Noise reduction method, system and device and storage medium

Abstract

The embodiment of the invention provides a noise reduction method, system and device and a storage medium. The method comprises the steps of determining noise frequency characteristics of environmental noise; determining a target frequency group based on the noise frequency characteristics; the target frequency group comprises one or more target frequencies; for a target frequency, generating a plurality of first candidate signals based on the plurality of candidate phases; acquiring a first feedback signal corresponding to the first candidate signal; based on the first feedback signal and the noise frequency characteristic, determining a target phase corresponding to the target frequency and an initial target signal; determining a target single-frequency signal corresponding to the target frequency based on the plurality of candidate amplitudes and the initial target signal; and based on the target single-frequency signal group corresponding to the target frequency group, completing noise reduction.

Description

Technical Field

[0001] The present invention relates to the field of noise reduction technology, and in particular to a noise reduction method, system, device and storage medium. Background Art

[0002] With the development of science and technology, environmental noise generated by social activities such as transportation and industry has gradually affected people's daily lives. ANC (Active Noise Control) noise reduction technology is widely used in the noise reduction processing of consumer electronic products. It is a technology that cancels noise by emitting sound waves with a phase opposite to the environmental noise. At present, ANC noise reduction technology is mainly good at eliminating low-frequency continuous noise (such as aircraft engine noise, air conditioning noise, etc.), but the noise reduction effect on high-frequency noise (such as human voice, sudden sharp sounds, etc.) is poor. And because the wavelength of high-frequency noise is short and changes rapidly, it is difficult to capture and generate accurate reverse sound waves in real time.

[0003] Therefore, it is desired to provide a noise reduction method, system, device and storage medium. In a scenario where the environmental noise is single and has a strong periodicity, a distributed single-point noise reduction scheme is adopted to achieve a stronger suppression capability for the main body of noise energy, avoid the influence of the system timeliness generated by real-time capture and generation on the noise reduction effect, and improve the upper limit of the noise reduction effect. Summary of the invention

[0004] One or more embodiments of the present specification provide a noise reduction method, the method comprising: determining the noise frequency characteristics of environmental noise; determining a target frequency group based on the noise frequency characteristics; the target frequency group includes one or more target frequencies; for one target frequency: generating multiple first candidate signals based on multiple candidate phases; obtaining a first feedback signal corresponding to the first candidate signal; determining a target phase and an initial target signal corresponding to the target frequency based on the first feedback signal and the noise frequency characteristics; determining a target single-frequency signal corresponding to the target frequency based on multiple candidate amplitudes and the initial target signal; and completing noise reduction based on the target single-frequency signal group corresponding to the target frequency group.

[0005] One or more embodiments of the present specification provide a noise reduction system, the system comprising: a feature determination module, configured to determine the noise frequency feature of the ambient noise; a frequency group determination module, configured to determine a target frequency group based on the noise frequency feature; the target frequency group comprises one or more target frequencies; a signal determination module, configured to: for one target frequency: generate multiple first candidate signals based on multiple candidate phases; obtain a first feedback signal corresponding to the first candidate signal; determine a target phase and an initial target signal corresponding to the target frequency based on the first feedback signal and the noise frequency feature; determine a target single-frequency signal corresponding to the target frequency based on multiple candidate amplitudes and the initial target signal; and a noise reduction completion module, configured to complete noise reduction based on the target single-frequency signal group corresponding to the target frequency group.

[0006] One or more embodiments of the present specification provide a noise reduction device, which includes at least one memory and at least one processor, wherein the at least one memory is used to store computer instructions, and the at least one processor executes the computer instructions or part of the instructions to implement the noise reduction method described in the above embodiments.

[0007] One or more embodiments of the present specification provide a computer-readable storage medium, wherein the storage medium stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the noise reduction method described in the above embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] This specification will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0009] Figure 1 is a schematic diagram of an application scenario of a noise reduction system according to some embodiments of this specification;

[0010] Figure 2 is an exemplary module diagram of a noise reduction system according to some embodiments of this specification;

[0011] Figure 3 is an exemplary flow chart of a noise reduction method according to some embodiments of this specification;

[0012] Figure 4 is an exemplary flow chart of determining a target frequency group according to some embodiments of this specification;

[0013] Figure 5 is an exemplary flow chart of determining a target phase and an initial target signal according to some embodiments of this specification;

[0014] Figure 6 is an exemplary schematic diagram of determining a target single-frequency signal according to some embodiments of this specification. DETAILED DESCRIPTION

[0015] The following is a brief introduction to the drawings required for describing the embodiments. The drawings do not represent all implementation methods.

[0016] The "system", "device", "unit" and / or "module" used herein are a method for distinguishing different components, elements, parts, portions or assemblies at different levels. If other words can achieve the same purpose, the words can be replaced by other expressions.

[0017] When the operations performed in the embodiments of this specification are described in steps, unless otherwise specified, the order of the steps is interchangeable, steps may be omitted, and other steps may be included in the operation process.

[0018] Flowcharts are used in this specification to illustrate the operations performed by the system according to the embodiments of this specification. It should be understood that the preceding or following operations are not necessarily performed precisely in order. Instead, the steps may be processed in reverse order or simultaneously. At the same time, other operations may also be added to these processes, or one or more operations may be removed from these processes.

[0019] In people's daily lives, there have always been problems such as traffic noise and industrial noise. ANC noise reduction technology is a technology that cancels out noise by emitting a waveform opposite to the noise. This technology has a good noise reduction effect on low-frequency continuous noise, but there is a slight delay when processing high-frequency noise, so that the reverse sound wave used for cancellation and the noise sound wave do not completely correspond, so the noise reduction effect is poor.

[0020] The present invention provides a noise reduction method, system, device and storage medium, which avoids the influence of system timeliness on the noise reduction effect by predetermining the target phase corresponding to each sound frequency of the energy body of the ambient noise and playing the corresponding feedback signal, thereby improving the noise reduction effect.

[0021] Figure 1 It is a schematic diagram of an application scenario of a noise reduction system according to some embodiments of this specification.

[0022] The noise reduction system can be applied to a variety of noisy environments. In this specification, the noise reduction process is mainly implemented in scenarios where the environmental noise is single and has strong periodicity. Among them, the scenarios where the environmental noise is single and has strong periodicity can include the noise reduction of factory machinery and equipment, the noise reduction of air conditioner outdoor units or fans, and the noise reduction of aircraft engine noise. The embodiments of this specification can be explained by taking the example of a user wearing headphones on an airplane (the noise reduction system is configured in the headphones) to perform noise reduction.

[0023] In some embodiments, Figure 1 As shown, the application scenario 100 of the noise reduction system may include a sensor 110 , a control circuit 120 , a speaker 130 , a processor 140 , a storage device 150 , and a network 160 .

[0024] The sensor 110 may be a sound sensor, such as a microphone, etc. The sound sensor may capture sound and convert the sound signal into an analog electrical signal. In some embodiments, the sensor 110 may be used to obtain ambient noise and convert the ambient noise from a sound signal into an analog electrical signal.

[0025] The control circuit 120 refers to a circuit for controlling the device. The sensor 110 or the control circuit 120 may include a signal converter (eg, an A / D converter). The signal converter may convert an analog electrical signal into an analog digital signal.

[0026] In some embodiments, after the sensor 110 acquires the ambient noise, it transmits the ambient noise directly or through the network 130 to the control circuit 120. The control circuit 120 can generate inverse noise (in the form of an analog digital signal) of the ambient noise based on the ambient noise (in the form of an analog digital signal).

[0027] Among them, inverse noise refers to a noise signal generated by technical means to eliminate or reduce the interference of environmental noise. When the inverse noise of environmental noise meets the environmental noise, they will interfere with each other, thus producing a canceling effect. For more descriptions of environmental noise, please refer to Figure 3 Related instructions.

[0028] The speaker 130 is a device that converts analog digital signals into sound signals. In some embodiments, the speaker 130 can convert the inverse noise of the ambient noise generated by the control circuit 120 from analog digital signals into sound signals and play the sound signals.

[0029] In some embodiments, the processor 140 may process data and / or information obtained from the sensor 110, the control circuit 120, the storage device 150, the network 160, and / or other components of the noise reduction system application scenario 100. For example, the ambient noise obtained by the sensor 110, the inverse noise generated by the control circuit 120, etc. may be transmitted to the processor 140 via the network 160 and processed by the processor 140.

[0030] In some embodiments, the processor 140 may be a computer, a user console, a single processor, a processor group, etc. The processor group may be centralized or distributed. In some embodiments, the processor 140 may be implemented on a cloud platform. For example, the cloud platform may include one or any combination of a private cloud, a public cloud, a hybrid cloud, etc.

[0031] In some embodiments, the processor 140 may be integrated into the control circuit 120 , or the processor 140 may be communicatively connected to the control circuit 120 via the network 160 .

[0032] The storage device 150 is used to store data, instructions and / or any other information. For example, the storage device can store noise frequency characteristic data obtained by feedback.

[0033] The storage device 150 may include one or more storage components, each of which may be an independent device or a part of another device. In some embodiments, the storage device 150 may include a random access memory (RAM), a read-only memory (ROM), a removable memory, etc. or any combination thereof. In some embodiments, the storage device 150 may be connected to a network 160 to communicate with one or more components in the application scenario 100 of the noise reduction system.

[0034] The network 160 includes any suitable network capable of facilitating information and / or data exchange of the application scenario 100 of the noise reduction system. In some embodiments, one or more components of the application scenario 100 of the noise reduction system (e.g., the sensor 110, the control circuit 120, the processor 140, and the storage device 150, etc.) can exchange information and / or data with one or more components of the application scenario 100 of the noise reduction system through the network 160.

[0035] It should be noted that the application scenario is provided only for the purpose of illustration and is not intended to limit the scope of this specification. For those of ordinary skill in the art, various modifications or changes can be made according to the description of this specification. For example, the application scenario may also include a database. For another example, the application scenario may be implemented on other devices to achieve similar or different functions. However, changes and modifications will not deviate from the scope of this specification.

[0036] Figure 2 is an exemplary module diagram of a noise reduction system according to some embodiments of the present specification.

[0037] In some embodiments, Figure 2 As shown, the noise reduction system 200 may include a feature determination module 210 , a frequency group determination module 220 , a signal determination module 230 , and a noise reduction completion module 240 .

[0038] The feature determination module 210 refers to a module for extracting features of noise. In some embodiments, the feature determination module 210 may be configured to determine a noise frequency feature of the ambient noise.

[0039] The frequency group determination module 220 refers to a module for determining a target frequency group. In some embodiments, the frequency group determination module 220 may be configured to determine a target frequency group based on noise frequency characteristics. In some embodiments, the target frequency group may include one or more target frequencies.

[0040] In some embodiments, the frequency group determination module 220 may be further configured to: determine the noise energy of the initial frequency of the ambient noise based on the noise frequency characteristics; and determine the target frequency group based on the noise energy.

[0041] The signal determination module 230 refers to a module for determining a target single-frequency signal. In some embodiments, the signal determination module 230 can be configured to: generate multiple first candidate signals based on multiple candidate phases for a target frequency; obtain a first feedback signal corresponding to the first candidate signal; determine the target phase and the initial target signal corresponding to the target frequency based on the first feedback signal and the noise frequency characteristics; determine the target single-frequency signal corresponding to the target frequency based on multiple candidate amplitudes and the initial target signal.

[0042] In some embodiments, the signal determination module 230 can be further configured to: determine the type of ambient noise based on noise frequency characteristics; determine multiple noise reduction effects of multiple first candidate signals based on multiple first feedback signals and noise frequency characteristics; determine the target phase corresponding to the target frequency and the initial target signal based on multiple noise reduction effects and / or ambient noise types.

[0043] In some embodiments, the signal determination module 230 can be further configured to: generate multiple second candidate signals based on multiple candidate amplitudes and the initial target signal; obtain a second feedback signal corresponding to the second candidate signal; and determine a target single-frequency signal corresponding to the target frequency based on the second feedback signal.

[0044] The noise reduction completion module 240 refers to a module that completes noise reduction. In some embodiments, the noise reduction completion module 240 may be configured to complete noise reduction based on the target single frequency signal group corresponding to the target frequency group.

[0045] In some embodiments, the modules in the noise reduction system 200 may be partially or fully integrated on the processor 140 .

[0046] For details about the processor 140, please refer to Figure 1 For more information about the determination module 210, the frequency group determination module 220, the signal determination module 230 and the noise reduction completion module 240, please refer to Figure 3-Figure 6 Related description.

[0047] It should be noted that the above description of the noise reduction system 200 and its modules is only for convenience of description and does not limit the present specification to the scope of the embodiments. It is understandable that, after understanding the principle of the system, those skilled in the art may arbitrarily combine the modules or form a subsystem to connect with other modules without deviating from the principle.

[0048] Figure 3 is an exemplary flow chart of a noise reduction method according to some embodiments of this specification. Figure 3 As shown, the process 300 includes the following steps 310 to 330. In some embodiments, the process 300 may be executed by a processor.

[0049] Step 310: determine the noise frequency characteristics of the ambient noise.

[0050] Environmental noise refers to the noise in the surrounding environment. For example, environmental noise can include the sound of airplane engines, birds singing, construction site sounds, and the sound of people talking around.

[0051] In some embodiments, the ambient noise can be acquired by a sensor (e.g., a microphone). For more information about the sensor, see Figure 1 Related description.

[0052] In some embodiments, the environmental noise may be composed of a noise frequency signal, or may be composed of a plurality of noise frequency signals superimposed. The noise frequency refers to the number of times the noise sound wave is repeated per unit time. For example, 50 Hz, 300 Hz, 1000 Hz, etc. The noise frequency signal refers to the analog digital signal (such as a sound wave) of the sound corresponding to the noise frequency. One noise frequency signal corresponds to one noise frequency, that is, the environmental noise may correspond to one or more noise frequencies (i.e., one or more noise frequency signals).

[0053] Noise frequency characteristics refer to the relevant characteristics of environmental noise corresponding to the noise frequency, for example, the noise amplitude, noise energy, and noise sound pressure level of environmental noise at a certain noise frequency.

[0054] In some embodiments, a noise frequency signal corresponds to a noise amplitude, a noise energy, and a noise sound pressure level, forming a corresponding noise frequency feature. When an environmental noise includes multiple noise frequency signals, the noise frequency feature of the environmental noise can be represented by a sequence consisting of multiple noise frequency features corresponding to the multiple noise frequency signals included in the environmental noise.

[0055] Noise amplitude refers to the amplitude of the sound wave. For example, 50 decibels, 80 decibels, etc. Noise amplitude can be used to represent the intensity of the sound. The larger the noise amplitude, the greater the sound intensity.

[0056] Noise energy refers to the energy carried by sound waves when they propagate, and can be used to represent the intensity of sound. -13 W / cm 2 .

[0057] The noise sound pressure level refers to the level of sound pressure. In some embodiments, the noise sound pressure level can be represented by a logarithmic series of sound pressure relative to a reference value (e.g., 20 micropascals). The noise sound pressure level is related to the noise energy. The greater the noise sound pressure level, the greater the noise energy and the greater the sound intensity of the noise.

[0058] In some embodiments, the noise frequency characteristics of the ambient noise can be determined in a variety of ways. For example, the processor can obtain the noise frequency characteristics by performing spectrum analysis on the ambient noise. The spectrum analysis can be implemented by related equipment (such as a spectrum analyzer). For another example, the processor can obtain the noise frequency characteristics by performing frequency domain analysis on the ambient noise. The frequency domain analysis method can include but is not limited to Fourier transform, wavelet transform, fast Fourier transform, etc.

[0059] Step 320: Determine a target frequency group based on the noise frequency characteristics.

[0060] A target frequency group refers to a set of target frequencies. In some embodiments, a target frequency group may include one or more target frequencies.

[0061] The target frequency refers to the noise frequency that is desired to be reduced or eliminated in the noise reduction process.

[0062] In some embodiments, for a target frequency in the ambient noise, if it is desired to be reduced or eliminated during the noise reduction process, a target single-frequency signal corresponding to the target frequency can be generated and superimposed with the noise frequency signal corresponding to the target frequency in the ambient noise to achieve noise reduction. That is, at this time, the target frequency is also the sound frequency corresponding to the target single-frequency signal.

[0063] In some embodiments, the target frequency may refer to a noise frequency used to generate different noise frequency signals to achieve noise reduction. Noise reduction is achieved by transmitting a noise frequency signal corresponding to the target frequency to offset ambient noise.

[0064] For example, when noise reduction is performed, it is desired to reduce or eliminate sound waves within the frequency range of 500 Hz-4000 Hz to which the human ear is sensitive, and the target frequency group may include noise frequencies within the range of 500 Hz-4000 Hz (ie, target frequencies).

[0065] In some embodiments, the processor may determine the target frequency group based on the noise frequency characteristics in a variety of ways, such as determining the noise frequency in the frequency range that the human ear is sensitive to as the target frequency, and determining the set of target frequencies as the target frequency group.

[0066] In some embodiments, the processor may determine the noise energy of the initial frequency of the ambient noise based on the noise frequency characteristics; and determine the target frequency group based on the noise energy.

[0067] The initial frequency refers to the own frequency of the environmental noise, that is, the noise frequency of the environmental noise mentioned above.

[0068] In some embodiments, the processor may directly obtain the noise energy of each initial frequency based on the noise frequency characteristics.

[0069] In some embodiments, the processor may determine the target frequency group based on the noise energy in a variety of ways. For example, the processor may sort the noise energies of multiple initial frequencies from large to small, and use the initial frequencies corresponding to the top noise energy rankings (such as the top 10) as the target frequencies, thereby determining the target frequency group.

[0070] In some embodiments, the processor may obtain noise characteristic parameters based on the noise frequency characteristics; determine an energy threshold based on the noise characteristic parameters; and determine a target frequency group based on the energy threshold and the noise energy. For more information about this section, please refer to Figure 4 Related description.

[0071] In some embodiments of the present specification, by determining the target frequency group of ambient noise through noise energy, noise of specific frequencies with larger noise energy in the ambient noise can be further eliminated in a targeted manner, thereby achieving stronger suppression of the energy body of the ambient noise, which is beneficial to enhancing the noise reduction effect.

[0072] In some embodiments, for a target frequency, the processor may execute step 330 , which includes steps 331 - 333 .

[0073] Step 331: Generate multiple first candidate signals based on multiple candidate phases.

[0074] The candidate phase refers to a phase for a candidate target phase. In some embodiments, the candidate phase may refer to an initial position of the first candidate signal within one cycle (0°-360°).

[0075] The first candidate signal refers to a candidate single frequency signal for reducing or canceling environmental noise. In some embodiments, the first candidate signal may be an analog digital signal including a sound frequency, which may be used to reduce or cancel a noise frequency signal of the same sound frequency in environmental noise.

[0076] In some embodiments, a target frequency may correspond to multiple candidate phases, and a candidate phase corresponds to a first candidate signal, then a target frequency corresponds to multiple first candidate signals. One or more first candidate signals corresponding to a target frequency have the same or similar frequency, the same or similar amplitude, and a different phase as the noise frequency signal that needs to be denoised. Among them, frequency similarity may refer to the absolute value of the difference between the frequency of the noise frequency signal and the frequency of the first candidate signal being less than a preset difference threshold. The preset difference threshold may be a very small value set by the processor by default or pre-set by a technician, that is, at this time, frequency similarity can be regarded as frequency similarity. The same applies to amplitude similarity.

[0077] In some embodiments, the processor may determine the candidate phases in a variety of ways. For example, the processor may randomly select multiple phases as multiple candidate phases within a cycle (0°-360°). For another example, the processor may randomly select a phase as an initial candidate phase within a cycle, and translate the initial candidate phase based on a first preset step size and a preset direction to obtain multiple candidate phases. Among them, the first preset step size may be a fixed step size (such as 5°) set by the processor or a technician, or a random step size (such as 3°, 5°, etc.) randomly generated by the processor. The preset direction may include left and right, and the translation range is (0°, 360°).

[0078] In some embodiments of the present specification, by randomly selecting a phase for translation to obtain multiple candidate phases, the computational complexity can be reduced; at the same time, by replacing the opposite phase method in the ANC noise reduction technology with the translation method, the influence of the system timeliness on the noise reduction effect is avoided, which is conducive to improving the noise reduction effect.

[0079] In some embodiments, for a target frequency, the processor may generate a first candidate signal having the same or similar frequency, the same or similar amplitude, and a different phase as the noise frequency signal corresponding to the target frequency based on one or more candidate phases.

[0080] Step 332: Acquire a first feedback signal corresponding to the first candidate signal.

[0081] The first feedback signal refers to a signal obtained after noise reduction processing using the first candidate signal. In some embodiments, for a target frequency, one or more first candidate signals correspond to one or more first feedback signals, and a first feedback signal can be formed by superimposing a noise frequency signal of the target frequency and a corresponding first candidate signal.

[0082] In some embodiments, the processor may obtain the first feedback signal corresponding to the first candidate signal in a variety of ways. For example, the processor may obtain the sound of the first candidate signal mixed with the ambient noise after the first candidate signal is sent through a sensor (such as a sound sensor), and determine the first feedback signal through signal conversion. For another example, the processor may analyze and process the first candidate signal and the noise frequency signal (such as MATLAB visualization processing) to determine the first feedback signal.

[0083] Step 333: Determine a target phase corresponding to the target frequency and an initial target signal based on the first feedback signal and the noise frequency characteristics.

[0084] The target phase refers to the phase corresponding to the first candidate signal that can effectively reduce noise. One target frequency corresponds to one target phase.

[0085] The initial target signal refers to a single-frequency signal that is preliminarily determined to reduce or offset environmental noise. One target phase corresponds to one initial target signal. That is, the initial target signal is the first candidate signal corresponding to the target phase, and the initial target signal has the same or similar frequency, the same or similar amplitude, and a different phase (the phase is the target phase) as the noise frequency signal to be reduced.

[0086] In some embodiments, the processor can determine the target phase and the initial target signal corresponding to the target frequency in a variety of ways based on the feedback signal and the noise frequency characteristics. For example, the processor can perform spectrum analysis on one or more first feedback signals corresponding to one or more first candidate signals, and determine the first candidate signal and candidate phase corresponding to the feedback signal with the sound amplitude closest to 0 as the initial target signal and target phase corresponding to the target frequency.

[0087] In some embodiments, the processor may determine the type of ambient noise based on the noise frequency characteristics; determine multiple noise reduction effects of multiple first candidate signals based on multiple first feedback signals and noise frequency characteristics; determine the target phase corresponding to the target frequency and the initial target signal based on multiple noise reduction effects and / or ambient noise types. For more information about this section, please refer to Figure 6 Related description.

[0088] Step 334: Determine a target single frequency signal corresponding to the target frequency based on the multiple candidate amplitudes and the initial target signal.

[0089] The candidate amplitude refers to an amplitude candidate value used to determine the target single frequency signal.

[0090] In some embodiments, the processor may determine the candidate amplitudes in a variety of ways. For example, the processor may randomly select multiple amplitudes within a preset amplitude range as multiple candidate amplitudes. The preset amplitude range may be set by the processor or a technician. For example, 30 decibels to 100 decibels.

[0091] For another example, for a target frequency, the processor can increase the amplitude of the initial target signal according to the second preset step size to obtain multiple candidate amplitudes. The amplitude of the initial target signal is the amplitude of the noise frequency signal corresponding to the target frequency. The second preset step size can be a fixed step size (such as 2 decibels) set by the processor or a technician, or a random step size (such as 2 decibels, 5 decibels, etc.) randomly generated by the processor.

[0092] The target single frequency signal refers to the single frequency signal finally determined to reduce or cancel the ambient noise. In some embodiments, the target single frequency signal may be an analog digital signal including a sound frequency, which may be used to reduce or cancel the noise frequency signal of the same sound frequency in the ambient noise.

[0093] In some embodiments, the phase of the initial target signal (i.e., the target phase) is the phase with the best noise reduction effect, that is, the target phase is the phase of the target single-frequency signal; and the amplitude of the initial target signal may not be the amplitude corresponding to the best noise reduction effect, that is, the amplitude of the target single-frequency signal has not yet been determined; therefore, it is necessary to further determine the amplitude corresponding to the target single-frequency signal.

[0094] That is, the target single-frequency signal has the same or different amplitude, the same or similar frequency, and the same phase as the initial target signal; the target single-frequency signal has the same or similar frequency, different phase (the phase is the target phase), and the same or different amplitude as the noise frequency signal corresponding to the target frequency.

[0095] In some embodiments, the processor may generate multiple second candidate signals based on multiple candidate amplitudes and the initial target signal; obtain a second feedback signal corresponding to the second candidate signal; and determine a target single frequency signal corresponding to the target frequency based on the second feedback signal. For more information about this part, please refer to Figure 6 Related description.

[0096] In actual situations, the noise reduction of environmental noise is achieved by superimposing target single-frequency signals corresponding to all target frequencies of the environmental noise. Therefore, after determining the target phase and target single-frequency signal corresponding to a target frequency, when determining the target phase and target single-frequency signal corresponding to the next target frequency, the target single-frequency signal of the previous target frequency can be superimposed and output with multiple first candidate signals and multiple second candidate signals of the next target frequency to determine multiple first feedback signals and multiple second feedback signals.

[0097] The superposition process is implemented based on the order of the target frequencies from low frequency to high frequency. In some embodiments, the processor can sort one or more target frequencies in the target frequency group from small to large, and then determine the target phase and target single-frequency signal corresponding to the target frequency in sequence according to the sorting (from low frequency to high frequency).

[0098] For example, the target frequency group includes two target frequencies of 500 Hz and 600 Hz. The target phase and target single-frequency signal corresponding to the first target frequency (i.e., 500 Hz) should be determined first. When determining the target phase and target single-frequency signal corresponding to the second target frequency (i.e., 600 Hz), the target single-frequency signal of the first target frequency (i.e., 500 Hz) and multiple first candidate signals of the second target frequency (i.e., 600 Hz) can be superimposed and output one by one to determine multiple first feedback signals corresponding to the second target frequency.

[0099] Step 340: completing noise reduction based on the target single-frequency signal group corresponding to the target frequency group.

[0100] A target single-frequency signal group refers to a set of one or more target single-frequency signals corresponding to one or more target frequencies. One target frequency corresponds to one target single-frequency signal, and one target frequency group corresponds to one target single-frequency signal group.

[0101] In some embodiments, the processor can output all target single-frequency signals in the target single-frequency signal group, and based on the speaker, all target single-frequency signals are converted into sound signals and broadcast, which meet and cancel the ambient noise to achieve noise reduction. For more information about the speaker, please refer to Figure 1 Related description.

[0102] In some embodiments of the present specification, a target frequency group is determined by the noise frequency characteristics of the ambient noise, and a target single-frequency signal group corresponding to the target frequency group is emitted to perform noise reduction. Noise reduction can be performed based on the characteristics of the ambient noise. For scenarios where the ambient noise is single and has a strong periodicity, a single-frequency signal can be used for distributed single-point noise reduction to improve the noise reduction effect.

[0103] Figure 4 FIG. 1 is an exemplary flow chart of determining a target frequency group according to some embodiments of this specification. Figure 4 As shown, the process 400 includes the following steps 410 to 430. In some embodiments, the process 400 may be executed by a processor.

[0104] Step 410: Acquire noise characteristic parameters based on noise frequency characteristics.

[0105] Noise characteristic parameters refer to parameter information related to noise frequency characteristics. For example, noise characteristic parameters may include noise frequency distribution, noise sound pressure level range, noise amplitude distribution, etc.

[0106] The noise frequency distribution refers to the distribution of the noise frequency. For example, the noise frequency in the range of 500 Hz-4000 Hz accounts for 80%, the noise frequency in other ranges accounts for 20%, and the average noise frequency is 2000 Hz.

[0107] The noise sound pressure level range refers to the numerical range of the noise sound pressure level in the noise frequency characteristics. For example, the noise sound pressure level range is (SPL min , SPL max ), SPL min 、SPL max They respectively refer to the minimum and maximum values ​​of the noise sound pressure level in the noise frequency characteristics.

[0108] The noise amplitude distribution refers to the distribution of the noise amplitude. For example, the noise amplitude distribution may include a mean of the noise amplitude of

[0109] In some embodiments, the processor may perform statistical analysis on the noise frequency characteristics to obtain noise characteristic parameters.

[0110] Step 420: Determine an energy threshold based on the noise characteristic parameter.

[0111] The energy threshold refers to a preset threshold for screening noise energy. In some embodiments, the processor may determine the initial frequency at which the noise energy exceeds the energy threshold as the target frequency, thereby screening out the initial frequency corresponding to the noise frequency signal of the energy subject in the ambient noise, avoiding increasing the processing load or introducing unnecessary feedback due to processing all frequencies.

[0112] In some embodiments, the processor may determine the energy threshold in a variety of ways based on the noise characteristic parameters. For example, the processor may determine the energy threshold based on the noise amplitude distribution and the noise sound pressure level range in the noise characteristic parameters. Exemplarily, the larger the noise amplitude mean value and the larger the noise sound pressure level maximum value, the larger the energy threshold.

[0113] For another example, the processor may determine the energy threshold by searching a database based on noise characteristic parameters, power information, and user motion information.

[0114] As an example only, the processor may construct a first database. The first database may include multiple first samples and their corresponding multiple first labels. Among them, the first sample may be a first sample vector composed of sample noise characteristic parameters, sample power information, and sample user motion information, and the first label may be a sample energy threshold corresponding to the first sample when the noise reduction effect meets the requirements. Among them, the first sample and its corresponding first label may be determined based on historical data. Meeting the requirements may refer to meeting the user's auditory needs, reducing the ambient noise to a preset value, etc.

[0115] For example, a group of first samples may be historical noise characteristic parameters of a historical environmental noise, historical power information of a historical sound device, and historical user movement information in a historical noise reduction process. The processor may determine the historical energy threshold corresponding to the final noise reduction effect that meets the requirements in the historical noise reduction process as the sample energy threshold.

[0116] In some embodiments, the processor may construct a first vector to be matched based on the current noise characteristic parameters, power information, and user motion information, search the first database, calculate multiple first similarities between the first vector to be matched and multiple first sample vectors, and determine the first label corresponding to the first sample vector with the highest first similarity as the current energy threshold. The first similarity may be represented by cosine similarity, Euclidean distance, etc.

[0117] Among them, power information refers to relevant information reflecting the power usage of the sound device (such as headphones, etc.). For example, the remaining power percentage, power consumption rate, etc. In some embodiments, the processor can obtain power information by reading the battery information of the device. The remaining power percentage refers to the ratio of the current remaining power of the battery to the maximum capacity of the battery. The power consumption rate refers to the power consumed by the battery in a unit of time.

[0118] A user refers to a person wearing a sound device (such as headphones, etc.). User motion information refers to relevant information reflecting the user's motion situation when using the sound device. In some embodiments, the user motion information may include the user's motion speed and the user's motion type. The user's motion speed refers to the speed of the user's motion. The user's motion type refers to the user's motion state. For example, walking state, running state, stationary state, etc. In some embodiments, the user motion information can be monitored and obtained by the processor through related devices built into the sound device (such as motion sensors), or can be obtained by the user by inputting it himself.

[0119] In determining the energy threshold, considering the power information can avoid excessive power consumption caused by over-processing of signals, which is beneficial to improving user experience and extending the use time of audio equipment. At the same time, since the auditory sensitivity is lower when exercising than when stationary, considering the user's motion information can select a larger energy threshold (that is, select fewer target frequencies for processing) when the user is exercising, so as to ensure the user's auditory experience while reducing power consumption.

[0120] Step 430: determine the target frequency group based on the energy threshold and the noise energy.

[0121] In some embodiments, the processor can determine the target frequency group in a variety of ways based on the energy threshold and the noise energy. For example, the processor can use all initial frequencies corresponding to the noise energy greater than the energy threshold as the target frequency, and determine the set of target frequencies as the target frequency group. For another example, the processor can use the initial frequency whose noise energy is greater than the energy threshold and whose noise energy ranks high (such as the top 10) as the target frequency, and determine the set of target frequencies as the target frequency group.

[0122] In some embodiments of the present specification, by reasonably determining the energy threshold through noise characteristic parameters, system performance, noise reduction effect and user experience can be taken into account at the same time to improve user satisfaction.

[0123] In some embodiments, the processor may also obtain noise reduction frequency characteristics after completing a round of noise reduction; dynamically adjust the energy threshold based on the noise reduction frequency characteristics and the noise frequency characteristics, and determine to update the energy threshold.

[0124] Among them, a round of noise reduction refers to executing Figure 3 A complete process of the process 300 shown is: determining the noise frequency characteristics of the ambient noise; determining the target frequency group based on the noise frequency characteristics; and completing a round of noise reduction based on the target single-frequency signal group corresponding to the target frequency group.

[0125] One round of noise reduction corresponds to a target frequency group and an energy threshold. After completing one round of noise reduction, the processor can determine whether the energy threshold needs to be dynamically adjusted, and then the next round of noise reduction can be performed according to the updated energy threshold (determining a new target frequency group).

[0126] The noise reduction frequency feature refers to the frequency feature of the second feedback signal obtained after completing a round of noise reduction. In some embodiments, one second feedback signal corresponds to one noise reduction frequency feature, and one noise reduction frequency feature is composed of a noise amplitude after noise reduction, a noise energy after noise reduction, and a noise sound pressure level after noise reduction. For more information about the second feedback signal, please refer to Figure 6 Related description.

[0127] In some embodiments, the processor can obtain the noise reduction frequency characteristics by performing spectrum analysis or frequency domain analysis on the second feedback signal. Figure 3 The description of the frequency characteristics of the noise will not be repeated here.

[0128] The updated energy threshold refers to the energy threshold after the energy threshold is adjusted. In some embodiments, the processor may re-determine the target frequency group based on the updated energy threshold.

[0129] In some embodiments, the processor may determine a noise reduction score based on the noise reduction frequency characteristics and the noise frequency characteristics; and adjust the energy threshold based on the noise reduction score to determine an updated energy threshold.

[0130] The noise reduction score refers to the score of the noise reduction effect. In some embodiments, the processor can determine the sound pressure drop corresponding to each noise frequency based on the noise reduction frequency characteristics and the noise frequency characteristics. Since non-target frequencies may be affected when using the target frequency for noise reduction, the sound pressure drop corresponding to the target frequency and the sound pressure drop corresponding to the non-target frequency can be weighted and averaged to obtain the mean sound pressure drop, and the mean sound pressure drop is used as the noise reduction score. The sound pressure drop refers to the reduction in the noise sound pressure level after noise reduction in the noise reduction frequency characteristics compared to the noise sound pressure level in the noise frequency characteristics.

[0131] The weights may be determined by a technician or a processor in a variety of ways. For example, a higher weight may be set for the sound pressure drop corresponding to the target frequency. For another example, a higher weight may be set for the sound pressure drop corresponding to the noise frequency within the frequency range of 500 Hz-4000 Hz to which the human ear is sensitive.

[0132] In some embodiments, the processor may increase the energy threshold in response to the noise reduction score being greater than a first score threshold; decrease the energy threshold in response to the noise reduction score being less than a second score threshold; or not adjust the energy threshold in response to the noise reduction score being between the first score threshold and the second score threshold.

[0133] Among them, the first scoring threshold is greater than the second scoring threshold. The first scoring threshold and the second scoring threshold can be set by the system default or pre-set by the user. If the noise reduction score is greater than the first scoring threshold, it means that the noise reduction score is high, and the corresponding noise reduction effect has met the noise reduction requirements. The energy threshold can be appropriately increased (that is, fewer target frequencies are selected for processing in the next round) to reduce the amount of data that needs to be processed later, and the larger the noise reduction score, the greater the increase in the energy threshold; if the noise reduction score is less than the second scoring threshold, it means that the noise reduction score is low, and the corresponding noise reduction effect does not meet the noise reduction requirements. The energy threshold can be appropriately reduced (that is, more target frequencies are selected for processing in the next round) until the ideal noise reduction effect is achieved, and the smaller the noise reduction score, the greater the reduction in the energy threshold.

[0134] In some embodiments of the present specification, by dynamically adjusting the energy threshold, the main noise frequency can be more accurately identified and processed, thereby improving the noise reduction effect and avoiding the waste of operating resources.

[0135] In some embodiments, the processor may also obtain playback frequency characteristics based on the playback audio; dynamically adjust the energy threshold based on the playback frequency characteristics and the noise frequency characteristics, and determine to update the energy threshold.

[0136] Playing audio refers to the audio played in the sound device, for example, music played in headphones, etc. In some embodiments, playing audio may refer to the converted analog-to-digital signal of the playing audio.

[0137] The playback frequency feature refers to the frequency feature of the playback audio. The content of the playback frequency feature is similar to the noise frequency feature and is not described in detail. In some embodiments, an audio amplitude, an audio energy, and an audio sound pressure level corresponding to a playback audio constitute a corresponding playback frequency feature. In some embodiments, the processor can obtain the playback frequency feature by performing spectrum analysis or frequency domain analysis on the playback audio. For more information about the playback frequency feature, see Figure 3 The description of the frequency characteristics of the noise will not be repeated here.

[0138] In some embodiments, the processor can determine the update energy threshold in a variety of ways based on the playback frequency characteristics and the noise frequency characteristics. For example, the processor can calculate the similarity between the playback frequency characteristics and the noise frequency characteristics. The greater the similarity, the greater the update energy threshold. For another example, the processor can match and compare the playback frequency characteristics and the noise frequency characteristics. The greater the proportion of the same frequencies in the two (the ratio of the number of the same frequencies to the total number of frequencies), the greater the update energy threshold, so as to avoid more noise reduction processing on the noise frequencies close to the frequency of the playback audio during the noise reduction process, so as to avoid a greater impact on the playback audio.

[0139] In some embodiments, the processor may also determine the frequencies to be deleted in the target frequency group based on the playback frequency characteristics.

[0140] The frequency to be deleted refers to the target frequency that needs to be deleted in the target frequency group. Since some frequencies of the played audio may be consistent with some noise frequencies of the ambient noise, if the target frequency group is directly subjected to noise reduction processing, it may cause erroneous noise reduction of some frequencies of the played audio. Therefore, it is necessary to identify the target frequencies that need to be deleted in the target frequency group to reduce the interference of the noise reduction processing on the played audio.

[0141] In some embodiments, the processor may compare the frequencies in the playback frequency feature with the target frequencies in the current target frequency group one by one to determine whether there is a target frequency in the target frequency group that overlaps with the frequency of the playback frequency feature. If so, the target frequency is determined as the frequency to be deleted.

[0142] In some embodiments of the present specification, by confirming an energy threshold or a target frequency group based on the playback frequency characteristics of the played audio, it is possible to avoid the noise reduction system from performing noise reduction processing on the key frequencies of the played audio, reduce the interference of the noise reduction processing on the played audio, and thus maintain the high fidelity and clarity of the played audio.

[0143] Figure 5 FIG. 1 is an exemplary flow chart of determining a target phase and an initial target signal according to some embodiments of this specification. Figure 5 As shown, the process 500 includes the following steps 510 to 530. In some embodiments, the process 500 may be executed by a processor.

[0144] Step 510: Determine the type of environmental noise based on the noise frequency characteristics.

[0145] The environmental noise type refers to the classification type of environmental noise. For example, the environmental noise type may include traffic noise, human voice, industrial noise, and natural noise.

[0146] In some embodiments, the processor can determine the type of environmental noise based on the noise frequency characteristics in a variety of ways. For example, the processor can determine the type of environmental noise by searching a database.

[0147] As an example only, the processor may construct a second database. The second database may include a plurality of second samples and a plurality of second labels corresponding thereto. The second sample may be a sample noise frequency feature of the sample environmental noise, and the second label may be a corresponding sample environmental noise type.

[0148] In some embodiments, the processor may obtain multiple sample environmental noises and their corresponding sample noise frequency characteristics and sample environmental noise types. For a sample environmental noise, the processor may determine the noise frequency characteristics of the preprocessed sample environmental noise, use it as a second sample, and use the environmental noise type of the sample environmental noise as the corresponding second label.

[0149] In some embodiments, the sample environmental noise may be collected manually, and its corresponding sample environmental noise type may be manually labeled. The sample environmental noise may also be historical environmental noise in historical data, and the historical environmental noise type determined historically is used as the sample environmental noise type.

[0150] In some embodiments, the multiple sample environmental noises (i.e., multiple second samples) used to construct the second database should include multiple sample environmental noise types and cover various possible environmental conditions to ensure the data richness of the second database. The environmental conditions may refer to the collection environment of the environmental noise, such as roads, factories, etc.

[0151] The preprocessing process may include normalization to improve the sample quality. Normalization refers to unifying the sample environment noise to facilitate subsequent analysis and processing.

[0152] In some embodiments, the method for determining the noise frequency characteristics of the preprocessed sample environmental noise can be found in Figure 3 The method of determining the noise frequency characteristics of the ambient noise in step 310 is not described in detail here.

[0153] In some embodiments, the processor may search in the second database based on the noise frequency feature of the current environmental noise, calculate multiple second similarities between the noise frequency feature of the current environmental noise and multiple second samples, and determine the second label corresponding to the second sample with the highest second similarity as the environmental noise type of the current environmental noise. The second similarity may be represented by cosine similarity, Euclidean distance, etc.

[0154] For more information on ambient noise and noise frequency characteristics, see Figure 3 Related description.

[0155] Step 520: Determine multiple noise reduction effects of multiple first candidate signals based on the multiple first feedback signals and noise frequency characteristics.

[0156] The noise reduction effect refers to the effect of reducing the ambient noise. Noise reduction can improve the purity of the sound, improve the clarity of the sound, etc. In some embodiments, the noise reduction effect can be represented by a noise reduction score, and the higher the noise reduction score, the better the noise reduction effect.

[0157] In some embodiments, a target frequency may correspond to one or more first candidate signals, a first candidate signal may correspond to a first feedback signal, and a first feedback signal may correspond to a noise reduction effect of a first candidate signal.

[0158] In some embodiments, the processor may determine a noise reduction effect of a corresponding first candidate signal in a variety of ways based on a first feedback signal and noise frequency characteristics.

[0159] For example, for a first feedback signal, the processor can determine the sound pressure level of the sound corresponding to the first feedback signal through spectrum analysis, recorded as the first sound pressure level; determine the noise sound pressure level of the corresponding target frequency in the ambient noise based on the noise frequency characteristics, recorded as the second sound pressure level; based on the first sound pressure level and the second sound pressure level, determine the noise reduction effect of a corresponding first candidate signal.

[0160] Exemplarily, when the first sound pressure level is less than the second sound pressure level, it means that the sound pressure level is reduced after the first candidate signal is superimposed, which means that the first candidate signal has a counteracting effect on the noise frequency signal, that is, superimposing the first candidate signal has a noise reduction effect.

[0161] When the first sound pressure level is less than the second sound pressure level, the greater the numerical difference between the second sound pressure level and the first sound pressure level, the better the noise reduction effect. At this time, the noise reduction effect of the first feedback signal corresponding to the target frequency can be represented by the difference between the second sound pressure level and the first sound pressure level.

[0162] For another example, for a first feedback signal, the processor can determine the sound energy corresponding to the first feedback signal and the noise energy of the corresponding target frequency in the ambient noise, and then determine the noise reduction effect of the corresponding first candidate signal. This part is similar to the process of determining the noise reduction effect based on the first sound pressure level and the second sound pressure level, and will not be repeated here.

[0163] For more information about the first feedback signal and the first candidate signal, see Figure 3 Related description.

[0164] Step 530: Determine a target phase corresponding to a target frequency and an initial target signal based on multiple noise reduction effects and / or environmental noise types.

[0165] For more information on target frequency, target phase, and initial target signal, see Figure 3 and Figure 4 Related description.

[0166] In some embodiments, for a target frequency, the processor may determine the target phase and the initial target signal in multiple ways based on multiple noise reduction effects of multiple first candidate signals and / or types of environmental noise.

[0167] For example, the processor may determine one or more first candidate signals whose noise reduction effect is greater than a preset threshold, and then determine the corresponding one or more candidate phases, and randomly select one of them as the target phase. The preset threshold may be related to the type of environmental noise. Exemplarily, the processor may determine the preset threshold by querying a preset relationship table based on the type of environmental noise.

[0168] Among them, the preset relationship table includes environmental noise types and corresponding preset thresholds. The preset relationship table can be constructed based on historical data. For example, for a historical environmental noise type in the historical data, the processor can count one or more corresponding historical preset thresholds. The historical preset threshold corresponding to the highest number of occurrences and the corresponding noise reduction effect that meets the requirements is determined as the historical preset threshold corresponding to the historical environmental noise type, and multiple historical environmental noise types are traversed to construct a preset relationship table.

[0169] For another example, the processor may determine the candidate phase corresponding to the first candidate signal with the best noise reduction effect as the target phase.

[0170] In some embodiments, the processor can also obtain environmental characteristic data and user wearing characteristics; based on the target frequency characteristics, environmental noise type, environmental characteristic data and user wearing characteristics, determine the target phase corresponding to the target frequency and the initial target signal through a phase prediction model.

[0171] Environmental characteristic data refers to data related to the environment in which the user is located. For example, environmental characteristic data may include ambient temperature, ambient humidity, etc. The environment in which the user is located may refer to the physical environment in which the user is located when wearing the audio device.

[0172] In some embodiments, the processor can obtain environmental characteristic data in a variety of ways. For example, the processor can obtain environmental characteristic data based on a built-in sensor device in the sound device (such as a temperature sensor and a humidity sensor in the earphone). The processor can also obtain environmental characteristic data through a third platform (such as a weather forecast APP).

[0173] User wearing characteristics refer to characteristic data related to the user wearing the sound device. For example, the user wearing characteristics may include the user's ear canal shape, ear pressure, and current wearing pressure, etc. Among them, the user's ear canal shape and ear pressure can be obtained by measuring the user in advance.

[0174] The current wearing pressure refers to the pressure on the ear caused by the current wearing of the earphone. In some embodiments, the current wearing pressure can be obtained based on a micro pressure sensor inside the earphone. The micro pressure sensor can contact the ear canal or the auricle to directly measure the pressure on the ear.

[0175] The target frequency characteristics refer to the characteristics related to the target frequency. For example, the target frequency characteristics may include the sound amplitude, sound energy, sound pressure level, initial phase, etc. of the target frequency. For more information about sound amplitude, sound energy, sound pressure level, and initial phase, please refer to Figure 3 Related descriptions of noise amplitude, noise energy, noise sound pressure level and initial phase.

[0176] In some embodiments, the processor can analyze the ambient noise of the target frequency based on fast Fourier transform and spectrum analysis to obtain the target frequency characteristics. Figure 3 The description of obtaining noise frequency characteristics in is not repeated here.

[0177] The phase prediction model refers to a model for determining the target phase of the target frequency. In some embodiments, the phase prediction model may be a machine learning model. For example, the phase prediction model may include convolutional neural networks (CNN), deep neural networks (DNN), and recurrent neural networks (RNN).

[0178] In some embodiments, the input of the phase prediction model may include target frequency characteristics, ambient noise type, environmental feature data, user wearing characteristics, current time point and user motion information, and the output may be a target phase corresponding to the target frequency.

[0179] For more information about user motion information, see Figure 4 Related description.

[0180] In some embodiments, the processor may train the phase prediction model in a variety of ways. For example, a set of training samples used for model training may include sample target frequency features, sample environmental noise types, sample environmental feature data, sample user wearing features, sample time points, sample user motion information, etc. The training labels corresponding to a set of training samples may be the target phases corresponding to the sample target frequencies in the set of training samples.

[0181] In some embodiments, for a sample target frequency, the processor can generate multiple sample candidate phases, and generate multiple corresponding sample first candidate signals based on the multiple sample candidate phases, perform multiple noise reduction experiments based on the multiple sample first candidate signals, determine the experimental noise reduction effects corresponding to the multiple sample first candidate signals, and select the candidate phase with the best noise reduction effect as the sample target phase corresponding to the sample target frequency, that is, the training label corresponding to the training sample.

[0182] The candidate phase with the best noise reduction effect can be selected and determined by a technician based on the sense of hearing, or the processor can determine the noise reduction effect based on the first feedback signal and the noise frequency characteristics, and then select the candidate phase with the best noise reduction effect. For determining the noise reduction effect based on the first feedback signal and the noise frequency characteristics, please refer to the relevant description of step 520, which will not be repeated here.

[0183] In some embodiments, the processor may input sample target frequency characteristics, sample environmental noise types, sample environmental feature data, sample user wearing characteristics, sample time points, sample user motion information, etc. into the initial phase prediction model, construct a loss function based on the target phase output by the initial phase prediction model and the training label, iteratively update the initial phase prediction model based on the loss function, and complete the training when the preset conditions are met to obtain a trained phase prediction model. The preset conditions may be that the loss function converges, the number of iterations reaches a threshold, etc.

[0184] In some embodiments, the processor may output a target phase corresponding to a target frequency based on a phase prediction model, and further determine a corresponding initial target signal.

[0185] In some embodiments of this specification, the machine learning model (phase prediction model) can significantly reduce the time required to determine the target phase compared to the traditional method, thereby speeding up the noise reduction process, improving the noise reduction efficiency, and enabling the device to adapt to environmental noise more quickly. In addition, the user's personalized data such as the user's wearing characteristics is used as the input of the phase prediction model, fully considering the user's personalized needs, and realizing the personalized customization of noise reduction, so as to ensure that each user can experience the best noise reduction effect.

[0186] In some embodiments of the present specification, by comparing the first feedback signal and the noise frequency characteristics, the noise reduction effect of the first candidate signal can be accurately evaluated, and then according to the noise reduction effect and / or the type of environmental noise, the target phase and the initial target signal corresponding to the target frequency can be accurately determined, which is beneficial to improving the noise reduction level and noise reduction efficiency.

[0187] It should be noted that the above descriptions of process 300, process 400 and process 500 are only for example and explanation, and do not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to process 300, process 400 and process 500 under the guidance of this specification. However, these modifications and changes are still within the scope of this specification.

[0188] Figure 6 is an exemplary schematic diagram of determining a target single-frequency signal according to some embodiments of this specification.

[0189] In some embodiments, Figure 6 As shown, the processor can generate multiple second candidate signals 630 based on multiple candidate amplitudes 610 and an initial target signal 620; obtain a second feedback signal 640 corresponding to the second candidate signal 630; and determine a target single-frequency signal 650 corresponding to the target frequency based on the second feedback signal 640.

[0190] For more information about candidate amplitudes, initial target signals, target frequencies, and target single-frequency signals, see Figure 3 Related description.

[0191] The second candidate signal refers to a candidate single-frequency signal for reducing or offsetting environmental noise. In some embodiments, the processor may assign multiple candidate amplitudes to the initial target signal to generate multiple second candidate signals. One initial target signal corresponds to multiple candidate amplitudes, and one candidate amplitude corresponds to one second candidate signal. That is, for a target frequency, its corresponding initial target signal corresponds to multiple second candidate signals, and the initial target signal has the same or similar frequency as the second candidate signal, the same phase, and different amplitudes.

[0192] The second feedback signal refers to a signal obtained after noise reduction processing using the second candidate signal. In some embodiments, for a target frequency, multiple second candidate signals correspond to multiple second feedback signals, and a second feedback signal can be formed by superposition of environmental noise and a corresponding second candidate signal.

[0193] In some embodiments, the processor may obtain the second feedback signal corresponding to the second candidate signal in a variety of ways. For example, the processor may obtain the sound of the second candidate signal mixed with the ambient noise after the second candidate signal is sent through a sensor (such as a sound sensor), and determine the second feedback signal through signal conversion.

[0194] In some embodiments, the processor may determine the target single-frequency signal corresponding to the target frequency in a variety of ways based on the second feedback signal. For example, for a target frequency, the processor may perform spectrum analysis on multiple second feedback signals corresponding to multiple second candidate signals, and determine the second candidate signal corresponding to the second feedback signal with the sound amplitude closest to 0 as the target single-frequency signal corresponding to the target frequency.

[0195] In some embodiments of the present specification, the amplitude of the initial target signal is changed to generate multiple second candidate signals, thereby determining a target single-frequency signal with the best noise reduction effect (the sound amplitude after noise reduction is closest to 0).

[0196] One or more embodiments of the present specification provide a noise reduction device, which includes at least one memory and at least one processor, wherein the at least one memory is used to store computer instructions, and the at least one processor executes the computer instructions or part of the instructions to implement a noise reduction method.

[0197] One or more embodiments of the present specification provide a computer-readable storage medium, wherein the storage medium stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes a noise reduction method.

[0198] Furthermore, certain features, structures or characteristics in one or more embodiments of this specification may be appropriately combined.

[0199] In some embodiments, numbers describing the number of components and attributes are used. It should be understood that such numbers used in the description of the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Unless otherwise specified, "about", "approximately" or "substantially" indicate that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may change according to the required features of individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and adopt the general method of retaining digits. Although the numerical domains and parameters used to confirm the breadth of their range in some embodiments of this specification are approximate values, in specific embodiments, the setting of such numerical values ​​is as accurate as possible within the feasible range.

[0200] If there is any inconsistency or conflict between the descriptions, definitions, and / or usage of terms in the materials cited in this specification and the contents described in this specification, the descriptions, definitions, and / or usage of terms in this specification shall prevail.

Claims

1. A noise reduction method, characterized in that: The method comprises: Determine the noise frequency characteristics of the ambient noise; Based on the noise frequency characteristics, determining a target frequency group; the target frequency group includes one or more target frequencies; For a target frequency: Based on the plurality of candidate phases, generating a plurality of first candidate signals; Acquire a first feedback signal corresponding to the first candidate signal; Determining a target phase and an initial target signal corresponding to the target frequency based on the first feedback signal and the noise frequency characteristic; Determining a target single-frequency signal corresponding to the target frequency based on a plurality of candidate amplitudes and the initial target signal; Noise reduction is performed based on the target single-frequency signal group corresponding to the target frequency group.

2. The method according to claim 1, characterized in that The step of determining a target frequency group based on the noise frequency characteristics comprises: Based on the noise frequency characteristics, determining the noise energy of the initial frequency of the ambient noise; Based on the noise energy, the target frequency group is determined.

3. The method according to claim 1, characterized in that The determining, based on the first feedback signal and the noise frequency characteristic, a target phase and an initial target signal corresponding to the target frequency includes: Determining the type of environmental noise based on the noise frequency characteristics; Determining multiple noise reduction effects of the multiple first candidate signals based on the multiple first feedback signals and the noise frequency characteristics; Based on the multiple noise reduction effects and / or the environmental noise type, the target phase corresponding to the target frequency and the initial target signal are determined.

4. The method according to claim 1, characterized in that The step of determining a target single-frequency signal corresponding to the target frequency based on a plurality of candidate amplitudes and the initial target signal comprises: Based on the plurality of candidate amplitudes and the initial target signal, generating a plurality of second candidate signals; Acquire a second feedback signal corresponding to the second candidate signal; Based on the second feedback signal, a target single-frequency signal corresponding to the target frequency is determined.

5. A noise reduction system, characterized in that: The system comprises: a characteristic determination module configured to determine a noise frequency characteristic of the ambient noise; A frequency group determination module is configured to determine a target frequency group based on the noise frequency characteristics; the target frequency group includes one or more target frequencies; The signal determination module is configured as follows: For a target frequency: Based on the plurality of candidate phases, generating a plurality of first candidate signals; Acquire a first feedback signal corresponding to the first candidate signal; Determining a target phase and an initial target signal corresponding to the target frequency based on the first feedback signal and the noise frequency characteristic; Determining a target single-frequency signal corresponding to the target frequency based on a plurality of candidate amplitudes and the initial target signal; The noise reduction completion module is configured to complete noise reduction based on the target single-frequency signal group corresponding to the target frequency group.

6. The system according to claim 5, characterized in that The frequency group determination module is further configured to: Based on the noise frequency characteristics, determining the noise energy of the initial frequency of the ambient noise; Based on the noise energy, the target frequency group is determined.

7. The system according to claim 5, characterized in that The signal determination module is further configured to: Determining the type of environmental noise based on the noise frequency characteristics; Determining multiple noise reduction effects of the multiple first candidate signals based on the multiple first feedback signals and the noise frequency characteristics; Based on the multiple noise reduction effects and / or the environmental noise type, the target phase corresponding to the target frequency and the initial target signal are determined.

8. The system according to claim 5, characterized in that The signal determination module is further configured to: Based on the plurality of candidate amplitudes and the initial target signal, generating a plurality of second candidate signals; Acquire a second feedback signal corresponding to the second candidate signal; Based on the second feedback signal, a target single-frequency signal corresponding to the target frequency is determined.

9. A noise reduction device, characterized in that: The device comprises at least one memory and at least one processor, wherein the at least one memory is used to store computer instructions, and the at least one processor executes the computer instructions or part of the instructions to implement the noise reduction method according to any one of claims 1 to 4.

10. A computer-readable storage medium, characterized in that: The storage medium stores computer instructions. When a computer reads the computer instructions, the computer executes the noise reduction method according to any one of claims 1 to 4.

Images