Head-mounted equipment environment sound noise reduction method and device based on perception priority

Abstract

The invention provides a head-mounted equipment environment sound noise reduction method and device based on perception priority. The method comprises the following steps: continuously collecting environment noise through a microphone; performing spectral analysis on the environmental noise to obtain sensing parameters of the environmental noise; determining the interference degree of the environmental noise based on the sensing parameters; if the interference degree is greater than a preset interference degree threshold value, dividing the interference degree into primary interference noise and secondary interference noise according to the size of the interference degree; dynamically allocating noise reduction resources according to the interference degree levels of the primary interference noise and the secondary interference noise so as to preferentially suppress the primary interference noise, and performing adaptive noise reduction on the secondary interference noise; when the main interference noise disappears, the secondary interference noise is updated to be the current main interference noise, and the noise reduction strategy is adjusted to preferentially process the updated main interference noise, so that the problems of low processing efficiency and large resource waste of a traditional noise reduction technology in a complex noise environment are effectively solved, and the optimal matching of a noise suppression effect and processing resources is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of environmental noise processing, and in particular to a head-mounted device environmental sound noise reduction method and device based on perception priority. BACKGROUND

[0002] In today's society, with the rapid development of technology, intelligent head-mounted devices such as AR / MR / AI glasses have gradually become an indispensable tool in people's daily life and work. Such devices are widely used in education, entertainment, office, industry and other fields, providing users with immersive visual experience and convenient information interaction.

[0003] However, in actual use, the environment in which the user is located often has various complex noises, such as noisy voices in a shopping mall, vehicle noise on the road, wind and rain sounds outdoors, and equipment running sounds in an office, etc. These noises not only interfere with the user's focus on the device content, reducing the use experience, but also may potentially harm the user's hearing health. In addition, traditional noise reduction techniques mostly use fixed noise reduction modes, which cannot be dynamically adjusted according to the characteristics of environmental noise, resulting in limited noise reduction effect and failing to meet the diverse needs of users in different scenarios. SUMMARY

[0004] The present application provides a head-mounted device environmental sound noise reduction method and device based on perception priority, to solve the technical problems of poor noise reduction effect and unreasonable resource allocation of head-mounted devices in the prior art.

[0005] In one aspect, the present application provides a head-mounted device environmental sound noise reduction method based on perception priority, the head-mounted device being provided with a microphone, the method comprising: continuously collecting environmental noise through the microphone; performing spectral analysis on the environmental noise to obtain perception parameters of the environmental noise; determining the interference degree of the environmental noise based on the perception parameters; if the interference degree is greater than a preset interference degree threshold, then according to the size of the interference degree, the interference degree is divided into primary interference noise and secondary interference noise; dynamically allocating noise reduction resources according to the interference degree levels of the primary interference noise and the secondary interference noise, to preferentially suppress the primary interference noise and adaptively reduce the secondary interference noise; when the primary interference noise disappears, updating the secondary interference noise to the current primary interference noise, and adjusting the noise reduction strategy to preferentially process the updated primary interference noise.

[0006] In another aspect, the present application also provides a headset ambient sound noise reduction device based on perceptual priority, the headset being provided with a microphone, the device comprising: a noise acquisition module configured to continuously collect ambient noise through the microphone; a spectrum analysis module configured to perform spectrum analysis on the ambient noise to obtain perceptual parameters of the ambient noise; an interference determination module configured to determine an interference degree of the ambient noise based on the perceptual parameters; a noise division module configured to, if the interference degree is greater than a preset interference degree threshold, divide the interference degree into primary interference noise and secondary interference noise according to the size of the interference degree; a hierarchical noise reduction module configured to dynamically allocate noise reduction resources according to the interference degree levels of the primary interference noise and the secondary interference noise, to preferentially suppress the primary interference noise and adaptively reduce the secondary interference noise; a level adjustment module configured to, when the primary interference noise disappears, update the secondary interference noise as the current primary interference noise and adjust the noise reduction strategy to preferentially process the updated primary interference noise.

[0007] The headset ambient sound noise reduction method and device based on perceptual priority provided by the present application can intelligently identify and preferentially process primary noise that has a significant impact on human ear perception by establishing a noise interference degree grading mechanism, can automatically adjust the processing strategy according to noise dynamic changes, optimizes resource utilization while ensuring noise reduction effect, effectively solves the problems of low processing efficiency and large resource waste of traditional noise reduction technology in a complex noise environment, and realizes optimal matching of noise suppression effect and processing resources. The scheme can automatically adjust the processing priority according to the characteristics of ambient noise, quickly responds when strong noise bursts, maintains noise reduction continuity when the noise environment changes, significantly improves user auditory comfort and reduces device energy consumption. BRIEF DESCRIPTION OF DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the present application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0009] Figure 1 is a flowchart of the headset ambient sound noise reduction method based on perceptual priority provided by the embodiments of the present application; Figure 2 is a structural schematic diagram of the headset ambient sound noise reduction device based on perceptual priority provided by the embodiments of the present application; Figure 3 is a structural schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0010] To make the objectives, technical solutions, and advantages of the present application clearer, the technical solutions in the present application will be described below in conjunction with the accompanying drawings in the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort belong to the scope of protection of the present application.

[0011] Figure 1 is a flowchart of a head-mounted device ambient sound noise reduction method based on a perception priority provided by an embodiment of the present application. The head-mounted device is provided with a microphone.

[0012] Referring to Figure 1 , the head-mounted device ambient sound noise reduction method based on a perception priority comprises the following steps 101 to 106.

[0013] Step 101, continuously collecting ambient noise through the microphone.

[0014] Step 102, performing spectrum analysis on the ambient noise to obtain a perception parameter of the ambient noise.

[0015] In this step, the spectrum analysis refers to a process of time-frequency domain conversion and feature extraction on the noise signal, which can be implemented by using fast Fourier transform combined with a psychoacoustic model, and is used to accurately reflect the influence characteristics of the noise on human ear perception.

[0016] Step 103, determining an interference degree of the ambient noise based on the perception parameter.

[0017] In this step, the interference degree refers to a quantitative index of the comprehensive noise physical characteristics and human ear perception sensitivity, which can be obtained by weighted summation after normalization processing on the frequency, loudness, and other parameters, and is used to objectively evaluate the interference degree of different noises.

[0018] Step 104, if the interference degree is greater than a preset interference degree threshold, then according to the size of the interference degree, the interference degree is divided into a main interference noise and a secondary interference noise.

[0019] Step 105, dynamically allocating noise reduction resources according to the interference degree levels of the main interference noise and the secondary interference noise, to preferentially suppress the main interference noise, and to adaptively reduce the secondary interference noise.

[0020] In this step, dynamically allocating the noise reduction resource refers to adjusting the resource configuration of the signal processing channel according to the noise priority adjustment signal, which can be implemented by using a multi-channel filter bank combined with a resource scheduling algorithm to ensure that the main noise obtains the maximum suppression effect.

[0021] In step 106, when the main interference noise disappears, the secondary interference noise is updated as the current main interference noise, and the noise reduction strategy is adjusted to preferentially process the updated main interference noise.

[0022] In this embodiment, specifically, the environmental sound signals are collected in real time by a microphone, the continuous audio stream is divided into frame units of fixed time length for spectral analysis, and parameters such as frequency distribution and energy intensity reflecting human ear perception characteristics are extracted. Based on the psychoacoustic model, the parameters are weighted and calculated to obtain a quantization value representing the noise interference degree. When the value exceeds a set threshold, the noise is divided into main and secondary categories according to the interference degree, and the active noise reduction module is preferentially called to process the main noise, while the remaining processing capacity is used to implement filtering and suppression on the secondary noise. When the main noise is detected to disappear, the system automatically upgrades the secondary noise to the main processing object, and reconfigures the noise reduction parameters to ensure the continuity of processing.

[0023] The embodiment can intelligently identify and preferentially process the main noise that significantly affects human ear perception by establishing a noise interference degree grading mechanism, and can automatically adjust the processing strategy according to the dynamic changes of the noise, thereby optimizing the resource utilization rate while ensuring the noise reduction effect. Through the above technical solutions, the embodiment effectively solves the problems of low processing efficiency and large resource waste of traditional noise reduction technology in a complex noise environment, and realizes the optimal matching of noise suppression effect and processing resources. The scheme can automatically adjust the processing priority according to the characteristics of the environmental noise, quickly respond when a strong noise occurs, maintain noise reduction continuity when the noise environment changes, significantly improve the user's auditory comfort, and reduce the energy consumption of the device.

[0024] In an embodiment of the present application, the perception parameters include frequency, loudness, sharpness and transient degree, and based on the perception parameters, the interference degree of the environmental noise is determined, including: Based on the physical characteristics of frequency, loudness, sharpness and transient degree and the human ear perception law, the frequency, loudness, sharpness and transient degree are normalized respectively to obtain frequency normalization result, loudness normalization result, sharpness normalization result and transient degree normalization result. The frequency normalization result, the loudness normalization result, the sharpness normalization result and the transient degree normalization result are weighted and summed to obtain the interference degree of the environmental noise.

[0025] In the embodiment, by introducing a multi-dimensional perception parameter normalization and weighted fusion mechanism, the perception characteristics of the human ear auditory system in a complex noise environment can be more accurately simulated, thereby improving the accuracy of the interference degree evaluation. For example, in a scenario where both low-frequency continuous noise and high-frequency transient noise exist, the traditional method may underestimate the interference degree of the transient noise, while the present solution can effectively identify the interference characteristics of such noise through nonlinear mapping of the transient degree parameter. Through the above technical solution, the embodiment can solve the interference degree evaluation deviation problem caused by the single perception dimension of the traditional noise reduction technology. By integrating the perception parameters of four dimensions, including frequency distribution, loudness intensity, high-frequency component proportion, and noise duration, and performing normalization processing and weighted fusion based on human ear perception rules, the actual interference degree of different noise components on the human ear can be accurately quantified. This provides an accurate decision basis for subsequent dynamic allocation of noise reduction resources, enabling the noise reduction system to prioritize processing noise components that have a greater impact on human ear perception, thereby achieving a more optimal overall noise reduction effect under limited computing resources. For example, in an office environment, the method can effectively identify the perception difference between the transient characteristics of keyboard tapping sound and the continuous characteristics of air conditioner noise, and adjust the noise reduction strategy to balance the auditory comfort and speech clarity requirements.

[0026] In an embodiment of the present specification, the loudness is normalized in the following manner: The loudness is normalized by logarithmic compression, as shown in the following formula (1): (1); In formula (1), is the normalized value of noise perception loudness. is the loudness. Set The input range of is [20, 120], and the unit is phon. is the minimum value, which can be 20. is the maximum value, which can be 120.

[0027] The frequency is normalized in the following manner: First, the frequency is converted to a Bark scale value; wherein the Bark scale value reflects the actual frequency band division of the human ear; The Bark scale value is linearly scaled, as shown in the following formula (2): (2); In formula (2), is the normalized value of noise frequency. The original frequency can be mapped by Bark scale and then directly linearly mapped to [1, 100]. B is the Bark scale mapping value, ranging from about (0, 24), corresponding to the critical frequency band of human hearing.

[0028] The sharpness is normalized in the following manner: The sharpness is linearly mapped to a preset interval range in proportion; When the sharpness is greater than a preset sharpness threshold, the sharpness is mapped to a maximum value of the preset interval range, as shown in the following formula (3): (3); In formula (3), is a normalized value of the noise sharpness, and the sharpness is usually less than 5 acum, and the extreme sound can be close to 6-8 acum. S is the original sharpness.

[0029] The transient degree is normalized in the following manner: The duration frame number of the transient degree is logarithmically transformed to obtain a logarithmically transformed value; The logarithmically transformed value is offset compensated and taken as an inverse; Based on the monotone decreasing characteristic of the inverse, the inverse is mapped to a normalized result of the transient degree, as shown in the following formula (4): (4); In formula (4), is a normalized value of the transient degree, is the duration frame number, and the range is .

[0030] The calculation method of the interference degree is shown in the following formula (5): (5); In formula (5), is the interference degree of the current noise. 、 、 and are weight values of the frequency, the loudness, the sharpness and the transient degree, respectively. is a sum of all the weights.

[0031] In the embodiment, specifically, in the noise processing process, the loudness parameter is normalized by logarithmic compression to more accurately reflect the nonlinear perception characteristics of the human ear to sound pressure level, avoiding the weight of high loudness noise being excessively amplified. The frequency parameter is converted by Bark scale and linearly scaled, and its numerical distribution matches the human ear auditory critical band division, which can effectively distinguish the perception priority of different frequency band noises. The sharpness parameter is processed by linear mapping in the preset interval, which not only retains the distinguishing degree of conventional sharp noise, but also limits the amplitude of abnormal sharp sound that exceeds the threshold to prevent a single parameter from dominating the overall interference degree evaluation result. The transient degree parameter is mapped to a high normalized value for burst noise with a short duration and a low normalized value for steady noise with a long duration through logarithmic transformation and reciprocal operation of the number of continuous frames, accurately representing the interference characteristics of transient noise on human attention.

[0032] The embodiment accurately simulates the human ear frequency band perception characteristics through Bark scale conversion, effectively retains key noise characteristics through logarithmic compression and segmented mapping processing, and significantly improves the accuracy of the interference degree evaluation. Through the above technical solutions, the embodiment can more accurately quantify the influence degree of different noise characteristics on human perception, making the interference degree evaluation result more consistent with the actual auditory perception. By establishing a scientific parameter conversion relationship, the evaluation deviation caused by the dimensional difference of the parameters in the traditional method is effectively avoided, providing a reliable decision basis for subsequent noise reduction resource allocation. For the different characteristics of transient burst noise and continuous steady noise, a differentiated normalization processing method is adopted to ensure that the interference degree of all types of noise can be accurately represented, thereby improving the adaptability of the noise reduction system to complex environmental noise.

[0033] In an embodiment of the present specification, according to the interference degree levels of the main interference noise and the secondary interference noise, the noise reduction resources are dynamically allocated to preferentially suppress the main interference noise and adaptively reduce the secondary interference noise, including: Generating reverse sound waves for active cancellation of the main interference noise; Specifically, the reverse sound wave active cancellation refers to achieving acoustic energy cancellation by generating sound waves opposite in phase to the noise. Specifically, a digital signal processor can be used to calculate the noise waveform in real time and generate a reverse waveform to achieve this, which can accurately eliminate the noise energy in a specific frequency band.

[0034] According to the remaining resources, the secondary interference noise is selected for partial suppression or filtering processing.

[0035] Specifically, the residual resource refers to the operation capability and hardware resource remaining in the system after the main noise is processed, which can be quantitatively evaluated by monitoring the processor load and memory occupancy in real time. The mechanism ensures that the resource allocation conforms to the actual processing capability. The partial suppression refers to the incomplete elimination of the secondary noise, which can be realized by reducing the noise reduction depth or narrowing the effective frequency width. The measure can maintain the basic noise reduction effect under limited resources. The filtering processing refers to weakening the noise energy through frequency domain filtering, which can be realized by using a finite impulse response filter or an adaptive filter. This method is suitable for wide-band suppression of the secondary noise.

[0036] In the embodiment, specifically, when the main interference noise is detected, the reverse sound wave matching the spectral characteristics of the noise is first generated, and the active noise reduction is realized through sound wave superposition. In this process, the processor resource is preferentially allocated to the reverse sound wave generation module to ensure the real-time performance and accuracy of the main noise elimination. For the secondary interference noise, a dynamic decision is made according to the real-time residual operation capability: when the residual resource is sufficient, the reverse sound wave counteraction is performed on the secondary noise in a partial frequency band; when the resource is insufficient, the filtering mode is switched to, and the secondary noise is attenuated in a wide frequency band by using a preset filter set. The hierarchical processing mechanism enables the noise reduction system to maintain the basic performance when the resource is limited, and avoids system overload caused by excessive resource allocation.

[0037] The embodiment establishes a resource allocation priority mechanism to flexibly adjust the processing mode of the secondary noise while ensuring the main noise elimination effect, thereby significantly improving the system resource utilization. Through the above technical solution, the embodiment effectively solves the problem of unreasonable noise reduction resource allocation in a complex noise environment, and can adaptively adjust the secondary noise processing strategy according to the real-time load of the system while maintaining the main noise elimination effect. The dynamic resource management mechanism not only avoids the performance degradation of the system caused by excessive resource consumption in the traditional scheme, but also ensures the differentiated processing needs of noises with different priorities, thereby significantly improving the noise reduction stability of the head-mounted device in a variable noise environment.

[0038] In an embodiment of the present specification, when the main interference noise disappears, the secondary interference noise is updated as the current main interference noise, and the noise reduction strategy is adjusted to preferentially process the updated main interference noise, including: Step one, the noise reduction priority of the secondary interference noise is raised to the level of the main interference noise; Specifically, the noise reduction priority raising refers to adjusting the processing level of the secondary noise to the highest level, which can be realized by modifying the weight coefficient in the algorithm. A priority mapping table is established in the noise classification module, and the priority adjustment mechanism is automatically triggered when the main noise disappears.

[0039] Step two, re-allocate the noise reduction resources, switch the main noise reduction processing channel to the updated main interference noise; Specifically, re-allocating the noise reduction resources refers to adjusting the operation channel allocation of the digital signal processor, which can be realized by using a dynamic memory allocation algorithm. The resources of the original main noise processing channel are released and re-allocated to the new main noise processing task.

[0040] Step three, according to the frequency characteristics of the updated main interference noise, select the corresponding noise reduction mode for noise reduction processing.

[0041] Specifically, the noise reduction mode selection refers to matching the preset suppression strategy according to the noise spectrum characteristics, which can be realized by establishing a mapping relationship between the frequency characteristics and the filter parameters, for example, selecting a band-stop filter for medium-high frequency noise and selecting phase inversion cancellation for low frequency noise.

[0042] In this embodiment, specifically, when it is detected that the sound pressure level of the current main noise is continuously lower than the set threshold, the noise state updating mechanism is triggered. At this time, the noise with the highest interference degree in the secondary noise queue is marked as the new main noise, and its priority parameter is set to the highest level. The signal processor immediately releases the adaptive filter and active noise reduction module resources occupied by the original main noise, and re-allocates these resources to the newly marked main noise processing channel. At the same time, according to the distribution characteristics of the new main noise on the Bark scale, the corresponding filter bank configuration is automatically selected, for example, for noise concentrated in the 2-5 kHz frequency band, a composite processing mode combining high-frequency band-stop filtering and phase compensation is enabled.

[0043] This embodiment can quickly reconfigure the processing channel when the noise environment changes by establishing a dynamic priority adjustment mechanism, ensure that the noise reduction resources always focus on the current most interfering noise component, and optimize the filter configuration through the frequency characteristic matching mechanism to improve the suppression accuracy of the noise in a specific frequency band. Through the above technical solutions, this embodiment can quickly establish a new noise reduction focus after the main noise source disappears, avoiding the noise reduction lag problem caused by the dynamic change of the environmental noise. Through the real-time resource re-allocation mechanism, the processing efficiency is effectively improved, ensuring that the noise reduction system always optimizes the processing of the current highest priority noise. At the same time, the noise reduction mode selection based on the frequency characteristics can implement accurate suppression for different noise types, maintaining the continuity of the noise reduction effect.

[0044] In an embodiment of the present application, the environmental noise is subjected to spectrum analysis to obtain the perceptual parameters of the environmental noise, including: Step one, frame processing of continuously collected environmental noise, and extracting spectrum characteristics; In this step, the frame processing refers to dividing the continuous ambient noise signal into data frames of fixed length, which can be implemented in an overlapping frame manner, for example, the frame length is 20 milliseconds and the frame shift is 10 milliseconds, and the spectral leakage is suppressed by a window function. This processing method can effectively capture the transient characteristics of the noise signal and avoid the response delay caused by long-time signal analysis.

[0045] Step two, based on the spectral features, the perceptual parameters of the ambient noise are calculated.

[0046] In this step, the spectral feature extraction refers to converting the time domain signal into frequency domain energy distribution, which can be implemented by fast Fourier transform, and the energy proportion of each frequency band, peak frequency position and harmonic component are calculated to provide basic data for perceptual parameter calculation. This step solves the problem of ignoring the dynamic change characteristics of noise in traditional methods, making the subsequent interference degree evaluation more in line with the actual auditory perception.

[0047] In this embodiment, through frame processing combined with window function, both the transient characteristics of sudden noise and the continuity of steady-state noise analysis can be captured. In addition, based on the calculation method of dynamic spectral features, compared with the fixed frequency band energy integration method, the noise components with auditory salience can be more accurately identified. Through the above technical solutions, the embodiment realizes the fine analysis of noise features, so that the subsequent interference degree classification has reliable quantitative basis. Frame processing ensures the real-time of noise analysis, and spectral feature extraction lays a foundation for multi-dimensional perceptual parameter calculation, finally supports the noise reduction system to implement differentiated processing strategies for different noise characteristics, significantly improves the noise reduction efficiency and resource utilization rate in complex acoustic environment.

[0048] In an embodiment of the present application, after the interference degree is divided into main interference noise and secondary interference noise, it further includes: Step one, detecting whether there is a frequency band coupling relationship between the secondary interference noise and the main interference noise; In this step, the frequency band coupling relationship refers to the energy overlap or spectral overlap phenomenon of the secondary interference noise and the main interference noise in a specific frequency range, which can be realized by calculating the spectral energy correlation coefficient of the two in the preset frequency band, for example, calculating the energy distribution similarity after extracting the frequency domain features by fast Fourier transform.

[0049] Step two, if there is, the secondary interference noise is included in the noise reduction processing channel of the main interference noise, and the independent noise reduction channel of the secondary noise is closed, and the main noise processing resources are completely shared; In this step, the noise reduction processing channel refers to an independent hardware module or software thread resource for generating a reverse sound wave or performing a filtering algorithm, which can be implemented by using a multi-path parallel computing unit in a digital signal processor. The shared main noise processing resource refers to dynamically switching the operation capability, storage space or filter group originally allocated to the secondary noise to the processing link corresponding to the main noise, for example, completing resource scheduling by register remapping or memory address switching.

[0050] In this embodiment, specifically, when it is detected that the secondary interference noise and the main interference noise have an energy overlap exceeding a preset threshold in a frequency band of 1 kHz to 4 kHz, the system automatically triggers the coupling processing mechanism. At this time, the spectral characteristics of the secondary noise are highly related to the main noise, and the independent noise reduction channel will produce redundant calculation. By inputting the spectral data of the secondary noise into the active noise reduction module of the main noise, the same set of adaptive filters is used to process the common frequency band of the two types of noise. The power supply of the secondary noise independent channel is cut off, the cache space occupied by the secondary noise is released and allocated to the real-time operation demand of the main channel. The process is triggered by a hardware interrupt signal, and the resource switching is quickly completed to ensure that the noise reduction continuity is not affected.

[0051] This embodiment dynamically merges noise sources with spectral correlation by real-time detection of the frequency band coupling relationship, eliminates redundant operation units, and avoids phase cancellation or gain abnormality problems that may occur when multiple channels are processed in parallel. Through the above technical solutions, the embodiment effectively solves the problem of low hardware resource utilization in the multi-noise source noise reduction scene, while maintaining the main noise suppression effect, reducing the overall power consumption of the system, reducing the algorithm processing delay, and avoiding the secondary noise pollution phenomenon caused by the cross interference of multiple channels.

[0052] In an embodiment of the present application, the secondary interference noise is included in the noise reduction processing channel of the main interference noise, and the independent noise reduction channel of the secondary noise is closed, completely sharing the main noise processing resource, including: Step one, calculate the spectral energy overlap ratio of the main noise and the secondary noise in a preset frequency band (such as 1 kHz~4kHz) as the coupling coefficient; In this step, the preset frequency band (1 kHz~4kHz) refers to the range of human ear sensitive frequency band, which can be realized by dividing the noise signal into frequency bands through a band-pass filter or a spectrum analyzer. The selection of this frequency band is based on the human ear hearing characteristics, which can reflect the key area of the influence of noise on human perception. The coupling coefficient refers to the similarity of the energy distribution of the two noise signals in the same frequency band, which can be calculated by using the spectral energy integration method to calculate the proportion of the energy in the overlap region to the total energy. This parameter is used to quantify the frequency band correlation between the noises, and provides a basis for resource allocation.

[0053] Step two, when the coupling coefficient is greater than the first preset coupling threshold, it is determined as strong coupling, the secondary interference noise is included in the main interference noise noise reduction processing channel, and the secondary noise independent noise reduction channel is closed, and the main noise processing resource is completely shared; In this step, the first preset coupling threshold and the second preset coupling threshold refer to the critical value for judging the coupling strength of the noise, which can be realized by setting the grading threshold value through experimental statistics of noise coupling data in different scenes. Threshold division can distinguish between strong coupling and weak coupling states, guiding the merging or retention of the noise reduction channel.

[0054] Step three, when the coupling coefficient is less than or equal to the first preset coupling threshold and greater than or equal to the second preset coupling threshold, it is determined as weak coupling, and the secondary noise noise reduction channel is retained.

[0055] In this embodiment, after detecting the secondary interference noise, the spectral energy distribution of the main noise and the secondary noise is first analyzed in real time, and the energy overlap ratio of both in the 1kHz~4kHz frequency band is calculated. For example, if the main noise occupies 60% of the energy in the frequency band, the secondary noise occupies 40% of the energy, and the overlapping part accounts for 80% of the total energy of the secondary noise, then the coupling coefficient is 80%. When the coefficient exceeds the first preset coupling threshold (for example, 70%), it is determined as a strong coupling state, at this time the frequency band characteristics of the secondary noise are highly overlapped with the main noise, and joint noise reduction can be realized by sharing the main noise processing channel, and the secondary noise independent channel is closed to save resources. If the coupling coefficient is between the second preset coupling threshold (for example, 30%) and the first threshold, the secondary noise independent channel is retained for separate processing.

[0056] This embodiment dynamically evaluates the noise coupling state, merges the processing channel when the frequency band overlap is significant, effectively reduces redundant calculation, and improves resource utilization. Through the above technical solution, this embodiment can avoid repeated noise reduction processing of the frequency band overlap noise, reduce the system operation load, and at the same time ensure the cooperative suppression effect of the coupled noise. In a complex acoustic scene, this scheme can adaptively adjust the channel configuration, solving the problem of low noise reduction efficiency caused by rigid resource allocation in traditional methods.

[0057] In an embodiment of the present application, after dynamically allocating noise reduction resources according to the interference levels of the main interference noise and the secondary interference noise, it further comprises: Step one, detecting whether the current main interference noise has an acoustic shadowing effect on the secondary interference noise; In this step, the acoustic shadowing effect refers to the phenomenon that when two sound sources exist at the same time, the stronger sound source will mask the weaker sound source. This feature can be realized by calculating the spectral energy difference between the main noise and the secondary noise, which is used to identify whether the secondary noise is naturally suppressed by the main noise.

[0058] Step two, if yes, evaluate the degree of masking of the current primary interference noise to the secondary interference noise; In this step, the degree of masking refers to quantifying the covering strength of the primary noise to the secondary noise, which can be realized by calculating the energy attenuation ratio of the secondary noise in the masking frequency band. This feature is used to determine whether active noise reduction is needed.

[0059] Step three, if the degree of masking is higher than the preset threshold, dynamically reduce the noise reduction strength of the secondary interference noise; In this step, dynamically reducing the noise reduction strength refers to adjusting the processing strength of the secondary noise according to the degree of masking, which can be realized by adjusting the filter bandwidth or reducing the amplitude of the reverse sound wave. This feature is used to optimize the efficiency of noise reduction resource allocation.

[0060] Step four, allocate the saved noise reduction resources to the primary interference noise processing.

[0061] In this embodiment, specifically, when there is an overlap between the primary noise and the secondary noise in the frequency band, the system detects whether the primary noise forms energy suppression to the secondary noise in a specific frequency band through real-time spectrum analysis. If the masking effect is detected, the energy attenuation ratio of the frequency band of the secondary noise that is masked is calculated, and when the ratio exceeds the preset threshold, it is determined that the secondary noise has been naturally suppressed by the primary noise. At this time, the system automatically reduces the active noise reduction strength for the secondary noise, such as reducing the number of filter channels or reducing the generation rate of the reverse sound wave, and concentrates the released computing resources on enhancing the cancellation processing of the primary noise. This dynamic adjustment process realizes the optimization of noise reduction efficiency through real-time monitoring and resource reallocation mechanism.

[0062] In some specific embodiments, the degree of masking can be realized by calculating the energy attenuation amplitude of the secondary noise in the 1kHz to 4kHz frequency band, and when the amplitude exceeds 60%, the noise reduction strength adjustment is triggered. The noise reduction strength adjustment can be manifested as reducing the computing resources of the secondary noise processing channel by 50%, and using the saved resources to enhance the algorithm iteration frequency of the primary noise processing channel.

[0063] This embodiment can effectively improve the resource utilization rate while ensuring the noise reduction effect by real-time detection of the degree of masking and dynamic adjustment of the processing strength, avoiding redundant processing of noises that have been naturally masked. Through the above technical solutions, this embodiment can intelligently optimize the allocation of noise reduction resources according to the actual acoustic environment, reduce invalid energy consumption in complex noise scenes, improve the suppression accuracy of the primary noise, and at the same time avoid the problem of sound field distortion caused by excessive noise reduction processing. This scheme is particularly suitable for environments such as shopping malls and traffic intersections where there are sudden strong noises, and can prolong the device's battery life while ensuring user experience.

[0064] In some other embodiments of this specification, after dynamically allocating noise reduction resources according to the interference levels of primary and secondary interference noise, the following steps may also be included: Step 1: Monitor the user's physiological state in real time, as shown below: The device utilizes a variety of biosensors integrated into the head-mounted device (such as heart rate sensors, skin conductance sensors, and electroencephalogram (EEG) sensors) to monitor the user's physiological state in real time. Heart rate sensors are used to detect changes in a user's heart rate and identify whether the user is in a state of tension or relaxation; Electrodermal response sensors are used to measure skin conductivity and assess a user's stress level. Electroencephalography (EEG) sensors are used to capture brain activity patterns and determine a user's level of concentration or fatigue.

[0065] Step 2: Assess the actual impact of current environmental noise on users based on physiological state data and adjust the interference weight accordingly, as shown below: The collected physiological state data is input into a pre-trained machine learning model, which can identify the relationship between different physiological parameters and the user's psychological state. By analyzing users' physiological state data through models, the user's current psychological state (such as tension, relaxation, focus, etc.) can be determined. Assess the actual level of disturbance caused by current environmental noise based on the user's psychological state, for example: For users under stress or with high attention requirements, increase the interference weight of the main interfering noise to enhance the noise reduction effect. For users in a relaxed state, reduce the interference weight of secondary noise and reduce the noise reduction processing, thereby saving resources.

[0066] In this embodiment, the interference weights of primary and secondary noise are dynamically adjusted by combining the user's physiological state data to more accurately reflect the real impact of noise on the user.

[0067] Based on the same general inventive concept, this invention also protects a head-mounted device for ambient sound noise reduction based on perception priority, such as... Figure 2 As shown, Figure 2 This is a schematic diagram of the structure of the head-mounted device ambient sound noise reduction device based on perception priority provided in an embodiment of the present invention. The following describes the head-mounted device ambient sound noise reduction device based on perception priority provided by the present invention. The head-mounted device ambient sound noise reduction device based on perception priority described below can be referred to in correspondence with the head-mounted device ambient sound noise reduction method based on perception priority described above.

[0068] The headset environmental sound noise reduction device based on perceptual priority comprises a noise acquisition module 201, a spectrum analysis module 202, an interference determination module 203, a noise division module 204, a hierarchical noise reduction module 205 and a level adjustment module 206.

[0069] The noise acquisition module 201 is used for continuously collecting environmental noise through a microphone; The spectrum analysis module 202 is used for performing spectrum analysis on the environmental noise to obtain perceptual parameters of the environmental noise; The interference determination module 203 is used for determining the interference degree of the environmental noise based on the perceptual parameters; The noise division module 204 is used for, if the interference degree is greater than a preset interference degree threshold, dividing the interference degree into main interference noise and secondary interference noise according to the size of the interference degree; The hierarchical noise reduction module 205 is used for dynamically allocating noise reduction resources according to the interference degree levels of the main interference noise and the secondary interference noise, to preferentially suppress the main interference noise and adaptively reduce the secondary interference noise; The level adjustment module 206 is used for, when the main interference noise disappears, updating the secondary interference noise as the current main interference noise and adjusting the noise reduction strategy to preferentially process the updated main interference noise.

[0070] Figure 3 is a structural schematic diagram of an electronic device provided by an embodiment of the present application.

[0071] As shown in Figure 3 , the electronic device can include a processor 310, a communications interface 320, a memory 330 and a communications bus 340, wherein the processor 310, the communications interface 320 and the memory 330 complete mutual communication through the communications bus 340. The processor 310 can invoke logical instructions in the memory 330 to execute a headset environmental sound noise reduction method based on perceptual priority.

[0072] In addition, the logic instructions in the memory 330 described above can be implemented in the form of a software function unit and sold or used as an independent product, and can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the parts that contribute to the prior art or parts of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0073] In another aspect, the present application also provides a computer program product, which comprises a computer program, the computer program can be stored on a non-transitory computer readable storage medium, and the computer program can be executed by a processor to enable a computer to execute the perception priority-based headset ambient sound noise reduction method provided by the above-mentioned methods.

[0074] In yet another aspect, the present application also provides a non-transitory computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the perception priority-based headset ambient sound noise reduction method provided by the above-mentioned methods.

[0075] The device embodiments described above are only schematic, wherein the units illustrated as separate components can or can not be physically separate, and the components illustrated as units can or can not be physical units, i.e., can be located in one place, or can be distributed on a plurality of network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the present embodiment. Those skilled in the art can understand and implement without creative labor.

[0076] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be implemented by means of software and the necessary universal hardware platform, and of course can also be implemented by hardware. Based on such understanding, the technical solutions described above essentially or the parts that contribute to the prior art can be embodied in the form of a software product. The computer software product can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments.

[0077] Finally, it should be noted that the above examples are only used to illustrate the technical solutions of the present application, and are not intended to limit the same; although the present application has been described in detail with reference to the foregoing examples, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing examples, or make equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for perceptual priority based head-worn device ambient sound noise reduction, the method comprising: The head-mounted device is provided with a microphone, and the method comprises: continuously collecting environmental noise through the microphone; performing spectral analysis on the environmental noise to obtain a perceptual parameter of the environmental noise; determining an interference degree of the environmental noise based on the perceptual parameter; if the interference degree is greater than a preset interference degree threshold, dividing the interference degree into a main interference noise and a secondary interference noise according to the size of the interference degree; dynamically allocating noise reduction resources according to the interference degree levels of the main interference noise and the secondary interference noise to preferentially suppress the main interference noise and adaptively reduce the secondary interference noise; when the main interference noise disappears, updating the secondary interference noise as the current main interference noise and adjusting the noise reduction strategy to preferentially process the updated main interference noise.

2. The perceptual priority based headset ambient sound noise reduction method of claim 1, wherein, The perceptual parameter comprises frequency, loudness, sharpness and transient degree, and the determination of the interference degree of the environmental noise based on the perceptual parameter comprises: based on the physical characteristics of the frequency, the loudness, the sharpness and the transient degree and the human ear perception law, respectively normalizing the frequency, the loudness, the sharpness and the transient degree to obtain frequency normalization result, loudness normalization result, sharpness normalization result and transient degree normalization result; weighting and summing the frequency normalization result, the loudness normalization result, the sharpness normalization result and the transient degree normalization result to obtain the interference degree of the environmental noise.

3. The perceptual priority based headset ambient sound noise reduction method of claim 2, wherein, The loudness is normalized in the following way: the loudness is normalized by logarithmic compression; The frequency is normalized in the following way: firstly, convert the frequency into Bark scale value; wherein, the Bark scale value reflects the actual frequency band division of the human ear; linearly scale the Bark scale value; The sharpness is normalized in the following way: linearly map the sharpness to a preset interval range in proportion; when the sharpness is greater than a preset sharpness threshold, map the sharpness to the maximum value of the preset interval range; The transient degree is normalized in the following way: logarithmically transform the continuous frame number of the transient degree to obtain a logarithmically transformed value; offset compensate the logarithmically transformed value and take the reciprocal; based on the monotonic decreasing characteristic of the reciprocal, map it into the normalization result of the transient degree.

4. The perceptual priority based headset ambient sound noise reduction method of claim 1, wherein, Dynamically allocating noise reduction resources according to the interference degree levels of the main interference noise and the secondary interference noise to preferentially suppress the main interference noise and adaptively reduce the secondary interference noise comprises: generating reverse sound waves for the main interference noise to actively cancel it out; selectively suppressing or filtering the secondary interference noise according to the remaining resources.

5. The perceptual priority based headset ambient sound noise reduction method of claim 1, wherein, When the main interference noise disappears, updating the secondary interference noise as the current main interference noise and adjusting the noise reduction strategy to preferentially process the updated main interference noise comprises: raising the noise reduction priority of the secondary interference noise to the level of the main interference noise; redistributing the noise reduction resources and switching the main noise reduction processing channel to the updated main interference noise; According to the frequency characteristics of the updated main interference noise, a corresponding noise reduction mode is selected for noise reduction processing.

6. The perceptual priority based headset ambient sound noise reduction method of claim 1, wherein, The environmental noise is subjected to spectral analysis to obtain the perceptual parameters of the environmental noise, including: The continuously collected environmental noise is subjected to frame processing and spectral feature extraction; Based on the spectral features, the perceptual parameters of the environmental noise are calculated.

7. The perceptual priority based headset ambient sound noise reduction method of claim 1, wherein, After the interference degree is divided into main interference noise and secondary interference noise, it further includes: Detecting whether there is a frequency band coupling relationship between the secondary interference noise and the main interference noise; If there is, the secondary interference noise is included in the noise reduction processing channel of the main interference noise, and the independent noise reduction channel of the secondary noise is closed, and the main noise processing resources are completely shared.

8. The perceptual priority based headset ambient sound noise reduction method of claim 7, wherein, The secondary interference noise is included in the noise reduction processing channel of the main interference noise, and the independent noise reduction channel of the secondary noise is closed, and the main noise processing resources are completely shared, including: Calculate the spectral energy overlap ratio of the main noise and the secondary noise in the preset frequency band as the coupling coefficient; When the coupling coefficient is greater than the first preset coupling threshold, it is determined as strong coupling, the secondary interference noise is included in the noise reduction processing channel of the main interference noise, and the independent noise reduction channel of the secondary noise is closed, and the main noise processing resources are completely shared; When the coupling coefficient is less than or equal to the first preset coupling threshold and greater than or equal to the second preset coupling threshold, it is determined as weak coupling, and the secondary noise reduction channel is retained.

9. The perceptual priority based headset ambient sound noise reduction method of claim 1, wherein, After the noise reduction resources are dynamically allocated according to the interference degree levels of the main interference noise and the secondary interference noise, it further includes: Detecting whether the current main interference noise produces an acoustic shadowing effect on the secondary interference noise; If yes, evaluate the shadowing degree of the current main interference noise on the secondary interference noise; If the shadowing degree is higher than a preset threshold, the noise reduction intensity of the secondary interference noise is dynamically reduced; The saved noise reduction resources are allocated to the main interference noise processing.

10. A head-mounted device for ambient sound noise reduction based on perception priority, characterized in that, The head-mounted device is provided with a microphone, and the device includes: A noise acquisition module for continuously collecting environmental noise through the microphone; A spectral analysis module for performing spectral analysis on the environmental noise to obtain perceptual parameters of the environmental noise; An interference determination module for determining the interference degree of the environmental noise based on the perceptual parameters; A noise division module for dividing the interference degree into main interference noise and secondary interference noise according to the size of the interference degree if the interference degree is greater than a preset interference degree threshold; A hierarchical noise reduction module for dynamically allocating noise reduction resources according to the interference degree levels of the main interference noise and the secondary interference noise to preferentially suppress the main interference noise and adaptively reduce the secondary interference noise; A level adjustment module for updating the secondary interference noise to the current main interference noise when the main interference noise disappears, and adjusting the noise reduction strategy to preferentially process the updated main interference noise.

Images