction frequency bands based on the frequency domain amplitude of the residual sound in the in-ear microphone.

[0008] As a preferred embodiment of the noise reduction method of the Bluetooth headset noise reduction system of the present invention, the specific steps of determining the target noise reduction frequency band among multiple candidate noise reduction frequency bands are as follows: dividing the noise reduction frequency range to obtain candidate noise reduction frequency bands; calculating the residual sound rise of each candidate noise reduction frequency band in the current noise reduction update cycle; determining the candidate noise reduction frequency band with the largest residual sound rise as the candidate target noise reduction frequency band; when the residual sound rise of the candidate target noise reduction frequency band reaches the residual sound rise boundary, and the number of noise reduction update cycles that continuously reach the residual sound rise boundary is not less than the number of consecutive reaches, the candidate target noise reduction frequency band is determined as the target noise reduction frequency band; the residual sound rise is determined based on the root mean square amplitude of the residual sound in the in-ear microphone in the candidate noise reduction frequency band, the reference value of the residual sound amplitude under stable noise reduction state, and the residual sound fluctuation amplitude under stable noise reduction state.

[0009] In a preferred embodiment of the noise reduction method of the Bluetooth headset noise reduction system of the present invention, the step of determining the speaker electroacoustic load change value includes: performing a short-time Fourier transform on the speaker driving voltage sequence and speaker coil current sequence within the current noise reduction update cycle to obtain the current voltage frequency domain amplitude and the current frequency domain amplitude within the target noise reduction frequency band; determining the current speaker electroacoustic load value based on the ratio of the current voltage frequency domain amplitude and the current current frequency domain amplitude within the target noise reduction frequency band; determining the stable reference speaker electroacoustic load value based on the speaker driving voltage sequence and speaker coil current sequence in the stable reference data; and determining the speaker electroacoustic load change value based on the degree of change of the current speaker electroacoustic load value relative to the stable reference speaker electroacoustic load value.

[0010] As a preferred embodiment of the noise reduction method of the Bluetooth headset noise reduction system of the present invention, the specific steps of prohibiting the writing of the noise reduction adjustment amount formed by the residual sound of the current in-ear microphone are as follows: maintaining the currently effective anti-noise amplitude parameter and anti-noise output time parameter of the target noise reduction frequency band; not writing the noise reduction adjustment amount corresponding to the residual sound of the current in-ear microphone into the anti-noise real-time control parameter of the target noise reduction frequency band; determining the adjustment limit ratio according to the ratio of the load change boundary to the speaker electroacoustic load change value; and limiting the single-cycle change amount of the anti-noise amplitude parameter and the single-cycle change amount of the anti-noise output time parameter in subsequent noise reduction update cycles according to the adjustment limit ratio.

[0011] As a preferred embodiment of the noise reduction method of the Bluetooth headset noise reduction system of the present invention, the step of determining the speaker residual vibration time change value includes: when the residual sound of the in-ear microphone reaches the residual sound rise boundary and the speaker electroacoustic load change value does not reach the load change boundary, generating a load permission mark that allows correction; when the load permission mark allows correction, extracting the envelope of the speaker coil current sequence corresponding to the target noise reduction frequency band in the current noise reduction update cycle to obtain the current current envelope sequence; retrieving a falling segment in the current current envelope sequence that continuously decreases from a local peak and decreases to below a predetermined proportion of the local peak; determining the time length between the sampling point of the local peak and the sampling point at the end of the attenuation in the falling segment as the current residual vibration time; determining the reference residual vibration time based on the speaker coil current sequence in the stable reference data; and determining the speaker residual vibration time change value according to the degree of change of the current residual vibration time relative to the reference residual vibration time.

[0012] As a preferred embodiment of the noise reduction method of the Bluetooth headset noise reduction system of the present invention, the generation of the anti-noise adjustment scheme includes: when the change value of the speaker's residual vibration time is greater than the residual vibration extension boundary, keeping the currently effective anti-noise amplitude parameter of the target noise reduction frequency band unchanged, and advancing the currently effective anti-noise output time parameter by a fixed number of sampling points to generate candidate anti-noise output time parameters; when the change value of the speaker's residual vibration time is less than the residual vibration shortening boundary, keeping the currently effective anti-noise amplitude parameter of the target noise reduction frequency band unchanged, and delaying the currently effective anti-noise output time parameter by a fixed number of sampling points to generate candidate anti-noise output parameters. Timing parameters: When the change in the speaker's residual vibration time is not greater than the residual vibration extension boundary and not less than the residual vibration shortening boundary, the timing parameter of the currently effective anti-noise output in the target noise reduction frequency band remains unchanged, and the amplitude parameter of the currently effective anti-noise is enhanced by a limited amplitude to generate candidate anti-noise amplitude parameters; when there is no continuous sampling segment in the current current envelope sequence that satisfies the falling segment condition, the timing parameter of the currently effective anti-noise output in the target noise reduction frequency band remains unchanged, and the amplitude parameter of the currently effective anti-noise is enhanced by a low amplitude to generate candidate anti-noise amplitude parameters; the change in the low amplitude enhancement is less than the change in the limited amplitude enhancement.

[0013] As a preferred embodiment of the noise reduction method of the Bluetooth headset noise reduction system of the present invention, the generation of real-time anti-noise control parameters for the target noise reduction frequency band includes: after executing candidate anti-noise output time parameters or candidate anti-noise amplitude parameters, collecting the in-ear microphone residual sound sequence of the target noise reduction frequency band to form adjusted evaluation window data; determining the in-ear residual sound improvement value based on the pre-adjustment evaluation window data and the post-adjustment evaluation window data; when the in-ear residual sound improvement value reaches the improvement writing boundary, writing the candidate anti-noise output time parameters or candidate anti-noise amplitude parameters into the anti-noise real-time control parameters corresponding to the target noise reduction frequency band; when the in-ear residual sound improvement value does not reach the improvement writing boundary and is not lower than the deterioration boundary, not writing the candidate anti-noise parameters, and restoring the anti-noise amplitude parameters and anti-noise output time parameters of the target noise reduction frequency band in the stable reference period; when the in-ear residual sound improvement value is lower than the deterioration boundary, not writing the candidate anti-noise parameters, restoring the anti-noise amplitude parameters and anti-noise output time parameters of the target noise reduction frequency band in the stable reference period, and reducing the subsequent adjustment upper limit of the target noise reduction frequency band in the next noise reduction update period.

[0014] In a second aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the noise reduction method of the Bluetooth headset noise reduction system as described in the first aspect of the present invention.

[0015] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements any step of the noise reduction method of the Bluetooth headset noise reduction system as described in the first aspect of the present invention.

[0016] The beneficial effects of this invention are as follows: by identifying wearing disturbances or changes in ear canal load through changes in electroacoustic load, and preventing the erroneous writing of noise reduction adjustment values, the risk of erroneous updates of anti-noise parameters is reduced, and the stability of noise reduction is improved; by judging the change in residual resonance time through the attenuation process of speaker coil current, and adjusting the anti-noise output time, the matching accuracy between anti-noise and in-ear noise is improved; by comparing the changes in residual sound of the in-ear microphone before and after adjustment, only valid adjustment parameters are written, achieving closed-loop verification of the noise reduction effect and improving the reliability of noise reduction. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart of a noise reduction method for a Bluetooth headset noise reduction system.

[0019] Figure 2 A schematic diagram showing the determination of the target noise reduction frequency band.

[0020] Figure 3 This is a schematic diagram illustrating the relationship between changes in the speaker's electroacoustic load and write control determination.

[0021] Figure 4 This is a schematic diagram of speaker reverberation time adjustment and closed-loop writing.

[0022] Figure 5 A comparative data graph showing the changes in residual sound in the ear at the target noise reduction frequency band under perturbation conditions.

[0023] Figure 6 A comparative data chart showing the noise reduction effect of the target noise reduction frequency band after after-resonance time compensation. Detailed Implementation

[0024] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0025] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0026] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0027] Reference Figures 1-4 As an embodiment of the present invention, this embodiment provides a noise reduction method for a Bluetooth headset noise reduction system, comprising the following steps: S1. Collect the speaker drive voltage, speaker coil current and residual sound of the in-ear microphone of the Bluetooth headset during the noise reduction update cycle, and determine the target noise reduction frequency band based on the residual sound of the in-ear microphone.

[0028] After the Bluetooth headset enters the active noise cancellation mode, the digital signal processor collects the speaker drive voltage, speaker coil current and residual sound from the in-ear microphone according to the noise cancellation update cycle (e.g., 20ms) to form basic electroacoustic data.

[0029] The preferred audio sampling frequency is 48kHz.

[0030] It should be noted that the speaker driving voltage is collected from the output of the speaker driving circuit, the speaker coil current is collected from the precision sampling resistor connected in series in the speaker circuit, and the residual sound of the in-ear microphone is collected from the error microphone located inside the sound outlet.

[0031] To avoid mistaking Bluetooth audio playback for residual noise, the playback audio sequence after Bluetooth audio decoding is read, and the playback sound subtraction filter coefficient table generated by the factory calibration is called to calculate the predictable playback sound component formed by the playback audio being transmitted to the in-ear microphone through the speaker, sound outlet and ear canal entrance. Then, the predictable playback sound component is subtracted from the sound collected by the in-ear microphone to obtain the residual sound of the in-ear microphone.

[0032] Among them, the residual sound of the in-ear microphone is used to characterize the residual noise sound pressure that still exists in the ear canal after the anti-noise cancellation.

[0033] It should be noted that the playback sound subtraction filter coefficient table was obtained by performing frequency sweep calibration on the artificial ear using the same model of Bluetooth headset. During frequency sweep calibration, the Bluetooth headset speaker sequentially played a frequency sweep signal in the range of 80Hz to 800Hz, and the in-ear microphone synchronously collected the frequency sweep response signal. The digital signal processor formed the amplitude correction coefficient and delay correction coefficient for the corresponding frequency point according to the amplitude of the speaker playback signal, the amplitude of the in-ear microphone response, and the sampling delay between the two at each frequency point. The amplitude correction coefficient and delay correction coefficient of multiple frequency points were combined to form the playback sound subtraction filter coefficient table.

[0034] Furthermore, a short-time Fourier transform is performed on the residual sound from the in-ear microphone to obtain the frequency domain amplitude of the residual sound within the current noise reduction update cycle.

[0035] The frequency range from 80Hz to 800Hz was divided into nine non-overlapping candidate noise reduction bands, each with a bandwidth of 80Hz.

[0036] It should be noted that 80Hz to 800Hz is the main controllable range for low-frequency noise in active noise cancellation of in-ear Bluetooth headphones. Sound pressure changes below 80Hz are easily affected by slow changes in ear canal pressure, while sound signals above 800Hz are more sensitive to the earbud wearing angle and the position of the microphone in the ear, which can easily lead to phase instability. The bandwidth of each candidate noise cancellation frequency band is set to 80Hz in order to obtain stable frequency domain energy statistics within the noise cancellation update cycle of 10ms to 40ms. When the bandwidth is less than 80Hz, the residual sound rise is easily affected by single-point frequency fluctuations. When the bandwidth is greater than 80Hz, different low-frequency noise components are easily mixed, reducing the specificity of subsequent anti-noise output timing and anti-noise amplitude adjustments.

[0037] Candidate noise reduction frequency bands are numbered in ascending order of frequency. ,in Take 1 to 9, for the... There are 1 candidate noise reduction frequency bands, with a starting frequency of 1. Hz, the termination frequency is Hz.

[0038] Within each noise reduction update cycle, calculate the... The residual sound rise of each candidate noise reduction frequency band: ; ; in, Indicates the first Within the noise reduction update cycle, the first The residual sound rise of each candidate noise reduction frequency band Indicates the first Reference values ​​for residual sound amplitude of each candidate noise reduction frequency band under stable noise reduction state. Indicates the first The residual acoustic fluctuation amplitude of each candidate noise reduction frequency band under stable noise reduction conditions. Indicates the first The starting frequency of each candidate noise reduction band Indicates the first The termination frequency of each candidate noise reduction band. Indicates the first Within each noise reduction update cycle, the residual sound from the in-ear microphone at the frequency The amplitude in the frequency domain at that point, Indicates the first Within the noise reduction update cycle, the first The square root of the square mean of the residual sound amplitude at each frequency point within the candidate noise reduction frequency band.

[0039] It should be noted that, and Data was obtained through stable active noise cancellation tests performed on the same model of Bluetooth headphones in an artificial ear. The tests were conducted under three sound environments: pink noise, low-frequency subway noise, and office air conditioning noise. The digital signal processor statistically analyzed the data during the time periods when the residual sound at the in-ear microphone remained continuously below the residual sound rise threshold. The residual root mean square amplitude of each candidate noise reduction frequency band was used, and the median of the residual root mean square amplitude over multiple stable time periods was taken as the median. The interquartile range of the residual root mean square amplitude over multiple stable time periods is used as... Using the median and interquartile range can reduce the impact of accidental touches, instantaneous impact noises, and short-term low-frequency playback on stable reference data.

[0040] It should be noted that the residual sound rise boundary was obtained by performing stable noise reduction test and disturbed noise reduction test in an artificial ear using the same model of Bluetooth headphones. The median value between the upper quartile of the residual sound rise of each candidate noise reduction frequency band in the stable noise reduction test and the lower quartile of the residual sound rise of each candidate noise reduction frequency band in the disturbed noise reduction test was taken as the residual sound rise boundary, with a value range of 2 to 4. The disturbed noise reduction test was conducted by slightly changing the in-ear position of the Bluetooth headphones or increasing the intensity of low-frequency external noise in the target frequency band.

[0041] Furthermore, the residual sound rise of the nine candidate noise reduction frequency bands is compared, and the candidate noise reduction frequency band with the largest residual sound rise is selected as the candidate target noise reduction frequency band.

[0042] When the residual sound rise of the candidate target noise reduction frequency band reaches the residual sound rise boundary, and the number of noise reduction update cycles that continuously reach the residual sound rise boundary is not less than the number of consecutive times, the candidate target noise reduction frequency band is determined as the target noise reduction frequency band.

[0043] It should be noted that the optimal range for the number of consecutive occurrences is 2 to 4 noise reduction update cycles. The increase in residual sound within a single noise reduction update cycle is easily affected by instantaneous touch, low-frequency transients during playback, or sampling fluctuations, while more than 2 consecutive noise reduction update cycles indicate that the increase in residual sound is persistent.

[0044] Read the current effective anti-noise amplitude parameter and the current effective anti-noise output time parameter corresponding to the target noise reduction frequency band.

[0045] It should be noted that the initial values ​​of the anti-noise amplitude parameter and the anti-noise output timing parameter are obtained through the factory calibration of the same model of Bluetooth headset. During factory calibration, the Bluetooth headset is installed in the artificial ear, and standard low-frequency noise in the range of 80Hz to 800Hz is played. The propagation time of external noise entering the ear canal, the propagation time of speaker anti-noise reaching the microphone in the ear, and the fixed calculation delay of the digital signal processor are recorded. The gain coefficient that minimizes the residual sound in the microphone in the ear is determined as the initial value of the anti-noise amplitude parameter, and the sampling point offset that maximizes the cancellation of anti-noise and ear canal noise at the microphone in the ear is determined as the anti-noise output timing parameter. During real-time operation, the currently effective anti-noise amplitude parameter and anti-noise output timing parameter are read according to the target noise reduction frequency band.

[0046] Before determining the target noise reduction frequency band, the stable current of the speaker coil is calculated simultaneously to determine the stabilization period, expressed as follows: ; in, Indicates the first The stability of the speaker coil current within each noise reduction update cycle Indicates the first The maximum value of the speaker coil current envelope within each noise reduction update cycle. Indicates the first Minimum value of speaker coil current envelope within each noise reduction update cycle Indicates the first The average value of the speaker coil current envelope within each noise reduction update cycle.

[0047] When the residual sound rise of all nine candidate noise reduction frequency bands is below the residual sound rise boundary, and the current stability of the speaker coil is below the current stability boundary, the current noise reduction update cycle is determined as the stable cycle.

[0048] If the Bluetooth headset has not yet established a stable cycle, the initial stable reference data will be used as the stable reference data.

[0049] It should be noted that the initial stable reference data was obtained through the factory calibration stage of the same model of Bluetooth headset. The Bluetooth headset was installed in the artificial ear, and the Bluetooth headset was put into a stable active noise cancellation working state. Standard low-frequency noise was played in the range of 80Hz to 800Hz according to nine candidate noise cancellation frequency bands. The digital signal processor simultaneously collected the speaker drive voltage sequence, speaker coil current sequence and in-ear microphone residual sound sequence, and read the anti-noise amplitude parameter and anti-noise output time parameter, and saved it as the initial stable reference data.

[0050] It should be noted that the current stability boundary was obtained by continuously performing stable active noise cancellation tests on the same model of Bluetooth headset in the artificial ear. The upper quartile value of the stable value of the speaker coil current in the stable active noise cancellation test was taken, and the preferred value range was 0.04 to 0.08.

[0051] Whenever a stable period is formed, the speaker driving voltage sequence, speaker coil current sequence, anti-noise amplitude parameter, anti-noise output time parameter, and in-ear microphone residual sound sequence within the stable period are saved. The stable period closest to the current noise reduction update period is taken as the previous stable period, and the speaker driving voltage sequence, speaker coil current sequence, anti-noise amplitude parameter, anti-noise output time parameter, and in-ear microphone residual sound sequence of the previous stable period are taken as stable reference data. Before a real-time stable period is formed, the stable reference period is the calibration period corresponding to the initial stable reference data.

[0052] Once the target noise reduction frequency band is determined, the digital signal processor will cache the residual sound sequence of the in-ear microphone in the target noise reduction frequency band before load judgment as the evaluation window data before adjustment.

[0053] It should be noted that the length of the evaluation window before adjustment is one to three noise reduction update cycles, preferably two noise reduction update cycles, because one noise reduction update cycle can cover the most recent residual sound state of the in-ear microphone before the candidate anti-noise parameter is executed. More than three noise reduction update cycles will cause the evaluation data to cross a new noise change stage, resulting in the evaluation window before adjustment not corresponding to the evaluation window after the candidate anti-noise parameter is executed.

[0054] S2. Within the target noise reduction frequency band, determine the speaker electroacoustic load change value based on the speaker drive voltage and speaker coil current. When the residual sound of the in-ear microphone reaches the residual sound rise boundary and the speaker electroacoustic load change value reaches the load change boundary, it is forbidden to write the noise reduction adjustment amount formed by the current residual sound of the in-ear microphone.

[0055] Furthermore, a short-time Fourier transform is performed on the loudspeaker driving voltage sequence and loudspeaker coil current sequence within the current noise reduction update cycle to obtain the current voltage frequency domain amplitude and the current frequency domain amplitude within the target noise reduction frequency band. The current loudspeaker electroacoustic load value is then calculated according to the start and end frequencies of the target noise reduction frequency band, expressed as: ; in, Indicates the first The current speaker electroacoustic load value of the target noise reduction frequency band within each noise reduction update cycle. Indicates the starting frequency of the target noise reduction frequency band. Indicates the termination frequency of the target noise reduction frequency band. Indicates the first Within each noise reduction update cycle, the speaker drive voltage at frequency The amplitude in the frequency domain at that point, Indicates the first Within each noise reduction update cycle, the speaker coil current is at a frequency The amplitude in the frequency domain at that point, Represents a frequency variable.

[0056] Furthermore, the stable reference loudspeaker electroacoustic load value is calculated based on the stable reference data.

[0057] Calculate the speaker electroacoustic load variation based on the current speaker electroacoustic load value and the stable reference speaker electroacoustic load value: ; in, Indicates the first The change in loudspeaker electroacoustic load in the target noise reduction frequency band within each noise reduction update cycle. This indicates the stable reference loudspeaker electroacoustic load value.

[0058] When the residual sound of the in-ear microphone in the target noise reduction frequency band reaches the residual sound rise boundary, and the change value of the speaker electroacoustic load reaches the load change boundary, it is determined that the current acoustic load of the ear canal has changed relative to the stable reference state. At this time, the rise of the residual sound of the in-ear microphone cannot directly indicate that the anti-noise amplitude is insufficient. Therefore, it is prohibited to write the noise reduction adjustment amount corresponding to the current residual sound of the in-ear microphone into the anti-noise real-time control parameters, and the current effective anti-noise amplitude parameter and the current effective anti-noise output time parameter corresponding to the target noise reduction frequency band are maintained.

[0059] When the residual sound from the in-ear microphone within the target noise reduction frequency band reaches the residual sound rise boundary, and the change in the speaker electroacoustic load reaches the load change boundary, it is determined that the current acoustic conditions of the ear canal have changed relative to the stable reference data. The current residual sound from the in-ear microphone cannot be directly used as a basis for increasing the anti-noise amplitude or changing the anti-noise output time. The currently effective anti-noise amplitude parameters and anti-noise output time parameters for the target noise reduction frequency band are maintained. The noise reduction adjustment amount corresponding to the current residual sound from the in-ear microphone is not written into the real-time anti-noise control parameters. The load change boundary is divided by the change in the speaker electroacoustic load to obtain the adjustment limit ratio. The target noise reduction frequency band is then... The permissible single-cycle change of the anti-noise amplitude parameter under steady-state conditions is multiplied by the adjustment limit ratio to obtain the restricted change of the anti-noise amplitude parameter in subsequent noise reduction update cycles. The permissible single-cycle change of the anti-noise output time parameter under steady-state conditions in the target noise reduction frequency band is multiplied by the adjustment limit ratio to obtain the restricted change of the anti-noise output time parameter in subsequent noise reduction update cycles. When generating candidate anti-noise parameters in subsequent noise reduction update cycles, the single-cycle change of the anti-noise amplitude parameter is ensured not to exceed the restricted change of the anti-noise amplitude parameter, and the single-cycle change of the anti-noise output time parameter is ensured not to exceed the restricted change of the anti-noise output time parameter.

[0060] It should be noted that the allowable single-cycle variation of the anti-noise amplitude parameter and the single-cycle variation of the anti-noise output time parameter in the target noise reduction frequency band under steady-state conditions were obtained through stable active noise cancellation tests performed on the same model of Bluetooth headset in an artificial ear. Specifically, within a stable time period when the residual sound of the in-ear microphone is below the residual sound rise boundary and the stable current of the speaker coil is below the current stability boundary, the single-cycle variation of the anti-noise amplitude parameter is gradually increased, and the maximum variation that will not cause the residual sound of the in-ear microphone to reach the residual sound rise boundary is taken as the allowable single-cycle variation of the anti-noise amplitude parameter under steady-state conditions; the single-cycle variation of the anti-noise output time parameter is gradually increased, and the maximum sampling point variation that will not cause the residual sound of the in-ear microphone to reach the residual sound rise boundary is taken as the allowable single-cycle variation of the anti-noise output time parameter under steady-state conditions.

[0061] When the residual sound of the in-ear microphone in the target noise reduction frequency band reaches the residual sound rise boundary, and the change value of the speaker electroacoustic load does not reach the load change boundary, it is determined that the current acoustic load of the ear canal is still within the range that can be used for noise reduction correction, and a load permission flag is generated and marked as allowed to correct; otherwise, the load permission flag is marked as prohibited to correct.

[0062] When the residual sound from the in-ear microphone within the target noise reduction frequency band has not reached the residual sound rise boundary, but the change value of the speaker electroacoustic load has reached the load change boundary, the digital signal processor (DSP) determines that the acoustic conditions of the ear canal have changed relative to the stable reference data, but the residual sound from the in-ear microphone has not yet formed an increased state that requires writing the anti-noise adjustment amount. The DSP maintains the currently effective anti-noise amplitude parameter and anti-noise output time parameter of the target noise reduction frequency band and does not generate candidate anti-noise parameters for enhancing the anti-noise amplitude. At the same time, the load change boundary is divided by the change value of the speaker electroacoustic load to obtain the adjustment limit ratio, and in the next noise reduction update cycle, the single-cycle allowable change of the anti-noise amplitude parameter and the single-cycle allowable change of the anti-noise output time parameter are limited according to the adjustment limit ratio to obtain the limited amplitude allowable change and the limited time allowable change, which serve as the upper limit for subsequent adjustments.

[0063] It should be noted that the load variation boundary was obtained by conducting stable wearing tests and disturbed wearing tests on the same model of Bluetooth headset in the artificial ear. In the stable wearing test, the Bluetooth headset maintained a standard in-ear posture and performed active noise cancellation. In the disturbed wearing test, the Bluetooth headset was made to have a slight axial displacement or partially block the pressure relief hole in the artificial ear, while maintaining active noise cancellation. The speaker electroacoustic load variation value of each candidate noise cancellation frequency band in the two types of tests was calculated respectively. The median value between the upper quartile of the speaker electroacoustic load variation value in the stable wearing test and the lower quartile of the speaker electroacoustic load variation value in the disturbed wearing test was taken as the load variation boundary. The preferred value range of the load variation boundary is 0.03 to 0.10.

[0064] S3. When the residual sound of the in-ear microphone reaches the residual sound rise boundary and the change value of the speaker electroacoustic load does not reach the load change boundary, determine the speaker residual vibration time change value according to the speaker coil current decay process, and generate an anti-noise adjustment scheme based on the speaker residual vibration time change value.

[0065] When the load permission is marked as disabled for correction, the current effective anti-noise amplitude parameters and anti-noise output time parameters of the target noise reduction frequency band are maintained, the speaker residual time change value is not calculated, no candidate anti-noise adjustment methods are generated, and the next noise reduction update cycle is returned.

[0066] Furthermore, when the load permission is marked as allowing correction, the envelope of the speaker coil current sequence corresponding to the target noise reduction frequency band in the current noise reduction update cycle is extracted to obtain the current current envelope sequence.

[0067] The current envelope sequence is obtained by taking the absolute value of the speaker coil current sequence and then performing low-pass smoothing.

[0068] Furthermore, the falling segment is retrieved within the current current envelope sequence.

[0069] The falling segment is a continuous sampling segment in which the current current envelope sequence continuously decreases from a local peak and decreases to below a predetermined proportion of the local peak.

[0070] It should be noted that the predetermined ratio is obtained by performing a stable active noise cancellation test on the same model of Bluetooth headphones in the artificial ear. The peak ratio that can stably characterize the end of coil current decay in the stable active noise cancellation test is taken as the predetermined ratio. The preferred value range is one-fifteenth to one-fifth of the local peak value. In specific implementation, one-tenth of the local peak value can be taken as an example value.

[0071] The time length from the local peak sampling point to the attenuation end sampling point within the falling segment is determined as the current residual vibration time of the current noise reduction update cycle.

[0072] Perform the same envelope extraction and falling segment retrieval on the loudspeaker coil current sequence in the stable reference data as on the current current envelope sequence, and determine the time length of the corresponding falling segment in the stable reference data as the reference residual oscillation time.

[0073] If the stable reference data is the initial stable reference data, then the reference residual oscillation time is calculated from the speaker coil current sequence saved during the factory calibration stage; if the stable reference data is the data saved in the previous stable cycle, then the reference residual oscillation time is calculated from the speaker coil current sequence in the previous stable cycle.

[0074] The expression for the change in the speaker's residual vibration time is as follows: ; in, Indicates the first The change in speaker reverberation time in the target noise reduction frequency band within each noise reduction update cycle. This indicates the current reverberation time of the target noise reduction frequency band within the current noise reduction update cycle. This represents the reference residual vibration time calculated based on stable reference data.

[0075] When the change in the speaker's residual vibration time is greater than the residual vibration extension boundary, it is determined that the effective time of anti-noise action in the target noise reduction frequency band is lagging behind the stable reference data, and the first candidate anti-noise adjustment method is generated.

[0076] The first candidate anti-noise adjustment method is to keep the anti-noise amplitude parameter currently effective in the target noise reduction frequency band unchanged, and advance the output time parameter of the currently effective anti-noise in the target noise reduction frequency band by one sampling point to three sampling points to obtain the candidate anti-noise output time parameter; the number of advanced sampling points is limited by the subsequent adjustment upper limit. When the allowable change of the restricted time corresponding to the subsequent adjustment upper limit is less than three sampling points, the number of advanced sampling points shall not exceed the allowable change of the restricted time corresponding to the subsequent adjustment upper limit.

[0077] It should be noted that the first candidate anti-noise adjustment method is chosen because the extended residual vibration time of the speaker indicates that the anti-noise energy decays more slowly, and the time position at which the anti-noise effectively cancels out in the ear shifts backward. By advancing the anti-noise output time, the time shift can be compensated, thus avoiding low-frequency booming caused by simply increasing the anti-noise amplitude.

[0078] When the change in the speaker's residual vibration time is less than the residual vibration shortening boundary, it is determined that the effective time of anti-noise action in the target noise reduction frequency band occurs earlier than the stable reference data, and a second candidate anti-noise adjustment method is generated.

[0079] The second candidate anti-noise adjustment method is to keep the currently effective anti-noise amplitude parameter of the target noise reduction frequency band unchanged, and delay the currently effective anti-noise output time parameter of the target noise reduction frequency band by one to three sampling points to obtain the candidate anti-noise output time parameter; the number of delayed sampling points is limited by the subsequent adjustment upper limit; when the allowable change of the restricted time corresponding to the subsequent adjustment upper limit is less than three sampling points, the number of delayed sampling points shall not exceed the allowable change of the restricted time corresponding to the subsequent adjustment upper limit.

[0080] It should be noted that the second candidate anti-noise adjustment method is chosen because the shortening of the speaker's residual vibration time indicates that the anti-noise energy decays faster, and the effective anti-noise cancellation time position is advanced. By delaying the anti-noise output time, the anti-noise can be realigned with the noise in the ear, reducing the increase in residual sound caused by the phase advance.

[0081] When the change in the speaker's residual vibration time is not greater than the residual vibration extension boundary and not less than the residual vibration shortening boundary, it is determined that there is no directional mismatch in the anti-noise output time within the target noise reduction frequency band, and a third candidate anti-noise adjustment method is generated.

[0082] The third candidate anti-noise adjustment method is to keep the output time parameter of the currently effective anti-noise in the target noise reduction frequency band unchanged, and to perform a limited amplitude enhancement on the amplitude parameter of the currently effective anti-noise in the target noise reduction frequency band to obtain candidate anti-noise amplitude parameters. The amount of change of the limited amplitude enhancement is limited by the subsequent adjustment upper limit. When the limited amplitude change corresponding to the subsequent adjustment upper limit is less than the amplitude change in the normal stable state, the amount of change of the limited amplitude enhancement shall not exceed the limited amplitude change corresponding to the subsequent adjustment upper limit.

[0083] It should be noted that the third candidate noise reduction method is chosen because the speaker's residual time does not show an advance or lag direction, and the increase in the residual sound of the in-ear microphone is more likely to come from insufficient noise reduction intensity in the target noise reduction frequency band. Therefore, keeping the output time unchanged and only increasing the amplitude by a limited amount can avoid unnecessary time correction.

[0084] When there is no continuous sampling segment in the current current envelope sequence that can satisfy the falling segment condition, the speaker residual time change value is not calculated, and a fourth candidate anti-noise adjustment method is generated.

[0085] The fourth candidate anti-noise adjustment method is to keep the output time parameter of the currently effective anti-noise in the target noise reduction frequency band unchanged, and to perform a low-amplitude enhancement on the amplitude parameter of the currently effective anti-noise in the target noise reduction frequency band to obtain the candidate anti-noise amplitude parameter; the change in the low-amplitude enhancement is less than the limited amplitude enhancement change in the third candidate anti-noise adjustment method, and shall not exceed the limited amplitude allowable change corresponding to the upper limit of subsequent adjustments.

[0086] It should be noted that the fourth candidate noise reduction adjustment method is chosen because when there is no effective drop segment, it is impossible to reliably determine the direction of the advance or lag of the effective noise reduction action time. If the output time is adjusted further, it is easy to cause erroneous compensation. Using low-amplitude enhancement can perform conservative noise reduction correction without introducing time misadjustment.

[0087] It should be noted that the residual vibration extension boundary and residual vibration shortening boundary were obtained by performing stable active noise cancellation tests in an artificial ear using the same model of Bluetooth headset. Specifically, under the condition that the Bluetooth headset maintains the standard in-ear posture and continuously performs active noise cancellation, the speaker coil current sequence within multiple stable cycles was collected, and the residual vibration time change value distribution under stable conditions was calculated. The upper quartile value of the residual vibration time change value distribution was taken as the residual vibration extension boundary, and the lower quartile value of the residual vibration time change value distribution was taken as the residual vibration shortening boundary. The preferred value range for the residual vibration extension boundary is 0.05 to 0.15, and the preferred value range for the residual vibration shortening boundary is -0.15 to -0.05.

[0088] It should be noted that the range of one to three sampling points for adjusting the anti-noise output timing parameter in a single operation is because the target noise reduction frequency band is between 80Hz and 800Hz. The single adjustment of the anti-noise output timing needs to be less than the time length that can cause a significant phase change within the low-frequency cycle. When the audio sampling frequency is 48kHz, one to three sampling points correspond to approximately 20.8 microseconds to 62.5 microseconds, which can form a detectable timing correction and is less likely to cause a perceptible anti-noise change in the user's perception. If the audio sampling frequency changes, the number of sampling points is converted according to the audio sampling period to keep the single adjustment time within the same order of magnitude.

[0089] It should be noted that the limited amplitude enhancement and low amplitude enhancement were obtained through stable active noise cancellation tests performed on the same model of Bluetooth headset in the artificial ear. Specifically, during the stable period when the residual sound of the in-ear microphone is below the residual sound rise boundary and the stable current of the speaker coil is below the current stability boundary, the single-cycle change of the anti-noise amplitude parameter is gradually increased. The maximum change that will not cause the residual sound of the in-ear microphone to reach the residual sound rise boundary is taken as the allowable single-cycle change of the anti-noise amplitude parameter under normal stable conditions. The limited amplitude enhancement in the third candidate anti-noise adjustment method is capped at the allowable single-cycle change of the anti-noise amplitude parameter under normal stable conditions; the change in the low amplitude enhancement in the fourth candidate anti-noise adjustment method is half of the limited amplitude enhancement change in the third candidate anti-noise adjustment method.

[0090] It should be noted that the change in low amplitude enhancement is half of the change in limited amplitude enhancement because the fourth candidate anti-noise adjustment method occurs when the speaker residual time cannot be reliably calculated. Halving the enhancement can preserve the response to the increase in residual sound of the in-ear microphone, while reducing the risk of overcompensation of anti-noise due to unknown time direction.

[0091] S4. After implementing the anti-noise adjustment scheme, generate real-time anti-noise control parameters for the target noise reduction frequency band based on the direction of change of residual sound in the in-ear microphone before and after adjustment.

[0092] The candidate anti-noise adjustment methods to be implemented will be used as anti-noise adjustment schemes.

[0093] Furthermore, after implementing the anti-noise adjustment scheme, the residual sound sequence of the in-ear microphone in the target noise reduction frequency band is collected to form the adjusted evaluation window data.

[0094] The adjusted evaluation window length is the same as the original evaluation window length.

[0095] The improvement value of residual sound in the ear is calculated based on the evaluation window data before and after adjustment: ; in, Indicates the first The improvement value of residual sound in the ear in the target noise reduction frequency band within each noise reduction update cycle. Indicates the length of the evaluation window. This represents the in-ear microphone residual sound sequence within the target noise reduction frequency band in the evaluation window before adjustment. This represents the in-ear microphone residual sound sequence within the target noise reduction frequency band in the adjusted evaluation window. This indicates the in-ear microphone residual sound sequence in the evaluation window before adjustment at time [time value missing]. The absolute sound pressure level, This indicates the adjusted evaluation window's in-ear microphone residual sound sequence at time [time missing]. The absolute sound pressure level.

[0096] The improvement value of residual sound in the ear is used to characterize whether the residual sound of the in-ear microphone in the target noise reduction frequency band is reduced after the candidate anti-noise parameters are implemented. When the improvement value of residual sound in the ear is positive, it means that the residual sound of the in-ear microphone in the evaluation window after adjustment is lower than that in the evaluation window before adjustment. When the improvement value of residual sound in the ear is close to zero, it means that the candidate anti-noise parameters have not formed an effective improvement. When the improvement value of residual sound in the ear is negative, it means that the candidate anti-noise parameters have increased the residual sound of the in-ear microphone.

[0097] When the improvement value of residual sound in the ear reaches the improvement writing boundary, the candidate anti-noise parameter is written into the anti-noise real-time control parameter corresponding to the target noise reduction frequency band; if the candidate anti-noise parameter is the candidate anti-noise output time parameter, the candidate anti-noise output time parameter is written and the anti-noise amplitude parameter remains unchanged; if the candidate anti-noise parameter is the candidate anti-noise amplitude parameter, the candidate anti-noise amplitude parameter is written and the anti-noise output time parameter remains unchanged.

[0098] Among them, the candidate anti-noise parameters include the anti-noise output time parameter and the anti-noise amplitude parameter.

[0099] When the improvement value of residual sound in the ear does not reach the improvement writing boundary and is not lower than the deterioration boundary, the candidate anti-noise parameters are not written, and the anti-noise amplitude parameters and anti-noise output time parameters of the target noise reduction frequency band in the stable reference period are restored.

[0100] When the improvement value of residual sound in the ear is lower than the deterioration boundary, the candidate anti-noise parameters are not written, the anti-noise amplitude parameters and anti-noise output time parameters of the target noise reduction frequency band in the stable reference period are restored, and the subsequent adjustment upper limit of the target noise reduction frequency band in the next noise reduction update period is reduced.

[0101] The method for reducing the upper limit of subsequent adjustments is to calculate the allowable change in the current restricted amplitude and the allowable change in the current restricted time according to the step reduction ratio, and obtain the allowable change in the restricted amplitude and the allowable change in the restricted time used in the next noise reduction update cycle.

[0102] It should be noted that the step-down ratio is obtained by performing a deterioration adjustment and recovery test on the same model of Bluetooth headset in the artificial ear. The ratio is taken as the ratio that will not cause the residual sound of the microphone in the ear to reach the deterioration boundary again and can be restored to a stable cycle. The preferred value range is 0.4 to 0.7.

[0103] When the improvement value of residual sound in the ear is lower than the deterioration boundary in two consecutive noise reduction update cycles for the same target noise reduction frequency band, the generation of similar candidate anti-noise parameters is stopped during the pause time.

[0104] It should be noted that the pause time was obtained by performing a continuous deterioration adjustment test on the same model of Bluetooth headset in the artificial ear, and the median of the time required for the residual sound of the in-ear microphone to recover from the deteriorated state to below the residual sound rise boundary was taken, preferably 200ms to 500ms.

[0105] It should be noted that the improvement write boundary and the degradation boundary were obtained by performing effective adjustment tests and ineffective adjustment tests in an artificial ear using the same model of Bluetooth headset. In the effective adjustment test, the residual sound of the in-ear microphone decreased after the candidate anti-noise parameter was executed. In the ineffective adjustment test, the residual sound of the in-ear microphone did not decrease or increased after the candidate anti-noise parameter was executed. The lower quartile value of the improvement value of the residual sound in the in-ear microphone in the effective adjustment test was used as the improvement write boundary, and the lower quartile value of the improvement value of the residual sound in the in-ear microphone in the ineffective adjustment test was used as the degradation boundary. The value range of the improvement write boundary was 0.05 to 0.12, and the value range of the degradation boundary was -0.08 to -0.03.

[0106] After the candidate noise reduction parameters are updated to the real-time noise reduction control parameters corresponding to the target noise reduction frequency band, the residual sound rise of the nine candidate noise reduction frequency bands is recalculated in the next noise reduction update cycle, and the speaker coil current stability is calculated simultaneously. When the residual sound rise of the nine candidate noise reduction frequency bands is lower than the residual sound rise boundary, and the speaker coil current stability is lower than the current stability boundary, the next noise reduction update cycle is determined as the new stable cycle. The speaker driving voltage sequence, speaker coil current sequence, in-ear microphone residual sound sequence, noise reduction amplitude parameters, and noise reduction output time parameters in the new stable cycle are saved, and the data saved in the new stable cycle is updated as stable reference data.

[0107] like Figure 5This paper demonstrates the variation of residual sound level in the ear at the target noise reduction frequency band with the noise reduction update cycle under wearing disturbance conditions. As can be seen from the curves, the traditional direct residual sound update method shows a significant increase and large fluctuation in residual sound level after the introduction of wearing disturbance, indicating that it is prone to misinterpreting changes in wearing posture or ear canal load as insufficient noise reduction. Although the load-limiting method can suppress the sudden increase in residual sound, the recovery speed and stability level are still limited. The method of this invention identifies changes in ear canal load by changing the value of speaker electroacoustic load, and prevents the erroneous writing of noise reduction adjustment amount when the boundary conditions are met. This ensures that the residual sound level only increases briefly and then quickly falls back to a low level, demonstrating the technical effect of reducing the risk of erroneous update of anti-noise parameters and improving the stability of noise reduction.

[0108] like Figure 6 The changes in the residual sound level in the ear after residual vibration time compensation are shown. As can be seen from the curves, without compensation, the residual sound level decreases slowly and remains at a high level for a long time. Although amplitude enhancement alone can reduce some residual sound, plateauing and fluctuations still occur in the later stages, indicating that simple amplitude adjustment is insufficient to fully solve the problem of the offset of the anti-noise action time. The method of this invention determines the change value of the speaker residual vibration time based on the attenuation process of the speaker coil current, and adjusts the anti-noise output time accordingly. Then, it performs closed-loop writing by measuring the changes in residual sound in the ear microphone before and after adjustment, so that the residual sound level decreases rapidly and stabilizes at a low level, demonstrating the technical effect of improving phase matching accuracy and noise reduction reliability.

[0109] This embodiment also provides a computer device applicable to the noise reduction method of a Bluetooth headset noise reduction system, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the noise reduction method of the Bluetooth headset noise reduction system as proposed in the above embodiment.

[0110] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0111] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the noise reduction method for implementing a Bluetooth headset noise reduction system as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0112] In summary, this invention identifies wearing disturbances or changes in ear canal load by changing electroacoustic load values ​​and prevents the erroneous writing of noise reduction adjustment values, thereby reducing the risk of erroneous updates to anti-noise parameters and improving noise reduction stability; it judges the change in residual resonance time by the attenuation process of the speaker coil current and adjusts the anti-noise output time, thereby improving the matching accuracy between anti-noise and in-ear noise; by comparing the changes in residual sound of the in-ear microphone before and after adjustment and only writing valid adjustment parameters, it achieves closed-loop verification of noise reduction effect and improves noise reduction reliability.

[0113] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A noise reduction method for a Bluetooth headset noise reduction system, characterized in that, include: Collect the speaker drive voltage, speaker coil current and residual sound of the in-ear microphone of the Bluetooth headset during the noise reduction update cycle, and determine the target noise reduction frequency band based on the residual sound of the in-ear microphone. Within the target noise reduction frequency band, the speaker electroacoustic load change value is determined based on the speaker drive voltage and speaker coil current. When the residual sound of the in-ear microphone reaches the residual sound rise boundary and the speaker electroacoustic load change value reaches the load change boundary, the noise reduction adjustment amount formed by the current in-ear microphone residual sound is prohibited from being written. When the residual sound of the in-ear microphone reaches the residual sound rise boundary and the change value of the speaker electroacoustic load does not reach the load change boundary, the change value of the speaker residual vibration time is determined according to the speaker coil current decay process, and an anti-noise adjustment scheme is generated based on the change value of the speaker residual vibration time. After implementing the anti-noise adjustment scheme, real-time anti-noise control parameters for the target noise reduction frequency band are generated based on the direction of change in the residual sound of the in-ear microphone before and after the adjustment.

2. The noise reduction method of the Bluetooth headset noise reduction system as described in claim 1, characterized in that, The determined target noise reduction frequency band includes: Read the playback audio sequence after Bluetooth audio decoding, and determine the predictable playback sound component formed when the playback audio sequence is transmitted to the in-ear microphone according to the playback sound subtraction filter coefficient table; The residual sound from the in-ear microphone is obtained by subtracting the predictable playback sound component from the sound collected by the in-ear microphone. Perform a short-time Fourier transform on the residual sound from the in-ear microphone to obtain the frequency domain amplitude of the residual sound within the current noise reduction update cycle; Based on the amplitude of residual sound in the ear, the target noise reduction frequency band is determined from multiple candidate noise reduction frequency bands.

3. The noise reduction method of the Bluetooth headset noise reduction system as described in claim 2, characterized in that, The specific steps for determining the target noise reduction frequency band from multiple candidate noise reduction frequency bands are as follows: The noise reduction frequency range is divided to obtain candidate noise reduction frequency bands; Calculate the residual sound rise of each candidate noise reduction frequency band in the current noise reduction update cycle; The candidate noise reduction frequency band with the largest residual sound rise was determined as the candidate target noise reduction frequency band; When the residual sound rise of the candidate target noise reduction frequency band reaches the residual sound rise boundary, and the number of noise reduction update cycles that continuously reach the residual sound rise boundary is not less than the number of consecutive times, the candidate target noise reduction frequency band is determined as the target noise reduction frequency band. The residual sound rise is determined based on the root mean square amplitude of the residual sound in the in-ear microphone within the candidate noise reduction frequency band, the reference value of the residual sound amplitude under stable noise reduction, and the residual sound fluctuation amplitude under stable noise reduction.

4. The noise reduction method of the Bluetooth headset noise reduction system as described in claim 1, characterized in that, The determination of the loudspeaker electroacoustic load variation value includes: Perform a short-time Fourier transform on the speaker driving voltage sequence and speaker coil current sequence within the current noise reduction update cycle to obtain the current voltage frequency domain amplitude and the current frequency domain amplitude within the target noise reduction frequency band; The current loudspeaker electroacoustic load value is determined based on the ratio of the current voltage frequency domain amplitude to the current current frequency domain amplitude within the target noise reduction frequency band. Based on the loudspeaker drive voltage sequence and loudspeaker coil current sequence in the stable reference data, determine the stable reference loudspeaker electroacoustic load value; The change in loudspeaker electroacoustic load is determined based on the degree of change of the current loudspeaker electroacoustic load value relative to the stable reference loudspeaker electroacoustic load value.

5. The noise reduction method of the Bluetooth headset noise reduction system as described in claim 4, characterized in that, The specific steps for prohibiting the writing of noise reduction adjustment amounts formed by the residual sound of the current in-ear microphone are as follows: Maintain the current effective anti-noise amplitude parameters and anti-noise output timing parameters for the target noise reduction frequency band; The noise reduction adjustment corresponding to the residual sound of the current in-ear microphone is not written into the anti-noise real-time control parameters of the target noise reduction frequency band; The adjustment limit ratio is determined based on the ratio of the load change boundary to the loudspeaker electroacoustic load change value; Based on the adjustment limit ratio, the single-cycle change of the anti-noise amplitude parameter and the single-cycle change of the anti-noise output parameter are respectively limited within the subsequent noise reduction update cycle.

6. The noise reduction method of the Bluetooth headset noise reduction system as described in claim 1, characterized in that, The determination of the speaker's residual vibration time variation value includes: When the residual sound from the in-ear microphone reaches the residual sound rise boundary and the change in the speaker electroacoustic load does not reach the load change boundary, a load permission mark that allows correction is generated. When the load permission is marked as allowing correction, the envelope of the speaker coil current sequence corresponding to the target noise reduction frequency band in the current noise reduction update cycle is extracted to obtain the current current envelope sequence; Retrieve from the current current envelope sequence a decreasing segment that starts from a local peak and continuously decreases to below a predetermined proportion of the local peak; The time length between the local peak sampling point and the attenuation end sampling point in the falling segment is determined as the current residual oscillation time; The reference residual oscillation time is determined based on the speaker coil current sequence in the stable reference data; The change value of the speaker's residual vibration time is determined based on the degree of change of the current residual vibration time relative to the reference residual vibration time.

7. The noise reduction method of the Bluetooth headset noise reduction system as described in claim 6, characterized in that, The generated anti-noise adjustment scheme includes: When the change value of the speaker's residual vibration time is greater than the residual vibration extension boundary, the amplitude parameter of the anti-noise currently in effect in the target noise reduction frequency band remains unchanged, and the number of sampling points of the output time parameter of the currently in effect anti-noise is fixed in advance to generate candidate anti-noise output time parameters; When the change value of the speaker's residual vibration time is less than the residual vibration shortening boundary, the amplitude parameter of the anti-noise currently in effect in the target noise reduction frequency band remains unchanged, and the output time parameter of the currently in effect anti-noise is delayed by a fixed number of sampling points to generate candidate anti-noise output time parameters; When the change value of the speaker's residual vibration time is not greater than the residual vibration extension boundary and not less than the residual vibration shortening boundary, the output time parameter of the currently effective anti-noise in the target noise reduction frequency band remains unchanged, and the amplitude parameter of the currently effective anti-noise is limited to enhance the amplitude, generating candidate anti-noise amplitude parameters. When there is no continuous sampling segment in the current current envelope sequence that satisfies the falling segment condition, keep the current effective anti-noise output time parameter of the target noise reduction frequency band unchanged, and perform low-amplitude enhancement on the current effective anti-noise amplitude parameter to generate candidate anti-noise amplitude parameters. The amount of change in the low-amplitude enhancement is less than the amount of change in the defined amplitude enhancement.

8. The noise reduction method of the Bluetooth headset noise reduction system as described in claim 1 or 7, characterized in that, The real-time noise control parameters for generating the target noise reduction frequency band include: After executing the candidate anti-noise output time parameter or candidate anti-noise amplitude parameter, the in-ear microphone residual sound sequence of the target noise reduction frequency band is collected to form the adjusted evaluation window data; The improvement value of residual sound in the ear is determined based on the evaluation window data before and after adjustment; When the improvement value of residual sound in the ear reaches the improvement writing boundary, the candidate anti-noise output time parameter or candidate anti-noise amplitude parameter is written into the anti-noise real-time control parameter corresponding to the target noise reduction frequency band. When the improvement value of residual sound in the ear does not reach the improvement writing boundary and is not lower than the deterioration boundary, the candidate anti-noise parameters are not written, and the anti-noise amplitude parameters and anti-noise output time parameters of the target noise reduction frequency band in the stable reference period are restored. When the improvement value of residual sound in the ear is lower than the deterioration boundary, the candidate anti-noise parameters are not written, the anti-noise amplitude parameters and anti-noise output time parameters of the target noise reduction frequency band in the stable reference period are restored, and the subsequent adjustment upper limit of the target noise reduction frequency band in the next noise reduction update period is reduced.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the noise reduction method of the Bluetooth headset noise reduction system according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the noise reduction method of the Bluetooth headset noise reduction system according to any one of claims 1 to 8.

Images