WO2024113811A1 - 降噪方法、耳机、装置、存储介质及计算机程序产品 - Google Patents

降噪方法、耳机、装置、存储介质及计算机程序产品 Download PDF

Info

Publication number
WO2024113811A1
WO2024113811A1 PCT/CN2023/103264 CN2023103264W WO2024113811A1 WO 2024113811 A1 WO2024113811 A1 WO 2024113811A1 CN 2023103264 W CN2023103264 W CN 2023103264W WO 2024113811 A1 WO2024113811 A1 WO 2024113811A1
Authority
WO
WIPO (PCT)
Prior art keywords
frame
filter
target
noise reduction
filters
Prior art date
Application number
PCT/CN2023/103264
Other languages
English (en)
French (fr)
Inventor
李玉龙
李佳生
欧阳山
吴国鹏
杨培年
范泛
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Publication of WO2024113811A1 publication Critical patent/WO2024113811A1/zh

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
    • H04R1/00—Details of transducers, loudspeakers or microphones
    • H04R1/10—Earpieces; Attachments therefor ; Earphones; Monophonic headphones
    • H04R1/1083—Reduction of ambient noise
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
    • H04R1/00—Details of transducers, loudspeakers or microphones
    • H04R1/10—Earpieces; Attachments therefor ; Earphones; Monophonic headphones
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
    • H04R2201/00—Details of transducers, loudspeakers or microphones covered by H04R1/00 but not provided for in any of its subgroups
    • H04R2201/10—Details of earpieces, attachments therefor, earphones or monophonic headphones covered by H04R1/10 but not provided for in any of its subgroups
    • H04R2201/105—Manufacture of mono- or stereophonic headphone components

Definitions

  • the present application relates to the field of audio processing technology, and in particular to a noise reduction method, earphones, device, storage medium and computer program product.
  • the present application provides a noise reduction method, earphone, device, storage medium and computer program product, which can improve the active noise reduction effect of earphone.
  • the technical solution is as follows:
  • a noise reduction method is provided, which is applied to headphones, wherein the headphones include at least one reference microphone, one error microphone and multiple first speakers; the method includes: determining multiple groups of target noise reduction parameters corresponding one-to-one to the multiple first speakers; based on the multiple groups of target noise reduction parameters, generating multiple groups of target anti-phase noise corresponding one-to-one to the multiple first speakers, wherein the frequency band of each target anti-phase noise in the multiple groups of target anti-phase noise covers the sound frequency band of the multiple first speakers; and utilizing the multiple groups of target anti-phase noise to perform noise reduction through the multiple first speakers.
  • each target anti-phase noise in the multiple groups of target anti-phase noises covers the sound frequency band of the multiple first speakers, that is, each target anti-phase noise is a full-band anti-phase noise, so no matter whether the first speaker is a high-frequency speaker, a low-frequency speaker or a full-range speaker, when using the multiple groups of target anti-phase noises for noise reduction, the noise reduction ability of each first speaker can be fully utilized.
  • this solution can improve the noise reduction effect of the headphone through the full-band anti-phase noise of multiple noise reduction channels.
  • the noise reduction method provided in the present application can determine the multiple groups of target noise reduction parameters in units of frames, that is, multiple groups of target noise reduction parameters corresponding to the multiple first speakers are determined in each frame.
  • the target noise reduction parameters can also be determined in other time units, for example, multiple groups of target noise reduction parameters corresponding to the multiple first speakers are determined every two frames. The following is an introduction based on frames.
  • the headset also includes a plurality of feed forward (FF) filters corresponding one-to-one to the plurality of first speakers.
  • the plurality of target noise reduction parameters include the k-th frame filter coefficients of the plurality of FF filters, where k is an integer greater than or equal to 1.
  • the headset also includes a plurality of feedback (FB) filters corresponding one-to-one to the plurality of first speakers, that is, the plurality of FB filters correspond one-to-one to the plurality of FF filters.
  • the plurality of target noise reduction parameters also include the k-th frame filter coefficients of the plurality of FB filters.
  • the plurality of target noise reduction parameters also include the k-th frame filter coefficients of the downlink compensation filter.
  • the target noise reduction gear can also be determined. Therefore, these four parts will be introduced separately below.
  • the initial filter coefficients of the multiple FF filters are determined as the k-th frame filter coefficients of the multiple FF filters, that is, the first frame filter coefficients of the multiple FF filters are the initial filter coefficients of the corresponding FF filters, or, based on the initial noise reduction gear and the mapping relationship between the noise reduction gear and the FF filter coefficient, the k-th frame filter coefficients of the multiple FF filters are determined.
  • the k-th frame filter coefficients of the multiple FF filters are determined. That is, the k-th frame filter coefficients of the multiple FF filters are determined by an adaptive method, and the determination process is an adaptive process, which can also be called an iterative process.
  • the initial filter coefficients of the multiple FF filters may be the same or different, and the initial filter coefficient may be 0 or not, which is not limited in the embodiments of the present application.
  • the initial noise reduction gear may be a pre-set gear, which refers to a gear at which the corresponding noise reduction coefficient can perform noise reduction normally without introducing stability problems.
  • the initial noise reduction gear may also be a gear determined by prompt sounds such as "noise reduction on" and "ding dong" sent by the user terminal at the beginning of the noise reduction.
  • the noise reduction coefficient corresponding to the gear can better adapt to the current human ear and wearing posture, and adaptive iteration based on the noise reduction coefficient corresponding to the gear can reach a convergence state faster, which is also not limited in the embodiments of the present application.
  • the implementation process of determining the k-1th frame filter coefficients of the multiple FF filters based on the k-1th frame reference signal collected by the at least one reference microphone, the k-1th frame error signal collected by the error microphone, and the target noise reduction gear includes: determining the k-1th frame filter coefficients of multiple SPs based on the target noise reduction gear and the mapping relationship between the noise reduction gear and the filter coefficients of the secondary path (SP), wherein the multiple SPs refer to the paths from the multiple first speakers to the error microphone.
  • SP secondary path
  • the headset When determining the k-th frame filter coefficients of the multiple FF filters, it can be determined by multi-channel linkage. Moreover, when the headset includes multiple FF filters, it may also include multiple FB filters corresponding to the multiple first speakers one by one, or it may not include the multiple FB filters. In different situations, the methods for determining the k-th frame filter coefficients of the multiple FF filters are different. They will be introduced separately below.
  • the process of determining the k-th frame filter coefficient of each FF filter is the same based on the k-1th frame reference signal collected by the at least one reference microphone, the k-1th frame error signal collected by the error microphone, and the k-1th frame filter coefficient of the multiple SPs, one of them is used as an example for description. That is, one of the multiple FF filters is used as the target FF filter, and the k-th frame filter coefficient of the target FF filter is determined in the following manner, and the process of determining the k-th frame filter coefficients of other FF filters in the multiple FF filters can refer to the process of determining the k-th frame filter coefficient of the target FF filter.
  • the headset does not include the multiple FB filters. If the target FF filter is the first FF filter, the k-th frame filter coefficient of the target FF filter is determined based on the k-1th frame reference signal collected by the target reference microphone, the k-1th frame error signal collected by the error microphone, and the k-1th frame filter coefficient of the target SP, where the target reference microphone is the reference microphone corresponding to the target FF filter, and the target SP refers to the path from the first speaker corresponding to the target FF filter to the error microphone.
  • the k-th frame filter coefficient of the target FF filter is determined based on the k-1th frame reference signal collected by the target reference microphone, the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficient of the multiple SPs, and the k-th frame frequency response information and the k-1th frame frequency response information of each FF filter located before the target FF filter.
  • the residual error is determined based on the k-1th frame reference signal collected by the target reference microphone and the k-1th frame error signal collected by the error microphone, and the k-1th frame frequency response information of the target FF filter is determined based on the k-1th frame frequency response information of the target FF filter, the k-1th frame filter coefficient of the target SP and the residual error.
  • the k-th frame filter coefficient of the target FF filter is determined based on the k-th frame frequency response information of the target FF filter.
  • One FF filter among the plurality of FF filters corresponds to one reference microphone. That is, the target reference microphone includes one reference microphone. At this time, the residual error is determined based on the k-1th frame reference signal collected by the target reference microphone and the k-1th frame error signal collected by the error microphone.
  • One FF filter among the multiple FF filters corresponds to at least two reference microphones. That is, the target reference microphone includes at least two reference microphones. At this time, the k-1th frame reference signal collected by the at least two reference microphones included in the target reference microphone is mixed to obtain the k-1th frame mixed reference signal. Based on the k-1th frame mixed reference signal and the k-1th frame error signal collected by the error microphone, the residual error is determined. In this way, the signal-to-noise ratio of the reference signal can be improved.
  • the residual error is determined based on the k-1th frame reference signal collected by the target reference microphone and the k-1th frame error signal collected by the error microphone.
  • the k-th frame frequency response information of the target FF filter is determined based on the k-1th frame frequency response information of the target FF filter, the residual error, the k-1th frame filter coefficients of the multiple SPs, and the k-th frame frequency response information and the k-1th frame frequency response information of each FF filter located before the target FF filter.
  • the k-th frame filter coefficient of the target FF filter is determined based on the k-th frame frequency response information of the target FF filter.
  • the k-th frame frequency response information of the target FF filter can be determined based on the k-1-th frame frequency response information of the target FF filter, the residual error, the k-1-th frame filter coefficient of the target SP, the k-th frame frequency response information and the k-1-th frame frequency response information of each FF filter located before the target FF filter, and the k-1-th frame filter coefficient of the SP corresponding to each FF filter located before the target FF filter.
  • the implementation process of determining the k-th frame filter coefficient of the target FF filter includes: establishing a loss function between the filter coefficient variable of the target FF filter and the k-th frame frequency response information of the target FF filter. Based on the loss function, the value of the filter coefficient variable is determined by the gradient descent method, and the k-th frame filter coefficient of the target FF filter is determined based on the value of the filter coefficient variable. That is, a loss function is established between the filter coefficient variable of the target FF filter and the k-th frame frequency response information of the target FF filter. The optimal value of the variable is determined by the gradient descent method, so that the k-th frame filter coefficient of the target FF filter is determined by the optimal value of the variable.
  • the filter coefficients of each frame of the target FF filter are determined according to the gradient descent method.
  • a value of a loss function is determined.
  • the value of the loss function reaches the minimum threshold value, it is determined that the filter coefficients of the target FF filter have reached the convergence stability condition.
  • the filter coefficients of the kth frame of the target FF filter when the value of the loss function between the filter coefficient variable and the kth frame frequency response information of the target FF filter reaches the minimum threshold value, it is determined that the filter coefficients of the kth frame of the target FF filter have reached the convergence stability condition.
  • the value of the loss function does not reach the minimum threshold value, it is determined that the filter coefficients of the kth frame of the target FF filter have not reached the convergence stability condition.
  • the minimum threshold value is set in advance and can be adjusted according to different needs in different situations.
  • the filter coefficient of each FF filter includes at least one biquad filter coefficient and a gain.
  • the variables corresponding to the biquad filter coefficient include filter type, cutoff frequency and quality factor.
  • the filter coefficient of each FF filter may also include other more or fewer parameters, which is not limited in this application.
  • a quiet environment has a background noise problem, that is, background noise.
  • background noise For example, for semi-open headphones, the headphones are more prone to background noise problems in a quiet environment than in-ear headphones.
  • a quiet environment does not require strong noise reduction, and some people will feel uncomfortable with strong noise reduction in a quiet environment. And the greater the noise reduction intensity, the stronger the negative pressure people feel. Therefore, when the value of the filter coefficient variable is determined by the gradient descent method, the target noise reduction amplitude can be dynamically adjusted based on the ambient volume, so as to determine the k-th frame filter coefficient of the target FF filter according to the target noise reduction amplitude, thereby improving the subjective experience effect of adaptive noise reduction.
  • the target noise reduction amplitude is determined according to the ambient volume of the k-1th frame and the ambient volume of the t-frame before the k-1th frame, where t is greater than or equal to 1 and less than k-1.
  • the value of the filter coefficient variable is determined by the gradient descent method, and the k-th frame filter coefficient of the target FF filter is determined based on the value of the filter coefficient variable.
  • the earphone also includes the multiple FB filters. If the target FF filter is the first FF filter, the k-th frame filter coefficient of the target FF filter is determined based on the k-1th frame reference signal collected by the target reference microphone, the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficients of the multiple SPs, and the k-1th frame filter coefficients of the multiple FB filters.
  • the k-th frame filter coefficient of the target FF filter is determined based on the k-1th frame reference signal collected by the target reference microphone, the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficients of the multiple SPs, the k-1th frame filter coefficients of the multiple FB filters, and the k-th frame frequency response information and the k-1th frame frequency response information of each FF filter located before the target FF filter.
  • the residual error may be determined based on the k-1th frame reference signal collected by the target reference microphone and the k-1th frame error signal collected by the error microphone.
  • the k-th frame frequency response information of the target FF filter is determined based on the k-1th frame frequency response information of the target FF filter, the residual error, the k-1th frame filter coefficients of the multiple FB filters, and the k-1th frame filter coefficients of the multiple SPs.
  • the k-th frame filter coefficient of the target FF filter is determined based on the k-th frame frequency response information of the target FF filter.
  • the residual error may be determined based on the k-1th frame reference signal collected by the target reference microphone and the k-1th frame error signal collected by the error microphone.
  • the k-1th frame frequency response information of the target FF filter is determined based on the k-1th frame frequency response information of the target FF filter, the residual error, the k-1th frame filter coefficients of the multiple SPs, the k-1th frame filter coefficients of the multiple FB filters, and the k-1th frame frequency response information and the k-1th frame frequency response information of each FF filter located before the target FF filter.
  • the k-1th frame filter coefficient of the target FF filter is determined based on the k-1th frame frequency response information of the target FF filter.
  • the k-th frame frequency response information of the target FF filter is determined based on the k-1-th frame filter coefficient of the target SP, and the k-1-th frame filter coefficient of the target SP is determined based on the target noise reduction gear by querying the mapping relationship between the noise reduction gear and the filter coefficient of the SP, that is, the k-1-th frame filter coefficient of the target SP is an estimated value.
  • the filter coefficient of the FF filter can be adapted.
  • the initial filter coefficients of the multiple FB filters are determined as the k-th frame filter coefficients of the multiple FB filters, that is, the first frame filter coefficients of the multiple FB filters are the initial filter coefficients of the corresponding FB filters, or the k-th frame filter coefficients of the multiple FB filters are determined based on the initial noise reduction gear and the mapping relationship between the noise reduction gear and the FB filter coefficients.
  • the k-th frame filter coefficients of the multiple FB filters can be determined based on the target noise reduction gear.
  • initial filter coefficients of the multiple FB filters may be the same or different, and the initial filter coefficient may be 0 or not 0, which is not limited in the embodiment of the present application.
  • the process of determining the k-th frame filter coefficient of each FB filter based on the target noise reduction gear is the same, one of them is used as an example for introduction. That is, one of the multiple FB filters is used as the target FB filter, and the k-th frame filter coefficient of the target FB filter is determined in the following two ways.
  • the process of determining the k-th frame filter coefficients of other FB filters in the multiple FB filters can refer to the process of determining the k-th frame filter coefficient of the target FB filter. In other words, when k is greater than 1, the k-th frame filter coefficient of the target FB filter can be determined in the following two ways.
  • the first method is to determine the k-th frame filter coefficient of the target FB filter based on the target noise reduction gear and the mapping relationship between the noise reduction gear and the FB filter coefficient.
  • the k-th frame filter coefficient of the target FB filter determined by the first method has better stability, simpler operation and higher efficiency.
  • the first frame filter coefficient of the target FB filter can be determined by querying the mapping relationship between the noise reduction gear and the FB filter coefficient through the initial noise reduction gear, when k is greater than or equal to 1, it is equivalent to determining the kth frame filter coefficient of the target FB filter in three ways. That is, (1) the kth frame filter coefficient of the target FB filter is determined by querying the mapping relationship between the noise reduction gear and the FB filter coefficient. (2) If the target FB filter belongs to the first type of FB filter, the kth frame filter coefficient of the target FB filter is determined by querying the mapping relationship between the noise reduction gear and the FB filter coefficient.
  • the kth frame filter coefficient of the target FB filter is determined based on the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficient of the target FB filter and the target noise reduction gear. (3) If the target FB filter belongs to the first type of FB filter, or the target FB filter belongs to the second type of FB filter and k is equal to 1, the kth frame filter coefficient of the target FB filter is determined by querying the mapping relationship between the noise reduction gear and the FB filter coefficient.
  • the k-th frame filter coefficient of the target FB filter is determined based on the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficient of the target FB filter, and the target noise reduction gear.
  • the implementation process of determining the k-1th frame filter coefficient of the target FB filter based on the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficient of the target FB filter and the target noise reduction gear includes: determining the k-1th frame filter coefficient of the target SP based on the target noise reduction gear and the mapping relationship between the noise reduction gear and the filter coefficient of the SP, the target SP refers to the path from the first speaker corresponding to the target FB filter to the error microphone; determining the k-1th frame filter coefficient of the target FB filter based on the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficient of the target FB filter and the k-1th frame filter coefficient of the target SP.
  • the sound frequency band of the first speaker corresponding to the first type of FB filter is higher than the sound frequency band of the first speaker corresponding to the second type of FB filter. That is, the first speaker corresponding to the first type of FB filter is a high-frequency speaker, and the first speaker corresponding to the second type of FB filter is a low-frequency speaker.
  • the first type of FB filter and the second type of FB filter can also be distinguished in other ways instead of according to the sound frequency band, and this application does not limit this either.
  • the second and third methods mentioned above combine the method of querying the mapping relationship between the noise reduction gear and the FB filter coefficient with the adaptive method, which can not only improve the noise reduction effect, but also has relatively low complexity and relatively controllable stability.
  • the k-th frame filter coefficient of the target FB filter can be determined not only in the above three ways, but also in other ways. For example, regardless of whether the target FB filter belongs to the first type of FB filter or the second type of FB filter, the k-th frame filter coefficient of the target FB filter is determined based on the k-1-th frame error signal collected by the error microphone, the k-1-th frame filter coefficient of the target FB filter, and the target noise reduction gear. This embodiment of the application does not limit this.
  • the initial filter coefficient of the downlink compensation filter is determined as the filter coefficient of the kth frame of the downlink compensation filter, or, based on the initial noise reduction gear and the mapping relationship between the noise reduction gear and the downlink compensation filter coefficient, the kth frame of the downlink compensation filter is determined.
  • Filter coefficient When k is greater than 1, the filter coefficient of the kth frame of the downlink compensation filter is determined based on the target noise reduction level and the mapping relationship between the noise reduction level and the downlink compensation filter coefficient.
  • the mapping relationship between the noise reduction gear and the downlink compensation filter coefficient includes multiple noise reduction gears, each noise reduction gear has a mapping relationship with the filter coefficient of the downlink compensation filter, and the mapping relationship between different noise reduction gears and the filter coefficient of the downlink compensation filter may be different. Therefore, after determining the target noise reduction gear, based on the target noise reduction gear, the corresponding downlink compensation filter coefficient can be obtained from the mapping relationship between the noise reduction gear and the downlink compensation filter coefficient, and the obtained downlink compensation filter coefficient is used as the kth frame filter coefficient of the downlink compensation filter.
  • Determine the noise reduction level of the k-1th frame obtain the noise reduction levels of m frames before the k-1th frame, where m is greater than or equal to 1 and less than k-1. Determine the target noise reduction level based on the noise reduction level of the k-1th frame and the noise reduction levels of the m frames.
  • the k-1th frame may or may not have a valid downlink signal, and may be in a quiet environment or a non-quiet environment, and of course, there may be abnormal signals, the methods for determining the noise reduction level of the k-1th frame are different in different situations, which will be introduced below.
  • the noise reduction gear of the k-1th frame is determined based on the reference filter coefficients of the multiple FF filters and the mapping relationship between the noise reduction gear and the frequency response information of the FF filter.
  • the reference filter coefficient is the initial filter coefficient of the corresponding FF filter
  • the reference filter coefficient is the filter coefficient of the corresponding FF filter that most recently reached the convergence stability condition before the kth frame, or the filter coefficient of the k-1th frame of the corresponding FF filter.
  • the reference frequency response information of the multiple FF filters is determined. Based on the mapping relationship between the noise reduction level and the frequency response information of the FF filters, the noise reduction levels that match the reference frequency response information of the multiple FF filters are determined to obtain multiple reference noise reduction levels. Based on the multiple reference noise reduction levels, the noise reduction level of the k-1th frame is determined.
  • the noise reduction level of the k-1th frame based on the multiple reference noise reduction levels, for example, determining the noise reduction level of the k-1th frame according to the average value of the multiple reference noise reduction levels, or determining the noise reduction level of the k-1th frame according to the reference noise reduction level with the largest number among the multiple reference noise reduction levels.
  • the average value of the multiple reference noise reduction levels can be directly determined as the noise reduction level of the k-1 frame, or the average value of the multiple reference noise reduction levels can be adjusted to obtain the noise reduction level of the k-1 frame.
  • the reference noise reduction level with the largest number among the multiple reference noise reduction levels can be directly determined as the noise reduction level of the k-1 frame, or the reference noise reduction level with the largest number among the multiple reference noise reduction levels can be adjusted to obtain the noise reduction level of the k-1 frame.
  • the noise reduction gear of the k-1th frame is determined based on the valid downlink signal of the k-1th frame, the reference signal of the k-1th frame collected by the at least one reference microphone, and the error signal of the k-1th frame collected by the error microphone.
  • the headset when the headset is in a downlink enabled state and is not in a downlink intermittent period, it is determined that there is a valid downlink signal in the k-1th frame.
  • the valid downlink signal can be extracted from the k-1th frame error signal collected by the error microphone, thereby determining the k-1th frame noise reduction gear based on the extracted valid downlink signal.
  • the noise reduction level of the k-3 frame is determined to be the noise reduction level of the k-1 frame. That is, the noise reduction level is maintained unchanged.
  • the noise has basically not changed, and the noise reduction level can be maintained unchanged.
  • the noise reduction level is maintained unchanged to perform robust control, thereby avoiding the divergence of the noise reduction level.
  • the noise reduction level of the k-1th frame After the noise reduction level of the k-1th frame is determined through the above three situations, the noise reduction level of the k-1th frame and the noise reduction levels of m frames before the k-1th frame may be combined to determine the target noise reduction level.
  • the m-frame noise reduction gear can be any m-frame noise reduction gear located before the k-1th frame, or it can be the m-frame noise reduction gear located before the k-1th frame and closest to the k-1th frame.
  • the noise reduction effect is evaluated according to the relevant algorithm to determine the noise reduction probability corresponding to the k-1th frame noise reduction gear and the noise reduction probability corresponding to the m-frame noise reduction gears, and the noise reduction gear with the largest noise reduction probability is determined as the target noise reduction gear.
  • the arithmetic mean or weighted average of the k-1th frame noise reduction gear and the m-frame noise reduction gear is determined to obtain the target noise reduction gear.
  • the noise reduction gear that appears most frequently in the noise reduction gear of the k-1th frame and the noise reduction gear of the m frames is determined as the target noise reduction gear, and so on.
  • the multiple sets of target noise reduction parameters can be referred to as noise reduction parameters of multiple noise reduction channels, and thus, the multiple sets of target anti-phase noise generated can also be referred to as anti-phase noise of multiple noise reduction channels. Since the generation process of the anti-phase noise of each noise reduction channel is the same, one of the noise reduction channels is used as an example for explanation below.
  • the target noise reduction channel includes a target FF filter and a target first speaker, and the reference microphone corresponding to the target FF filter is called a target reference microphone.
  • the target anti-phase noise includes feedforward anti-phase noise. That is, based on the k-th frame filter coefficient of the target FF filter, the k-th frame reference signal collected by the target reference microphone is processed to obtain the feedforward anti-phase noise.
  • the target reference microphone may include one reference microphone or may include at least two reference microphones.
  • the k-th frame reference signal collected by the target reference microphone can be directly processed based on the k-th frame filter coefficient of the target FF filter to obtain feedforward inverted noise.
  • the target reference microphone includes at least two reference microphones, the k-th frame reference signal collected by the at least two reference microphones is mixed to obtain the k-th frame mixed reference signal, and then, based on the k-th frame filter coefficient of the target FF filter, the k-th frame mixed reference signal is processed to obtain feedforward inverted noise.
  • the target noise reduction channel also includes a target FB filter.
  • the target anti-phase noise also includes feedback anti-phase noise. That is, based on the k-th frame filter coefficient of the downlink compensation filter, the k-th frame downlink signal sent by the user terminal is downlink compensated. Then, the k-th frame downlink signal after downlink compensation is inverted and mixed with the k-th frame error signal collected by the error microphone to obtain the k-th frame noise signal collected by the error microphone. Based on the k-th frame filter coefficient of the target FB filter, the k-th frame noise signal collected by the error microphone is processed to obtain feedback anti-phase noise.
  • one frame may include one sample point or multiple sample points
  • a group of target inverted noise can be generated at each sample point, or a group of target inverted noise can be generated in one frame.
  • the downlink signal is not frequency-divided when determining the multiple sets of target noise reduction parameters, that is, the multiple sets of target noise reduction parameters are determined by the full-frequency downlink signal.
  • the frequency band of each target anti-phase noise in the multiple sets of target anti-phase noises covers the sound frequency band of the multiple first speakers, that is, the frequency band of each target anti-phase noise is full-band.
  • the multiple groups of target anti-phase noises are mixed with the k-th frame downlink signals to be played by the multiple first speakers respectively, and then played through the corresponding first speakers, thereby achieving the purpose of noise reduction.
  • Some of the multiple first speakers may be high-frequency speakers, and others may be low-frequency speakers. Alternatively, some may be full-range speakers, and others may be non-full-range speakers. That is, the sound frequency bands of the multiple first speakers may be different. Alternatively, all of the multiple first speakers are full-range speakers. Alternatively, all of the multiple first speakers are non-full-range speakers. In the case where all of the multiple first speakers are full-range speakers, the k-th frame downlink signal to be played by the multiple first speakers is the k-th frame downlink signal sent by the user terminal.
  • Two of the plurality of first speakers may include two first speakers formed by a double-diaphragm (or double-dynamic) speaker, or the plurality of first speakers may include a plurality of speakers of split speakers.
  • the headset may also include at least one second speaker, which does not participate in noise reduction.
  • the second speaker may participate in downlink compensation (i.e., downlink compensation is performed on the downlink signal sent by the user terminal, and the downlink signal is a full-band audio signal, including the audio signal of the sound frequency band of the second speaker).
  • the first speaker may be a medium-low frequency speaker, or a full-band speaker
  • the second speaker may be a high-frequency speaker, or a medium-frequency speaker or a low-frequency speaker.
  • the second speaker may not participate in downlink compensation.
  • the first speaker may be a medium-low frequency speaker, or a full-band speaker
  • the second speaker may be a high-frequency speaker.
  • a frame Since the process of determining the multiple sets of target noise reduction parameters by the adaptive method takes a certain amount of time, a frame includes multiple samples and When the duration of a frame is long, the determination duration of the multiple sets of target noise reduction parameters will be less than the duration of a frame. Therefore, through the relevant data of the k-1th frame, calculations can be performed in a partial time period starting from the kth frame to obtain multiple sets of target noise reduction parameters for the kth frame, and then active noise reduction is performed in a partial time period subsequent to the kth frame according to the multiple sets of target noise reduction parameters of the kth frame.
  • the determination duration of the multiple sets of target noise reduction parameters may be equal to the duration of a frame.
  • the multiple sets of target noise reduction parameters can be determined as multiple sets of target noise reduction parameters for the k+1th frame, and then active noise reduction is performed in the time period of the k+1th frame according to the multiple sets of target noise reduction parameters of the k+1th frame.
  • a headset comprising at least one reference microphone, an error microphone, a plurality of first speakers and a noise reduction processor, wherein the noise reduction processor is used to implement the steps of the method described in the first aspect.
  • the plurality of first speakers include two first speakers formed by a double-diaphragm speaker; or, the plurality of first speakers include a plurality of speakers that are split speakers.
  • the headset further includes at least one second speaker, and the at least one second speaker does not participate in noise reduction.
  • a noise reduction device which has the function of implementing the noise reduction method in the first aspect.
  • the noise reduction device includes one or more modules, which are used to implement the noise reduction method provided in the first aspect.
  • a computer-readable storage medium wherein instructions are stored in the computer-readable storage medium, and when the computer-readable storage medium is run on a computer, the computer executes the noise reduction method described in the first aspect.
  • a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the noise reduction method described in the first aspect.
  • FIG1 is a system architecture diagram of a noise reduction method provided in an embodiment of the present application.
  • FIG2 is a flow chart of a noise reduction method provided in an embodiment of the present application.
  • FIG3 is a flow chart of determining a target noise reduction range provided by an embodiment of the present application.
  • FIG4 is a schematic diagram of a frequency response curve of an FF filter at 16 noise reduction gears provided in an embodiment of the present application
  • FIG5 is a flow chart of determining a noise reduction level for the k-1th frame provided by an embodiment of the present application.
  • FIG6 is a flow chart of determining multiple groups of target noise reduction parameters provided by an embodiment of the present application.
  • FIG7 is a schematic diagram of the structure of an earphone provided in an embodiment of the present application.
  • FIG8 is a schematic diagram of the structure of another earphone provided in an embodiment of the present application.
  • FIG9 is a schematic diagram of the structure of another earphone provided in an embodiment of the present application.
  • FIG10 is a schematic diagram of the structure of another earphone provided in an embodiment of the present application.
  • FIG11 is a schematic diagram of the structure of another earphone provided in an embodiment of the present application.
  • FIG12 is a schematic structural diagram of a noise reduction device provided in an embodiment of the present application.
  • FIG. 13 is a schematic diagram of the structure of another earphone provided in an embodiment of the present application.
  • Semi-open headphones are widely accepted by users due to their good comfort. However, due to the poor sealing between the headphones and the human ears in the semi-open form, environmental noise is more easily felt. It is more challenging to achieve active noise reduction of headphones in a semi-open form. The reason is that when different people wear them, or even when the same person wears them at different times, the wearing postures will vary greatly. In technical terms, the response function and acoustic leakage degree between the headphones and the ear canal are very different. Therefore, how to achieve adaptive noise reduction is an urgent demand in a semi-open form, in order to cope with the problem of differentiated ear canal response, so as to achieve the best match between the headphones and the ear canal. Moreover, even in the in-ear or head-mounted form, the ear canal response is not absolutely consistent, and there are still differences of varying degrees. The industry is currently exploring the feasibility of adaptive noise reduction for headphones.
  • earphones come in many forms, including in-ear, head-mounted, semi-open, and open.
  • the audio performance of the speaker (i.e., horn) in the whole device is strongly related to the specific form of the earphone, especially the low-frequency performance.
  • the audio performance of high, medium, and low frequencies is generally guaranteed; in semi-open or open forms, due to severe acoustic leakage, the low-frequency response drops significantly, which not only affects the expressiveness of the low-frequency sound quality, but also seriously affects the effect of active noise reduction (not enough to generate reverse phase noise with sufficient energy).
  • an embodiment of the present application provides a noise reduction method to achieve adaptive active noise reduction (active noise cancellation, ANC) of headphones.
  • Figure 1 is a system architecture diagram involved in a noise reduction method provided in an embodiment of the present application.
  • the system can be called a headphone noise reduction system.
  • the system includes a headphone 101 and a user terminal 102.
  • the headphone 101 and the user terminal 102 are connected to communicate via a wired or wireless method.
  • the headphone 101 communicates with the user terminal 102 via Bluetooth, or communicates via other wireless networks.
  • Audio signals and control signals can be transmitted between the earphone 101 and the user terminal 102.
  • the user terminal 102 sends an audio signal such as music or voice to the earphone 101 for playback, and for another example, the user terminal 102 sends a control signal to the earphone 101 to control whether the active noise reduction function of the earphone 101 is turned on, etc.
  • the user terminal 102 may be a mobile phone, a computer (such as a laptop computer, a desktop computer, a handheld tablet computer, a car tablet computer) or other electronic devices, and may also be other electronic devices, such as a smart speaker, a car speaker, etc.
  • a computer such as a laptop computer, a desktop computer, a handheld tablet computer, a car tablet computer
  • other electronic devices such as a smart speaker, a car speaker, etc.
  • the embodiment of the present application does not limit the type and structure of the user terminal 102.
  • the earphone 101 provided in the embodiment of the present application may be wired or wireless.
  • the earphone 101 provided in the embodiment of the present application may be neck-worn, ear-hook/ear-clip, true wireless stereo (TWS), etc.
  • TWS true wireless stereo
  • the earphone 101 provided in the embodiment of the present application may be in-ear, semi-open, open, head-worn, etc.
  • the embodiment of the present application does not limit the communication method, wearing method and appearance of the earphone.
  • the hardware structure of the earphone provided in the embodiment of the present application is introduced below in combination with the wearing method of the earphone in the human ear.
  • the headset 101 includes a plurality of speakers (i.e., speakers), a plurality of microphones, a microcontroller (MCU), an ANC chip, and a memory.
  • the plurality of speakers include a plurality of first speakers, such as speaker 1 and speaker 2.
  • the plurality of first speakers need to participate in noise reduction, for example, the first speaker is a mid-low frequency speaker, and the mid-low frequency speaker needs to participate in noise reduction.
  • the plurality of speakers also include at least one second speaker, and the at least one second speaker does not participate in noise reduction, for example, the second speaker is a high frequency speaker, and the high frequency speaker does not need to participate in noise reduction.
  • the plurality of microphones include at least one reference microphone and one error microphone.
  • FIG1 is introduced by taking a reference microphone as an example.
  • the speaker is used to play downlink signals (such as music, voice and other audio signals), and each speaker is driven by an independent digital to analog converter (DAC) and power amplifier (PA), that is, one speaker corresponds to one DAC and one PA, and different speakers correspond to different DACs and PAs.
  • DAC digital to analog converter
  • PA power amplifier
  • the first speaker is also used to play anti-phase noise, which is used to reduce the noise signal in the user's ear canal, thereby achieving the effect of actively reducing noise.
  • the reference microphone is deployed outside the earphone. After the earphone is worn on the human ear, the reference microphone is located outside the human ear.
  • the reference microphone is used to collect noise signals from the external environment. In the embodiment of the present application, the noise signal collected by the reference microphone is referred to as a reference signal.
  • the error microphone is deployed inside the earphone. After the earphone is worn on the human ear, the error microphone is located inside the human ear.
  • the error microphone is used to collect noise signals in the ear canal. In the embodiment of the present application, the noise signal collected by the error microphone is referred to as an error signal.
  • the microcontroller is used to process the reference signal collected by the reference microphone, the error signal collected by the error microphone, the downlink signal, etc., so as to determine a set of target noise reduction parameters corresponding to each first speaker in the multiple first speakers, and write the set of target noise reduction parameters corresponding to each first speaker into the ANC chip.
  • the ANC chip is used to calculate the reference signal and error signal collected by the reference microphone based on a set of target noise reduction parameters corresponding to each first speaker.
  • the error signal collected by the microphone is processed to generate anti-phase noise, and the generated anti-phase noise is then mixed with the downlink signal to be played by the first speaker and output to the corresponding first speaker to weaken the noise signal in the ear canal.
  • the memory is used to store initial parameters involved in determining the target noise reduction parameters corresponding to each first speaker, as well as mapping relationships and the like.
  • microcontroller, ANC chip and memory can be integrated on the same circuit board or deployed on different circuit boards, and the embodiments of the present application do not limit this.
  • the microcontroller and the ANC chip are only distinguished in terms of logical functions. In actual physical form, the microcontroller and the ANC chip can be integrated on one chip or deployed on multiple chips respectively, for example, the microcontroller and the ANC chip are deployed on two chips.
  • the headset 101 may further include other components, such as a proximity light sensor, for detecting whether the headset 101 is in the ear. If the headset 101 is a wireless headset, the headset 101 may further include a wireless communication module, which may be a wireless LAN module or a Bluetooth module. The wireless communication module is used for the headset 101 to communicate with other devices.
  • a wireless communication module which may be a wireless LAN module or a Bluetooth module. The wireless communication module is used for the headset 101 to communicate with other devices.
  • the headset 101 may include more or fewer components than shown in the figure, or combine some components, or separate some components, or arrange the components differently.
  • the components shown in the figure may be implemented in hardware, software, or a combination of hardware and software.
  • Fig. 2 is a flow chart of a noise reduction method provided by an embodiment of the present application, the method is applied to a headset, the headset includes at least one reference microphone, an error microphone and a plurality of first speakers. Referring to Fig. 2, the method includes the following steps.
  • Step 201 Determine a plurality of groups of target noise reduction parameters corresponding one-to-one to the plurality of first speakers.
  • the noise reduction method provided in the embodiment of the present application can determine the multiple groups of target noise reduction parameters in units of frames, that is, each frame determines multiple groups of target noise reduction parameters corresponding to the multiple first speakers.
  • the target noise reduction parameters can also be determined in other time units, for example, multiple groups of target noise reduction parameters corresponding to the multiple first speakers are determined every two frames. The following is an introduction based on frames.
  • the headset also includes a plurality of FF filters corresponding one-to-one to the plurality of first speakers.
  • the plurality of sets of target noise reduction parameters include the kth frame filter coefficients of the plurality of FF filters, and k is an integer greater than or equal to 1.
  • the headset also includes a plurality of FB filters corresponding one-to-one to the plurality of first speakers, that is, the plurality of FB filters correspond one-to-one to the plurality of FF filters.
  • the plurality of sets of target noise reduction parameters also include the kth frame filter coefficients of the plurality of FB filters.
  • the plurality of sets of target noise reduction parameters also include the kth frame filter coefficients of the downlink compensation filter.
  • the target noise reduction gear can also be determined. Therefore, these four parts will be introduced separately below.
  • the multiple groups of target noise reduction parameters can also be referred to as noise reduction parameters of multiple noise reduction channels, and one noise reduction channel includes one FF filter and one first speaker.
  • one noise reduction channel also includes one FB filter.
  • the initial filter coefficients of the multiple FF filters are determined as the k-th frame filter coefficients of the multiple FF filters, that is, the 1st frame filter coefficients of the multiple FF filters are the initial filter coefficients of the corresponding FF filters, or the k-th frame filter coefficients of the multiple FF filters are determined based on the initial noise reduction gear and the mapping relationship between the noise reduction gear and the FF filter coefficients.
  • the k-th frame filter coefficients of the multiple FF filters are determined based on the k-1th frame reference signal collected by the at least one reference microphone, the k-1th frame error signal collected by the error microphone, and the target noise reduction gear. That is, the k-th frame filter coefficients of the multiple FF filters are determined by an adaptive method, and the determination process is an adaptive process, which can also be called an iterative process.
  • the initial filter coefficients of the multiple FF filters may be the same or different, and the initial filter coefficient may be 0 or not, which is not limited in the embodiments of the present application.
  • the initial noise reduction gear may be a pre-set gear, which refers to a gear at which the corresponding noise reduction coefficient can perform noise reduction normally without introducing stability problems.
  • the initial noise reduction gear may also be a gear determined by prompt sounds such as "noise reduction on" and "ding dong" sent by the user terminal at the beginning of the noise reduction.
  • the noise reduction coefficient corresponding to the gear can better adapt to the current human ear and wearing posture, and adaptive iteration based on the noise reduction coefficient corresponding to the gear can reach a convergence state faster, which is also not limited in the embodiments of the present application.
  • the implementation process of determining the k-th frame filter coefficients of the multiple FF filters based on the k-1th frame reference signal collected by the at least one reference microphone, the k-1th frame error signal collected by the error microphone, and the target noise reduction gear comprises: determining the k-th frame filter coefficients of the multiple FF filters based on the target noise reduction gear and the target noise reduction gear.
  • the mapping relationship between the noise level and the filter coefficient of the SP is used to determine the k-1th frame filter coefficient of the multiple SPs, where the multiple SPs refer to the paths from the multiple first speakers to the error microphones.
  • the k-1th frame filter coefficient of the multiple FF filters is determined based on the k-1th frame reference signal collected by the at least one reference microphone, the k-1th frame error signal collected by the error microphone, and the k-1th frame filter coefficient of the multiple SPs.
  • the multiple SPs can also be referred to as SPs of multiple noise reduction channels.
  • the mapping relationship between the noise reduction gear and the filter coefficient of the SP includes multiple noise reduction gears. Each noise reduction gear has a mapping relationship with the filter coefficient of the multiple SPs. The mapping relationship between different noise reduction gears and the filter coefficients of the multiple SPs may be different. Therefore, after determining the target noise reduction gear, based on the target noise reduction gear, the filter coefficients corresponding to the multiple SPs can be obtained from the mapping relationship between the noise reduction gear and the filter coefficient of the SP, and the obtained filter coefficients are used as the k-1th frame filter coefficients of the multiple SPs.
  • the initial noise reduction gear is the same.
  • the headset When determining the k-th frame filter coefficients of the multiple FF filters, it can be determined by multi-channel linkage. Moreover, when the headset includes multiple FF filters, it may also include multiple FB filters corresponding to the multiple first speakers one by one, or it may not include the multiple FB filters. In different situations, the methods for determining the k-th frame filter coefficients of the multiple FF filters are different. They will be introduced separately below.
  • the process of determining the k-th frame filter coefficient of each FF filter is the same based on the k-1th frame reference signal collected by the at least one reference microphone, the k-1th frame error signal collected by the error microphone, and the k-1th frame filter coefficient of the multiple SPs, one of them is used as an example for description. That is, one of the multiple FF filters is used as the target FF filter, and the k-th frame filter coefficient of the target FF filter is determined in the following manner, and the process of determining the k-th frame filter coefficients of other FF filters in the multiple FF filters can refer to the process of determining the k-th frame filter coefficient of the target FF filter.
  • the headset does not include the multiple FB filters. If the target FF filter is the first FF filter, the k-th frame filter coefficient of the target FF filter is determined based on the k-1th frame reference signal collected by the target reference microphone, the k-1th frame error signal collected by the error microphone, and the k-1th frame filter coefficient of the target SP, where the target reference microphone is the reference microphone corresponding to the target FF filter, and the target SP refers to the path from the first speaker corresponding to the target FF filter to the error microphone.
  • the k-th frame filter coefficient of the target FF filter is determined based on the k-1th frame reference signal collected by the target reference microphone, the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficient of the multiple SPs, and the k-th frame frequency response information and the k-1th frame frequency response information of each FF filter located before the target FF filter.
  • the residual error is determined based on the k-1th frame reference signal collected by the target reference microphone and the k-1th frame error signal collected by the error microphone, and the k-1th frame frequency response information of the target FF filter is determined based on the k-1th frame frequency response information of the target FF filter, the k-1th frame filter coefficient of the target SP and the residual error.
  • the k-th frame filter coefficient of the target FF filter is determined based on the k-th frame frequency response information of the target FF filter.
  • one FF filter among the plurality of FF filters corresponds to one reference microphone. That is, the target reference microphone includes one reference microphone. At this time, based on the k-1th frame reference signal collected by the target reference microphone and the k-1th frame error signal collected by the error microphone, the residual error is determined according to the following formula (1).
  • Res k-1 refers to the residual error
  • Ref k-1 refers to the k-1th frame reference signal collected by the target reference microphone
  • Err k-1 refers to the k-1th frame error signal collected by the error microphone.
  • one FF filter among the multiple FF filters corresponds to at least two reference microphones. That is, the target reference microphone includes at least two reference microphones. At this time, the k-1th frame reference signal collected by the at least two reference microphones included in the target reference microphone is mixed to obtain the k-1th frame mixed reference signal. Based on the k-1th frame mixed reference signal and the k-1th frame error signal collected by the error microphone, the residual error is determined. In this way, the signal-to-noise ratio of the reference signal can be improved.
  • the method of determining the residual error based on the k-1th frame mixing reference signal and the k-1th frame error signal collected by the error microphone is similar to the method of determining the residual error by formula (1) mentioned above, that is, the k-1th frame error signal collected by the error microphone is divided by the k-1th frame mixing reference signal to obtain the residual error.
  • the frequency response information of the k-1th frame filter coefficient of the target SP can be determined, and then, based on the k-1th frame frequency response information of the target FF filter, the frequency response information of the k-1th frame filter coefficient of the target SP and the residual error, the k-1th frame frequency response information of the target FF filter is determined according to the following formula (2).
  • FF k refers to the k-th frame frequency response information of the target FF filter
  • FF k-1 refers to the k-1-th frame frequency response information of the target FF filter
  • ⁇ refers to the step size, which is set in advance
  • SP k-1 refers to the frequency response information of the k-1-th frame filter coefficient of the target SP.
  • the residual error is determined based on the k-1th frame reference signal collected by the target reference microphone and the k-1th frame error signal collected by the error microphone.
  • the k-th frame frequency response information of the target FF filter is determined based on the k-1th frame frequency response information of the target FF filter, the residual error, the k-1th frame filter coefficients of the multiple SPs, and the k-th frame frequency response information and the k-1th frame frequency response information of each FF filter located before the target FF filter.
  • the k-th frame filter coefficient of the target FF filter is determined based on the k-th frame frequency response information of the target FF filter.
  • the method for determining the residual error is the same as above.
  • the k-th frame frequency response information of the target FF filter can be determined based on the k-1-th frame frequency response information of the target FF filter, the residual error, the k-1-th frame filter coefficient of the target SP, the k-th frame frequency response information and the k-1-th frame frequency response information of each FF filter located before the target FF filter, and the k-1-th frame filter coefficient of the SP corresponding to each FF filter located before the target FF filter.
  • the frequency response information of the k-1th frame filter coefficient of the target SP and the frequency response information of the k-1th frame filter coefficient of the SP corresponding to each FF filter located before the target FF filter may be determined. Then, based on the k-1th frame frequency response information of the target FF filter, the residual error, the frequency response information of the k-1th frame filter coefficient of the target SP, the k-th frame frequency response information and the k-1th frame frequency response information of each FF filter located before the target FF filter, and the frequency response information of the k-1th frame filter coefficient of the SP corresponding to each FF filter located before the target FF filter, the k-th frame frequency response information of the target FF filter is determined according to the following formula (3).
  • FF i,k refers to the k-th frame frequency response information of the target FF filter, that is, the target FF filter is the i-th FF filter among the multiple FF filters
  • FF i,k-1 refers to the k-1-th frame frequency response information of the target FF filter
  • Res i,k-1 refers to the residual error
  • SP i,k-1 refers to the frequency response information of the k-1-th frame filter coefficient of the target SP
  • FF j,k refers to the k-th frame frequency response information of the j-th FF filter located before the target FF filter
  • FF j,k-1 refers to the k-1-th frame frequency response information of the j-th FF filter located before the target FF filter
  • SP j,k-1 refers to the frequency response information of the k-1-th frame filter coefficient of the SP corresponding to the j-th FF filter located before the target FF filter.
  • the implementation process of determining the k-th frame filter coefficient of the target FF filter includes: establishing a loss function between the filter coefficient variable of the target FF filter and the k-th frame frequency response information of the target FF filter. Based on the loss function, the value of the filter coefficient variable is determined by the gradient descent method, and the k-th frame filter coefficient of the target FF filter is determined based on the value of the filter coefficient variable. That is, a loss function is established between the filter coefficient variable of the target FF filter and the k-th frame frequency response information of the target FF filter. The optimal value of the variable is determined by the gradient descent method, so that the k-th frame filter coefficient of the target FF filter is determined by the optimal value of the variable.
  • the filter coefficients of each frame of the target FF filter are determined according to the gradient descent method.
  • a value of a loss function is determined.
  • the value of the loss function reaches the minimum threshold value, it is determined that the filter coefficients of the target FF filter have reached the convergence stability condition.
  • the filter coefficients of the kth frame of the target FF filter when the value of the loss function between the filter coefficient variable and the kth frame frequency response information of the target FF filter reaches the minimum threshold value, it is determined that the filter coefficients of the kth frame of the target FF filter have reached the convergence stability condition.
  • the value of the loss function does not reach the minimum threshold value, it is determined that the filter coefficients of the kth frame of the target FF filter have not reached the convergence stability condition.
  • the minimum threshold value is set in advance and can be adjusted according to different needs in different situations.
  • the filter coefficient of each FF filter includes at least one biquad filter coefficient and a gain.
  • the variables corresponding to the biquad filter coefficient include filter type, cutoff frequency and quality factor.
  • the filter coefficient of each FF filter may also include other more or fewer parameters, which is not limited in the embodiments of the present application.
  • the k-th frame filter coefficient of the first FF filter can be determined based on the value of the filter coefficient variable according to a relevant algorithm, and the embodiment of the present application does not limit the algorithm.
  • a quiet environment has a background noise problem, that is, background noise.
  • background noise For example, for semi-open headphones, the headphones are more prone to background noise problems in a quiet environment than in-ear headphones.
  • a quiet environment does not require strong noise reduction, and some people will feel uncomfortable with strong noise reduction in a quiet environment. And the greater the noise reduction intensity, the stronger the negative pressure people feel. Therefore, when the value of the filter coefficient variable is determined by the gradient descent method, the target noise reduction amplitude can be dynamically adjusted based on the ambient volume, so as to determine the k-th frame filter coefficient of the target FF filter according to the target noise reduction amplitude, thereby improving the subjective experience effect of adaptive noise reduction.
  • the target noise reduction amplitude is determined according to the ambient volume of the k-1th frame and the ambient volume of the t-frame before the k-1th frame, where t is greater than or equal to 1 and less than k-1.
  • the value of the filter coefficient variable is determined by the gradient descent method, and the k-th frame filter coefficient of the target FF filter is determined based on the value of the filter coefficient variable.
  • the target ambient volume is determined according to the ambient volume of the k-1th frame and the ambient volume of the t-frame before the k-1th frame. If the target ambient volume is less than or equal to the first volume threshold, the first noise reduction amplitude is determined as the target noise reduction amplitude. If the target ambient volume is greater than the first volume threshold, it is determined whether the target ambient volume is significantly increased or significantly decreased. If the target ambient volume is significantly increased, the noise reduction amplitude of the k-1th frame is increased to obtain the target noise reduction amplitude. If the ambient volume is significantly decreased, the noise reduction amplitude of the k-1th frame is reduced to obtain the target noise reduction amplitude. If the target ambient volume is neither significantly increased nor significantly decreased, the noise reduction amplitude of the k-1th frame is determined as the target noise reduction amplitude, that is, the noise reduction amplitude is maintained unchanged.
  • the target ambient volume based on the ambient volume of the k-1th frame and the ambient volume of the t-frame before the k-1th frame, such as taking the arithmetic mean, the weighted mean, etc., which are not limited in the embodiments of the present application.
  • the t-frame can be any t-frame before the k-1th frame, or it can be a t-frame before the k-1th frame and closest to the k-1th frame, which are not limited in the embodiments of the present application.
  • the first volume threshold is set in advance, and the first volume threshold is used to characterize whether you are currently in a quiet environment. That is, if the target environment volume is less than or equal to the first volume threshold, it indicates that you are in a quiet environment. If the target environment volume is greater than the first volume threshold, it indicates that you are in a non-quiet environment.
  • the first noise reduction amplitude is set in advance for a quiet environment and is used for weak noise reduction to avoid over-amplifying the background noise or introducing subjective comfort issues. In actual applications, the first volume threshold and the first noise reduction amplitude can be adjusted according to different needs.
  • the target environment volume is used to determine whether the environment is in a quiet environment.
  • the first noise reduction amplitude is determined as the target noise reduction amplitude.
  • the target environment volume increases significantly, the noise reduction amplitude of the k-1th frame is increased to obtain the target noise reduction amplitude.
  • the target environment volume decreases significantly, the noise reduction amplitude of the k-1th frame is reduced to obtain the target noise reduction amplitude.
  • the target environment volume does not increase significantly and does not decrease significantly, the noise reduction amplitude of the k-1th frame is determined as the target noise reduction amplitude, that is, the noise reduction amplitude is maintained unchanged.
  • the target environment volume determined this time is greater than the target environment volume determined last time, and the difference between the target environment volume determined this time and the target environment volume determined last time is greater than the second volume threshold, it is determined that the target environment volume determined this time has increased significantly.
  • the target environment volume determined this time is less than the target environment volume determined last time, and the difference between the target environment volume determined this time and the target environment volume determined last time is greater than the second volume threshold, it is determined that the target environment volume determined this time has decreased significantly.
  • the second volume threshold is also set in advance, such as 3dB. In practical applications, the second volume threshold can also be adjusted according to different requirements.
  • the earphone also includes the multiple FB filters. If the target FF filter is the first FF filter, the k-th frame filter coefficient of the target FF filter is determined based on the k-1th frame reference signal collected by the target reference microphone, the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficients of the multiple SPs, and the k-1th frame filter coefficients of the multiple FB filters.
  • the k-th frame filter coefficient of the target FF filter is determined based on the k-1th frame reference signal collected by the target reference microphone, the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficients of the multiple SPs, the k-1th frame filter coefficients of the multiple FB filters, and the k-th frame frequency response information and the k-1th frame frequency response information of each FF filter located before the target FF filter.
  • the residual error may be determined based on the k-1th frame reference signal collected by the target reference microphone and the k-1th frame error signal collected by the error microphone.
  • the k-th frame frequency response information of the target FF filter is determined based on the k-1th frame frequency response information of the target FF filter, the residual error, the k-1th frame filter coefficients of the multiple FB filters, and the k-1th frame filter coefficients of the multiple SPs.
  • the k-th frame filter coefficient of the target FF filter is determined based on the k-th frame frequency response information of the target FF filter.
  • the frequency response information of the k-1th frame filter coefficients of the multiple FB filters and the frequency response information of the k-1th frame filter coefficients of the multiple SPs can be determined, and then, based on the k-1th frame frequency response information of the target FF filter, the residual error, the frequency response information of the k-1th frame filter coefficients of the multiple FB filters, and the frequency response information of the k-1th frame filter coefficients of the multiple SPs, the k-1th frame frequency response information of the target FF filter is determined according to the following formula (4).
  • FF 1,k refers to the k-th frame frequency response information of the target FF filter
  • FF 1,k-1 refers to the k-1-th frame frequency response information of the target FF filter
  • Res 1,k-1 refers to the residual error
  • SP 1,k-1 refers to the frequency response information of the k-1-th frame filter coefficient of the target SP
  • FB j,k-1 refers to the frequency response information of the k-1-th frame filter coefficient of the j-th FB filter among the multiple FB filters
  • SP j,k-1 refers to the frequency response information of the k-1-th frame filter coefficient of the SP corresponding to the j-th FB filter
  • n refers to the total number of the multiple FB filters, that is, the total number of the multiple noise reduction channels.
  • the residual error may be determined based on the k-1th frame reference signal collected by the target reference microphone and the k-1th frame error signal collected by the error microphone.
  • the k-1th frame frequency response information of the target FF filter is determined based on the k-1th frame frequency response information of the target FF filter, the residual error, the k-1th frame filter coefficients of the multiple SPs, the k-1th frame filter coefficients of the multiple FB filters, and the k-1th frame frequency response information and the k-1th frame frequency response information of each FF filter located before the target FF filter.
  • the k-1th frame filter coefficient of the target FF filter is determined based on the k-1th frame frequency response information of the target FF filter.
  • the frequency response information of the k-1th frame filter coefficients of the multiple SPs and the frequency response information of the k-1th frame filter coefficients of the multiple FB filters may be determined, and then, based on the k-1th frame frequency response information of the target FF filter, the residual error, the frequency response information of the k-1th frame filter coefficients of the multiple SPs, the frequency response information of the k-1th frame filter coefficients of the multiple FB filters, and the k-th frame frequency response information and the k-1th frame frequency response information of each FF filter located before the target FF filter, the k-th frame frequency response information of the target FF filter is determined according to the following formula (5).
  • FB j,k-1 refers to the frequency response information of the k-1th frame filter coefficient of the jth FB filter among the multiple FB filters
  • n refers to the total number of the multiple FB filters, that is, the total number of the multiple noise reduction channels, and the meanings of other letters are the same as those in the above formula (3).
  • the implementation process of determining the k-th frame filter coefficient of the target FF filter based on the k-th frame frequency response information of the target FF filter is the same as the first case above.
  • the frequency response information of the filter coefficient of the above SP can be determined based on the filter coefficient of the SP according to the relevant algorithm
  • the frequency response information of the FB filter coefficient can also be determined based on the filter coefficient of the FB filter according to the relevant algorithm.
  • the embodiment of the present application does not limit the algorithm.
  • the k-th frame frequency response information of the target FF filter is determined based on the k-1-th frame filter coefficient of the target SP, and the k-1-th frame filter coefficient of the target SP is determined based on the target noise reduction gear by querying the mapping relationship between the noise reduction gear and the filter coefficient of the SP, that is, the k-1-th frame filter coefficient of the target SP is an estimated value.
  • the filter coefficient of the FF filter can be adapted.
  • the initial filter coefficients of the multiple FB filters are determined as the k-th frame filter coefficients of the multiple FB filters, that is, the first frame filter coefficients of the multiple FB filters are the initial filter coefficients of the corresponding FB filters, or the k-th frame filter coefficients of the multiple FB filters are determined based on the initial noise reduction gear and the mapping relationship between the noise reduction gear and the FB filter coefficients.
  • the k-th frame filter coefficients of the multiple FB filters can be determined based on the target noise reduction gear.
  • initial filter coefficients of the multiple FB filters may be the same or different, and the initial filter coefficient may be 0 or not 0, which is not limited in the embodiment of the present application.
  • the process of determining the k-th frame filter coefficient of each FB filter based on the target noise reduction gear is the same, one of them is used as an example for introduction. That is, one of the multiple FB filters is used as the target FB filter, and the k-th frame filter coefficient of the target FB filter is determined in the following two ways.
  • the process of determining the k-th frame filter coefficients of other FB filters in the multiple FB filters can refer to the process of determining the k-th frame filter coefficient of the target FB filter. In other words, when k is greater than 1, the k-th frame filter coefficient of the target FB filter can be determined in the following two ways.
  • the first method is to determine the target FB filter's first position based on the target noise reduction gear and the mapping relationship between the noise reduction gear and the FB filter coefficient. k frame filter coefficients.
  • the mapping relationship between the noise reduction gear and the FB filter coefficient includes multiple noise reduction gears, each noise reduction gear has a mapping relationship with the filter coefficients of the multiple FB filters, and the mapping relationship between different noise reduction gears and the filter coefficients of the multiple FB filters may be different. Therefore, based on the target noise reduction gear, the filter coefficient corresponding to the target FB filter can be obtained from the mapping relationship between the noise reduction gear and the FB filter coefficient, and the obtained filter coefficient is used as the k-th frame filter coefficient of the target FB filter.
  • the k-th frame filter coefficient of the target FB filter determined by the first method has better stability, simpler operation and higher efficiency.
  • the second method is that if the target FB filter belongs to the first type of FB filter, the k-th frame filter coefficient of the target FB filter is determined based on the target noise reduction gear and the mapping relationship between the noise reduction gear and the FB filter coefficient. If the target FB filter belongs to the second type of FB filter, the k-1-th frame filter coefficient of the target FB filter is determined based on the k-1-th frame error signal collected by the error microphone, the k-1-th frame filter coefficient of the target FB filter and the target noise reduction gear.
  • the k-th frame filter coefficient of the target FB filter can be determined by an adaptive method.
  • the process of determining the k-th frame filter coefficient of the target FB filter is an adaptive process, which can also be called an iterative process.
  • the implementation process of determining the k-1th frame filter coefficient of the target FB filter based on the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficient of the target FB filter and the target noise reduction gear includes: determining the k-1th frame filter coefficient of the target SP based on the target noise reduction gear and the mapping relationship between the noise reduction gear and the filter coefficient of the SP, the target SP refers to the path from the first speaker corresponding to the target FB filter to the error microphone; determining the k-1th frame filter coefficient of the target FB filter based on the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficient of the target FB filter and the k-1th frame filter coefficient of the target SP.
  • the k-th frame filter coefficient of the target FB filter can be determined according to the relevant algorithm based on the k-1 frame error signal collected by the error microphone, the k-1 frame filter coefficient of the target FB filter and the k-1 frame filter coefficient of the target SP.
  • the embodiment of the present application does not limit the algorithm.
  • the first frame filter coefficient of the target FB filter can be the initial filter coefficient, or can be determined by querying the mapping relationship between the noise reduction gear and the FB filter coefficient through the initial noise reduction gear, when k is greater than or equal to 1, it is equivalent to determining the kth frame filter coefficient of the target FB filter in three ways. That is, (1) the kth frame filter coefficient of the target FB filter is determined by querying the mapping relationship between the noise reduction gear and the FB filter coefficient. (2) If the target FB filter belongs to the first type of FB filter, the kth frame filter coefficient of the target FB filter is determined by querying the mapping relationship between the noise reduction gear and the FB filter coefficient.
  • the kth frame filter coefficient of the target FB filter is determined based on the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficient of the target FB filter, and the target noise reduction gear. (3) If the target FB filter belongs to the first type of FB filter, or the target FB filter belongs to the second type of FB filter and k is equal to 1, the kth frame filter coefficient of the target FB filter is determined by querying the mapping relationship between the noise reduction gear and the FB filter coefficient.
  • the k-th frame filter coefficient of the target FB filter is determined based on the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficient of the target FB filter, and the target noise reduction gear.
  • the sound frequency band of the first speaker corresponding to the first type of FB filter is higher than the sound frequency band of the first speaker corresponding to the second type of FB filter. That is, the first speaker corresponding to the first type of FB filter is a high-frequency speaker, and the first speaker corresponding to the second type of FB filter is a low-frequency speaker.
  • the first type of FB filter and the second type of FB filter can also be distinguished in other ways instead of according to the sound frequency band, and the embodiment of the present application is also not limited to this.
  • the second and third methods mentioned above combine the method of querying the mapping relationship between the noise reduction gear and the FB filter coefficient with the adaptive method, which can not only improve the noise reduction effect, but also has relatively low complexity and relatively controllable stability.
  • the k-th frame filter coefficient of the target FB filter can be determined not only in the above three ways, but also in other ways. For example, regardless of whether the target FB filter belongs to the first type of FB filter or the second type of FB filter, the k-th frame filter coefficient of the target FB filter is determined based on the k-1 frame error signal collected by the error microphone, the k-1 frame filter coefficient of the target FB filter, and the target noise reduction gear. The embodiment of the present application does not limit this.
  • the initial downlink compensation filter coefficient is determined as the k-th frame filter coefficient of the downlink compensation filter, or the k-th frame filter coefficient of the downlink compensation filter is determined based on the initial noise reduction gear and the mapping relationship between the noise reduction gear and the downlink compensation filter coefficient.
  • the downlink compensation filter is determined based on the target noise reduction gear and the mapping relationship between the noise reduction gear and the downlink compensation filter coefficient. The filter coefficients of the kth frame of the filter.
  • the mapping relationship between the noise reduction gear and the downlink compensation filter coefficient includes multiple noise reduction gears, each noise reduction gear has a mapping relationship with the filter coefficient of the downlink compensation filter, and the mapping relationship between different noise reduction gears and the filter coefficient of the downlink compensation filter may be different. Therefore, after determining the target noise reduction gear, based on the target noise reduction gear, the corresponding downlink compensation filter coefficient can be obtained from the mapping relationship between the noise reduction gear and the downlink compensation filter coefficient, and the obtained downlink compensation filter coefficient is used as the kth frame filter coefficient of the downlink compensation filter.
  • Determine the noise reduction level of the k-1th frame obtain the noise reduction levels of m frames before the k-1th frame, where m is greater than or equal to 1 and less than k-1. Determine the target noise reduction level based on the noise reduction level of the k-1th frame and the noise reduction levels of the m frames.
  • the k-1th frame may or may not have a valid downlink signal, and may be in a quiet environment or a non-quiet environment, and of course, there may be abnormal signals, the methods for determining the noise reduction level of the k-1th frame are different in different situations, which will be introduced below.
  • the noise reduction gear of the k-1th frame is determined based on the reference filter coefficients of the multiple FF filters and the mapping relationship between the noise reduction gear and the frequency response information of the FF filter.
  • the reference filter coefficient is the initial filter coefficient of the corresponding FF filter
  • the reference filter coefficient is the filter coefficient of the corresponding FF filter that most recently reached the convergence stability condition before the kth frame, or the filter coefficient of the k-1th frame of the corresponding FF filter.
  • the user terminal When playing an audio signal through the headset, such as playing music or making a call, the user terminal will send a control signaling of playing the audio signal to the headset. Therefore, whether the headset receives the control signaling can be used to determine whether it is currently in a downlink enabled state.
  • the headset When the headset is not in a downlink enabled state, it is determined that there is no valid downlink signal in the k-1 frame.
  • the k-1 frame When the headset is in a downlink enabled state, the k-1 frame does not necessarily have continuous sound output. For example, there will be no sound output during the pause period of speaking, the transition period of changing music, etc., and this time is often not necessarily short. Therefore, when the headset is in a downlink enabled state, it can also be determined whether the k-1 frame is in a downlink intermittent period.
  • the k-1 frame is in a downlink intermittent period, it is determined that there is no valid downlink signal in the k-1 frame. If the k-1 frame is not in a downlink intermittent period, it is determined that there is a valid downlink signal in the k-1 frame.
  • the noise reduction level of the k-1th frame may vary with the environmental noise. Therefore, it is necessary to determine the noise reduction level of the k-1th frame based on the reference filter coefficients of the multiple FF filters and the mapping relationship between the noise reduction level and the frequency response information of the FF filters.
  • the reference frequency response information of the multiple FF filters is determined. Based on the mapping relationship between the noise reduction level and the frequency response information of the FF filters, the noise reduction levels that match the reference frequency response information of the multiple FF filters are determined to obtain multiple reference noise reduction levels. Based on the multiple reference noise reduction levels, the noise reduction level of the k-1th frame is determined.
  • the reference frequency response information of the multiple FF filters can be determined according to a relevant algorithm based on the reference filter coefficients of the multiple FF filters, and the embodiment of the present application does not limit the algorithm.
  • the frequency response information of the FF filter may also be different, so the mapping relationship between the noise reduction gear and the frequency response information of the FF filter can be stored in advance.
  • the reference frequency response information of the FF filter is matched with the frequency response information of the FF filter under different noise reduction gears in the mapping relationship, so as to determine the frequency response information matching the reference frequency response information of the FF filter from the mapping relationship, and then use the noise reduction gear corresponding to the matched frequency response information as a reference noise reduction gear.
  • the same processing is performed, so that multiple reference noise reduction gears can be obtained.
  • the frequency response information of the FF filter can be represented by a frequency response curve. Therefore, after determining the reference frequency response curves of the multiple FF filters, for any FF filter among the multiple FF filters, the reference frequency response curve of the FF filter is matched with the frequency response curve of the FF filter at different noise reduction levels in the mapping relationship.
  • the complete reference frequency response curve of the FF filter can be matched with the complete frequency response curve of the FF filter at different noise reduction gears in the mapping relationship.
  • the curve in the reference frequency response curve of the FF filter located in the target frequency band can also be matched with the curve in the frequency response curve of the FF filter at different noise reduction gears in the mapping relationship located in the target frequency band, which is not limited in the embodiments of the present application.
  • the target frequency band refers to the frequency band with obvious characteristics in the frequency response curve.
  • the target frequency band is set in advance.
  • the target frequency band is the frequency band from 100 to 200 Hz.
  • the value of the target frequency band may also be different under different acoustic conditions of the headphones.
  • the mapping relationship between the noise reduction gear and the frequency response information of the FF filter includes the frequency response curves of the FF filter under 16 noise reduction gears, and the frequency response curves of the FF filter under the 16 noise reduction gears are shown in FIG4. Since the characteristics in the frequency band of 100 to 200 Hz in FIG4 are clearly distinguished, 100 to 200 Hz is used as the target frequency band. Then, the curve located within 100 to 200 Hz in the reference frequency response curve of the FF filter is matched with the curve located within 100 to 200 Hz in the frequency response curve of the FF filter under the 16 noise reduction gears.
  • the noise reduction level of the k-1th frame based on the multiple reference noise reduction levels, for example, determining the noise reduction level of the k-1th frame according to the average value of the multiple reference noise reduction levels, or determining the noise reduction level of the k-1th frame according to the reference noise reduction level with the largest number among the multiple reference noise reduction levels.
  • the average value of the multiple reference noise reduction gears can be directly determined as the noise reduction gear of the k-1th frame, or the average value of the multiple reference noise reduction gears can be adjusted to obtain the noise reduction gear of the k-1th frame.
  • the reference noise reduction gear with the largest number among the multiple reference noise reduction gears can be directly determined as the noise reduction gear of the k-1th frame, or the reference noise reduction gear with the largest number among the multiple reference noise reduction gears can be adjusted to obtain the noise reduction gear of the k-1th frame.
  • the embodiments of the present application are not limited to this.
  • the above-mentioned convergence stability condition refers to the filter coefficient of the FF filter converging to basically unchanged.
  • the filter coefficient of the FF filter may be adaptively adjusted multiple times during the entire noise reduction process, when determining the filter coefficient of the kth frame of the FF filter, the filter coefficient of the FF filter that most recently reached the convergence stability condition before the kth frame can be used as the reference filter coefficient, and the filter coefficient of the k-1th frame of the FF filter can also be used as the reference filter coefficient.
  • the noise reduction gear of the k-1th frame is determined based on the valid downlink signal of the k-1th frame, the reference signal of the k-1th frame collected by the at least one reference microphone, and the error signal of the k-1th frame collected by the error microphone.
  • the headset when the headset is in a downlink enabled state and is not in a downlink intermittent period, it is determined that there is a valid downlink signal in the k-1th frame.
  • the valid downlink signal can be extracted from the k-1th frame error signal collected by the error microphone, thereby determining the k-1th frame noise reduction gear based on the extracted valid downlink signal.
  • the effective downlink signal can be extracted from the k-1th frame error signal collected by the error microphone, so as to determine the k-1th frame noise reduction gear based on the extracted effective downlink signal.
  • the embodiment of the present application does not limit the algorithm.
  • the noise reduction level of the k-3 frame is determined to be the noise reduction level of the k-1 frame. That is, the noise reduction level is maintained unchanged.
  • the noise has basically not changed, and the noise reduction level can be maintained unchanged.
  • the noise reduction level is maintained unchanged to perform robust control, thereby avoiding the divergence of the noise reduction level.
  • Abnormal noise signals refer to signals that have a relatively serious impact on the user's listening experience, such as howling, clipping, background noise, wind noise, etc.
  • howling is the phenomenon that the amplitude or energy of a single-frequency sound signal suddenly increases from a small value. It is usually caused by the earphone being squeezed or the user quickly changing the wearing posture of the earphone. The sound signal emitted during howling is called howling noise. Howling will cause discomfort to the user and interfere with the playback of the downlink signal, seriously affecting the audio playback effect.
  • Clipping is the phenomenon that low-frequency signals overflow and produce crackling noises. The crackling noises produced are called clipping noise.
  • Background noise is the background noise, which can also be called background noise.
  • Background noise is noise caused by the performance limitations of the device's hardware (such as circuits or other components in the earphone), such as the rustling sound in the TV sound in addition to the program sound.
  • the background noise In a noisy environment, the background noise is generally not perceived or heard by the user. When the environment is quiet, the user can perceive the background noise. Too much background noise will not only make people irritable, but also drown out the weaker details in the sound.
  • Wind noise is the whirring sound produced when there is wind in the environment. Wind noise affects the normal use of headphones by users. And because the direction of wind noise is relatively random, the impact of wind noise on the user's ears is different, that is, the left ear and the right ear have inconsistent hearing under the influence of wind noise.
  • the noise reduction gear of the k-1th frame is determined based on the valid downlink signal of the k-1th frame, the k-1th frame reference signal collected by the at least one reference microphone, and the k-1th frame error signal collected by the error microphone.
  • the noise reduction level of the k-1th frame After the noise reduction level of the k-1th frame is determined through the above three situations, the noise reduction level of the k-1th frame and the noise reduction levels of m frames before the k-1th frame may be combined to determine the target noise reduction level.
  • the m-frame noise reduction gear can be any m-frame noise reduction gear located before the k-1 frame, or it can be the m-frame noise reduction gear located before the k-1 frame and closest to the k-1 frame, and the embodiment of the present application does not limit this.
  • the noise reduction effect is evaluated according to the relevant algorithm to determine the noise reduction probability corresponding to the k-1 frame noise reduction gear and the noise reduction probability corresponding to the m-frame noise reduction gear respectively, and the noise reduction gear with the largest noise reduction probability is determined as the target noise reduction gear.
  • the arithmetic mean or weighted average of the k-1 frame noise reduction gear and the m-frame noise reduction gear is determined to obtain the target noise reduction gear.
  • the noise reduction gear with the largest number of occurrences in the k-1 frame noise reduction gear and the m-frame noise reduction gear is determined as the target noise reduction gear, and so on.
  • mapping relationships involved above are all determined in advance. For example, when one of the multiple first speakers is working and the other first speakers are not working, the mapping relationships are determined based on the reference signal collected by the at least one reference microphone and the error signal collected by the error microphone in each of the multiple leakage states.
  • the multiple leakage states are formed by the earphones and a variety of different ear canal environments, and the multiple leakage states correspond one-to-one to a plurality of noise reduction gears.
  • Figure 6 a simple exemplary summary of a determination process of the multiple sets of target noise reduction parameters is given through Figure 6. Please refer to Figure 6, and the initial value can be set offline, including the above-mentioned initial noise reduction gear, initial filter coefficient and various mapping relationships. Then determine whether there is a valid downlink signal in the k-1th frame, whether it is in a quiet environment, and whether there is an abnormal noise signal, so as to determine the noise reduction gear of the k-1th frame based on different situations.
  • the target noise reduction gear is determined by the noise reduction gear of the k-1th frame and the noise reduction gear of the previous m frames.
  • the target noise reduction amplitude is determined based on the ambient volume of the k-1th frame
  • the FB filter coefficient is adapted based on the target noise reduction gear to determine the kth frame filter coefficient of the multiple FB filters.
  • the FF filter coefficient is adapted based on the target noise reduction gear and the target noise reduction amplitude to determine the kth frame filter coefficient of the multiple FF filters.
  • Step 202 Based on the multiple groups of target noise reduction parameters, multiple groups of target reverse phase noise corresponding to the multiple first speakers are generated, and the frequency band of each target reverse phase noise in the multiple groups of target reverse phase noise covers the sound frequency band of the multiple first speakers.
  • the multiple sets of target noise reduction parameters can be referred to as noise reduction parameters of multiple noise reduction channels, and thus, the multiple sets of target anti-phase noise generated can also be referred to as anti-phase noise of multiple noise reduction channels. Since the generation process of the anti-phase noise of each noise reduction channel is the same, one of the noise reduction channels is used as an example for explanation below.
  • the target noise reduction channel includes a target FF filter and a target first speaker, and the reference microphone corresponding to the target FF filter is called a target reference microphone.
  • the target anti-phase noise includes feedforward anti-phase noise. That is, based on the k-th frame filter coefficient of the target FF filter, the k-th frame reference signal collected by the target reference microphone is processed to obtain the feedforward anti-phase noise.
  • the target reference microphone may include one reference microphone or may include at least two reference microphones.
  • the k-th frame reference signal collected by the target reference microphone can be directly processed based on the k-th frame filter coefficient of the target FF filter to obtain feedforward inverted noise.
  • the target reference microphone includes at least two reference microphones, the k-th frame reference signal collected by the at least two reference microphones is mixed to obtain the k-th frame mixed reference signal, and then, based on the k-th frame filter coefficient of the target FF filter, the k-th frame mixed reference signal is processed to obtain feedforward inverted noise.
  • the target noise reduction channel also includes a target FB filter.
  • the target anti-phase noise also includes feedback anti-phase noise. That is, based on the k-th frame filter coefficient of the downlink compensation filter, the k-th frame downlink signal sent by the user terminal is downlink compensated. Then, the k-th frame downlink signal after downlink compensation is inverted and mixed with the k-th frame error signal collected by the error microphone to obtain the k-th frame noise signal collected by the error microphone. Based on the k-th frame filter coefficient of the target FB filter, the k-th frame noise signal collected by the error microphone is processed to obtain feedback anti-phase noise.
  • one frame may include one sample point or multiple sample points
  • a group of target inverted noise can be generated at each sample point, or a group of target inverted noise can be generated in one frame.
  • the embodiment of the present application does not perform frequency division on the downlink signal when determining multiple sets of target noise reduction parameters, that is, the multiple sets of target noise reduction parameters are determined by the full-frequency downlink signal.
  • the frequency band of each target anti-phase noise in the multiple sets of target anti-phase noises covers the sound frequency band of the multiple first speakers, that is, the frequency band of each target anti-phase noise is full-band.
  • Step 203 Utilize the multiple groups of target anti-phase noises to perform noise reduction through the multiple first speakers.
  • the multiple groups of target anti-phase noises are mixed with the k-th frame downlink signals to be played by the multiple first speakers respectively, and then played through the corresponding first speakers, thereby achieving the purpose of noise reduction.
  • Some of the multiple first speakers may be high-frequency speakers, and others may be low-frequency speakers. Alternatively, some may be full-range speakers, and others may be non-full-range speakers. That is, the sound frequency bands of the multiple first speakers may be different. Alternatively, all of the multiple first speakers are full-range speakers. Alternatively, all of the multiple first speakers are non-full-range speakers. In the case where all of the multiple first speakers are full-range speakers, the k-th frame downlink signal to be played by the multiple first speakers is the k-th frame downlink signal sent by the user terminal.
  • Two of the plurality of first speakers may include two first speakers formed by a double-diaphragm (or double-dynamic) speaker, or the plurality of first speakers may include a plurality of speakers of split speakers.
  • the headset may also include at least one second speaker, which does not participate in noise reduction.
  • the second speaker may participate in downlink compensation (i.e., downlink compensation is performed on the downlink signal sent by the user terminal, and the downlink signal is a full-band audio signal, including the audio signal of the sound frequency band of the second speaker).
  • the first speaker may be a medium-low frequency speaker, or a full-band speaker
  • the second speaker may be a high-frequency speaker, or a medium-frequency speaker or a low-frequency speaker.
  • the second speaker may not participate in downlink compensation.
  • the first speaker may be a medium-low frequency speaker, or a full-band speaker
  • the second speaker may be a high-frequency speaker.
  • the sound frequency band of the at least one second speaker is higher than the sound frequency band of the at least one first speaker.
  • the sound frequency band of the at least one second speaker may also be lower than the sound frequency band of the at least one first speaker, and this embodiment of the application does not limit this.
  • the process of determining the multiple sets of target noise reduction parameters by the adaptive method mentioned above takes a certain amount of time, when a frame includes multiple sample points and the duration of a frame is long, the determination duration of the multiple sets of target noise reduction parameters will be less than the duration of a frame. Therefore, through the relevant data of the k-1th frame, calculations can be performed in a partial time period starting from the kth frame to obtain multiple sets of target noise reduction parameters for the kth frame, and then active noise reduction is performed in a partial time period subsequent to the kth frame according to the multiple sets of target noise reduction parameters of the kth frame.
  • the determination duration of the multiple sets of target noise reduction parameters may be equal to the duration of a frame.
  • the multiple sets of target noise reduction parameters can be determined as multiple sets of target noise reduction parameters for the k+1th frame, and then active noise reduction is performed in the time period of the k+1th frame according to the multiple sets of target noise reduction parameters of the k+1th frame.
  • each target anti-phase noise in the multiple groups of target anti-phase noises covers the sound frequency band of the multiple first speakers, that is, each target anti-phase noise is a full-band anti-phase noise, so no matter whether the first speaker is a high-frequency speaker, a low-frequency speaker or a full-range speaker, when using the multiple groups of target anti-phase noises for noise reduction, the noise reduction ability of each first speaker can be fully utilized.
  • this solution can improve the noise reduction effect of the headphone through the full-band anti-phase noise of multiple noise reduction channels.
  • FIG7 is a schematic diagram of the structure of an earphone provided by an embodiment of the present application.
  • the earphone includes f reference microphones, an error microphone, n FF filters, n FF adaptive engines corresponding to the n FF filters, n FB filters, n FB adaptive engines corresponding to the n FB filters, n first speakers (i.e., speakers 1 to n), and a downlink compensation filter.
  • a downlink compensation adaptive engine (not shown in the figure), a digital frequency divider, and n equalizers (EQ) calibrators.
  • f and n are both integers greater than or equal to 1, and f and n may be equal or unequal.
  • the f reference microphones are used to collect noise signals from the external environment, i.e., reference signals.
  • the error microphone is used to collect noise signals in the ear canal, i.e., error signals.
  • the n FF adaptive engines are used to determine the k-th frame filter coefficients of the corresponding FF filters, and refresh the determined k-th frame filter coefficients into the corresponding FF filters.
  • the n FB adaptive engines are used to determine the k-th frame filter coefficients of the corresponding FB filters, and refresh the determined k-th frame filter coefficients into the corresponding FB filters.
  • the downlink compensation adaptive engine is used to determine the k-th frame filter coefficients of the downlink compensation filter, and refresh the determined k-th frame filter coefficients into the downlink compensation filter.
  • the digital frequency divider is used to divide the k-th frame downlink signal sent by the user terminal according to the sound frequency band of the n first speakers to obtain the k-th frame downlink signal corresponding to each first speaker.
  • the n EQ calibrators are used to calibrate the mass production parameters of the corresponding first speakers to align the tolerance consistency of the mass production parameters of the n first speakers.
  • the n FF filters are used to process the kth frame reference signal collected by the corresponding reference microphone based on their respective kth frame filter coefficients to obtain feedforward inverted noise.
  • the downlink compensation filter is used to perform downlink compensation on the kth frame downlink signal sent by the user terminal based on its own kth frame filter coefficient.
  • the kth frame downlink signal after downlink compensation is inverted and mixed with the kth frame error signal collected by the error microphone to obtain the kth frame noise signal collected by the error microphone.
  • the n FB filters are used to process the kth frame noise signal collected by the error microphone based on their respective kth frame filter coefficients to obtain feedback inverted noise.
  • the feedforward inverted noise, feedback inverted noise, and kth frame downlink signal of the first speaker of each noise reduction channel are mixed and played through the corresponding first speaker to achieve noise reduction.
  • FIG8 is a schematic diagram of the structure of another headset provided in an embodiment of the present application.
  • the headset includes a reference microphone, an error microphone, two FF filters, two FF adaptive engines corresponding to the two FF filters, two FB filters, two FB adaptive engines corresponding to the two FB filters, two first speakers (i.e., speaker 1 and speaker 2), a downlink compensation filter, a downlink compensation adaptive engine (not shown in the figure), and two EQ calibrators.
  • the two FF filters correspond to the reference microphone, and the two first speakers are speakers formed by dual-diaphragm (or dual-dynamic) speakers.
  • the headset may not include a digital crossover.
  • the two first speakers can be considered as a combination of two dynamic speakers but with a common magnetic circuit, and are physically one speaker. Since the two dynamic speakers have good sounding capabilities in the full frequency band, they can be regarded as the superposition of two full-frequency ANC noise reduction modules, so that the noise reduction capabilities of the speakers can be fully utilized.
  • FIG9 is a schematic diagram of the structure of another headset provided in an embodiment of the present application.
  • the headset includes a reference microphone, an error microphone, two FF filters, two FF adaptive engines corresponding to the two FF filters, two FB filters, two FB adaptive engines corresponding to the two FB filters, two first speakers (i.e., speaker 1 and speaker 2), a downstream compensation filter, a downstream compensation adaptive engine (not shown in the figure), a digital crossover, and two EQ calibrators.
  • the difference from FIG8 is that there are two physical entities of the first speaker, and the two split first speakers can be different or the same.
  • each first speaker can be designed as a full-range noise reduction unit when viewed independently, so that the maximum noise reduction capability of each speaker is fully utilized.
  • Figure 10 is a schematic diagram of the structure of another headset provided in an embodiment of the present application.
  • the headset includes two reference microphones, an error microphone, two FF filters, two FF adaptive engines corresponding to the two FF filters, two FB filters, two FB adaptive engines corresponding to the two FB filters, two first speakers (i.e., speaker 1 and speaker 2), a second speaker (i.e., speaker 3), a downlink compensation filter, a downlink compensation adaptive engine (not shown in the figure), a digital divider, and three EQ calibrators.
  • Figure 10 uses three speakers to achieve high-definition sound quality requirements. The frequency responses of the three speakers focus on low, medium, and high audio bands, respectively.
  • the first speaker may be a medium-low frequency speaker or a full-frequency speaker
  • the second speaker may be a high-frequency speaker or a medium-frequency speaker or a low-frequency speaker.
  • the second speaker may not participate in the downlink compensation (that is, after the downlink signal sent by the user terminal is digitally divided, the downlink signals corresponding to the two first speakers are obtained, and the downlink signals corresponding to the two first speakers are downlink compensated, excluding the downlink signal corresponding to the second speaker).
  • the first speaker can be a medium and low frequency speaker or a full-range speaker
  • the second speaker is a high-frequency speaker.
  • the crossover point of the high-frequency speaker can be above 6kHz, that is, the audio signals above 6kHz are not compensated.
  • the embodiment of the present application does not limit the crossover frequency point of 6 kHz, and it can also be other high-frequency crossover frequency points.
  • the FF adaptive engine, FB adaptive engine, and downlink compensation adaptive engine mentioned above can be deployed on a microcontroller.
  • the FF filter, FB filter, and downlink compensation filter can be deployed on an ANC chip.
  • the microcontroller and the ANC chip can be collectively referred to as a noise reduction processor.
  • the microcontroller and the ANC chip can be integrated on one chip or deployed on two chips.
  • FIG12 is a schematic diagram of the structure of a noise reduction device provided in an embodiment of the present application, and the noise reduction device can be implemented by software, hardware or a combination of both to form part or all of a headset, and the headset can be the headset shown in FIG1.
  • the device includes: a noise reduction parameter determination module 1201, an anti-phase noise generation module 1202 and a noise reduction module 1203.
  • a noise reduction parameter determination module 1201 is used to determine multiple groups of target noise reduction parameters corresponding to multiple first speakers one by one;
  • the anti-phase noise generating module 1202 is used to generate multiple groups of target anti-phase noises corresponding to the multiple first speakers based on the multiple groups of target noise reduction parameters, wherein the frequency band of each target anti-phase noise in the multiple groups of target anti-phase noises covers the sound frequency band of the multiple first speakers;
  • the noise reduction module 1203 is used to use multiple groups of target anti-phase noise to perform noise reduction through multiple first speakers.
  • the headset further includes a plurality of FF filters corresponding one-to-one to the plurality of first speakers, and the plurality of sets of target noise reduction parameters include k-th frame filter coefficients of the plurality of FF filters, where k is an integer greater than or equal to 1;
  • the noise reduction parameter determination module 1201 includes:
  • a first FF filter coefficient determination submodule configured to determine the initial filter coefficients of the plurality of FF filters as the k-th frame filter coefficients of the plurality of FF filters when k is equal to 1, or determine the k-th frame filter coefficients of the plurality of FF filters based on the initial noise reduction gear and the mapping relationship between the noise reduction gear and the FF filter coefficient;
  • the second FF filter coefficient determination submodule is used to determine the kth frame filter coefficients of multiple FF filters based on the k-1th frame reference signal collected by at least one reference microphone, the k-1th frame error signal collected by the error microphone and the target noise reduction gear when k is greater than 1.
  • the second FF filter coefficient determination submodule is specifically used for:
  • the k-1th frame filter coefficients of multiple FF filters are determined.
  • the second FF filter coefficient determination submodule is further specifically used for:
  • One of the multiple FF filters is used as the target FF filter, and the k-th frame filter coefficient of the target FF filter is determined according to the following operation until the k-th frame filter coefficient of each FF filter is determined:
  • the k-th frame filter coefficient of the target FF filter is determined based on the k-1-th frame reference signal collected by the target reference microphone, the k-1-th frame error signal collected by the error microphone, and the k-1-th frame filter coefficient of the target SP, where the target reference microphone is the reference microphone corresponding to the target FF filter, and the target SP refers to the path from the first speaker corresponding to the target FF filter to the error microphone;
  • the k-th frame filter coefficient of the target FF filter is determined based on the k-1th frame reference signal collected by the target reference microphone, the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficients of multiple SPs, and the k-th frame frequency response information and the k-1th frame frequency response information of each FF filter located before the target FF filter.
  • the earphone further includes a plurality of FB filters corresponding one-to-one to the plurality of first speakers;
  • the second FF filter coefficient determination submodule is also specifically used for:
  • the k-1th frame filter coefficients of multiple FF filters are determined.
  • the second FF filter coefficient determination submodule is further specifically used for:
  • One of the multiple FF filters is used as the target FF filter, and the k-th frame filter coefficient of the target FF filter is determined according to the following operation until the k-th frame filter coefficient of each FF filter is determined:
  • the k-th frame filter coefficient of the target FF filter is determined based on the k-1th frame reference signal collected by the target reference microphone, the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficients of the multiple SPs, and the k-1th frame filter coefficients of the multiple FB filters, and the target reference microphone is the reference microphone corresponding to the target FF filter;
  • the k-th frame filter coefficient of the target FF filter is determined based on the k-1th frame reference signal collected by the target reference microphone, the k-1th frame error signal collected by the error microphone, the k-1th frame filter coefficients of multiple SPs, the k-1th frame filter coefficients of multiple FB filters, and the k-th frame frequency response information and the k-1th frame frequency response information of each FF filter located before the target FF filter.
  • the earphone further includes a plurality of feedback FB filters corresponding one-to-one to the plurality of first speakers, and the plurality of sets of target noise reduction parameters further include k-th frame filter coefficients of the plurality of FB filters, where k is an integer greater than or equal to 1;
  • the noise reduction parameter determination module also includes:
  • a first FB filter coefficient determination submodule configured to, when k is equal to 1, determine the initial filter coefficients of the plurality of FB filters as the k-th frame filter coefficients of the plurality of FB filters, or, based on the initial noise reduction gear and a mapping relationship between the noise reduction gear and the FB filter coefficient, determine the k-th frame filter coefficients of the plurality of FB filters;
  • the second FB filter coefficient determination submodule is used to determine the k-th frame filter coefficients of multiple FB filters based on the target noise reduction gear when k is greater than 1.
  • the second FB filter coefficient determination submodule is specifically used for:
  • One of the multiple FB filters is used as the target FB filter, and the k-th frame filter coefficient of the target FB filter is determined according to the following operation until the k-th frame filter coefficient of each FB filter is determined:
  • the k-th frame filter coefficient of the target FB filter is determined based on the target noise reduction gear and the mapping relationship between the noise reduction gear and the FB filter coefficient. If the target FB filter belongs to the second type of FB filter, the k-1-th frame filter coefficient of the target FB filter is determined based on the k-1-th frame error signal collected by the error microphone, the k-1-th frame filter coefficient of the target FB filter and the target noise reduction gear.
  • the second FB filter coefficient determination submodule is further specifically used for:
  • the target noise reduction gear Based on the target noise reduction gear and the mapping relationship between the noise reduction gear and the filter coefficient of the secondary path SP, determine the k-1th frame filter coefficient of the target SP, where the target SP refers to the path from the first speaker to the error microphone corresponding to the target FB filter;
  • the k-th frame filter coefficient of the target FB filter is determined based on the k-1-th frame error signal collected by the error microphone, the k-1-th frame filter coefficient of the target FB filter, and the k-1-th frame filter coefficient of the target SP.
  • a sound frequency band of the first speaker corresponding to the first type FB filter is higher than a sound frequency band of the first speaker corresponding to the second type FB filter.
  • the device further comprises:
  • a first noise reduction level determination module used to determine the noise reduction level of the k-1th frame
  • a noise reduction gear acquisition module is used to acquire the noise reduction gears of m frames before the k-1th frame, where m is greater than or equal to 1 and less than k-1;
  • the second noise reduction level determination module is used to determine a target noise reduction level based on the noise reduction level of the k-1th frame and the noise reduction level of the mth frame.
  • the first noise reduction gear determination module is specifically used to:
  • the noise reduction level of the k-1th frame is determined according to the reference filter coefficients of the multiple FF filters and the mapping relationship between the noise reduction level and the frequency response information of the FF filters;
  • the reference filter coefficient is the initial filter coefficient of the corresponding FF filter.
  • the reference filter coefficient is the filter coefficient of the corresponding FF filter that most recently reached the convergence stability condition before the kth frame, or the k-1th frame filter coefficient of the corresponding FF filter.
  • the first noise reduction gear determination module is specifically used to:
  • a noise reduction level for the k-1th frame is determined.
  • the first noise reduction gear determination module is further specifically configured to:
  • the noise reduction level of the k-1th frame is determined according to the reference noise reduction level with the largest number among the multiple reference noise reduction levels.
  • the first noise reduction gear determination module is specifically used to:
  • the noise reduction gear of the k-1th frame is determined based on the valid downlink signal of the k-1th frame, the k-1th frame reference signal collected by at least one reference microphone, and the k-1th frame error signal collected by the error microphone.
  • the filter coefficients of each FF filter include at least one biquad filter coefficient and a gain.
  • the target anti-phase noise is a full-band anti-phase noise, so no matter whether the first speaker is a high-frequency speaker, a low-frequency speaker or a full-range speaker, when using the target anti-phase noise for noise reduction, the noise reduction ability of each first speaker can be fully utilized.
  • this solution can improve the noise reduction effect of the headphone through the full-band anti-phase noise of multiple noise reduction channels.
  • the noise reduction device provided in the above embodiment performs noise reduction
  • only the division of the above functional modules is used as an example.
  • the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above.
  • the noise reduction device provided in the above embodiment and the noise reduction method embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.
  • FIG13 is a schematic diagram of another headset according to an embodiment of the present application.
  • the headset includes one or more processors 1301 , a communication bus 1302 , a memory 1303 and one or more communication interfaces 1304 .
  • Processor 1301 is a general-purpose central processing unit (CPU), a network processor (NP), a microprocessor, or one or more integrated circuits for implementing the solution of the present application, such as an application-specific integrated circuit (ASIC), a programmable logic device (PLD) or a combination thereof.
  • the PLD is a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL) or any combination thereof.
  • the communication bus 1302 is used to transmit information between the above components.
  • the communication bus 1302 is divided into an address bus, a data bus, a control bus, etc.
  • address bus a data bus
  • control bus a control bus
  • only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.
  • the memory 1303 is a read-only memory (ROM), a random access memory (RAM), an electrically erasable programmable read-only memory (EEPROM), an optical disc (including a compact disc read-only memory (CD-ROM), a compressed optical disc, a laser disc, a digital versatile disc, a Blu-ray disc, etc.), a disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited to this.
  • the memory 1303 exists independently and is connected to the processor 1301 through the communication bus 1302, or the memory 1303 is integrated with the processor 1301.
  • the communication interface 1304 uses any transceiver-like device for communicating with other devices or communication networks.
  • the communication interface 1304 includes a wired communication interface, and optionally, also includes a wireless communication interface.
  • the wired communication interface is, for example, an Ethernet interface.
  • the Ethernet interface is an optical interface, an electrical interface, or a combination thereof.
  • the wireless communication interface is a wireless local area network (WLAN) interface, a cellular network communication interface, or a combination thereof.
  • WLAN wireless local area network
  • the memory 1303 is used to store a program code 1305 for executing the solution of the present application, and the processor 1301 can execute the program code 1305 stored in the memory 1303.
  • the program code includes one or more software modules, and the headset can implement the noise reduction method provided in the embodiment of FIG. 2 above through the processor 1301 and the program code 1305 in the memory 1303.
  • the computer program product includes one or more computer instructions.
  • the computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable device.
  • the computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium.
  • the computer instructions can be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center.
  • the computer-readable storage medium can be any available medium that a computer can access, or a data storage device such as a server or data center that includes one or more available media integrated.
  • the available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a magnetic tape), an optical medium (e.g., a digital versatile disc (DVD)), or a semiconductor medium (e.g., a solid state disk (SSD)).
  • the computer-readable storage medium mentioned in the embodiment of the present application may be a non- Volatile storage media, in other words, may be non-transitory storage media.
  • An embodiment of the present application further provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method described above are implemented.
  • the embodiment of the present application further provides a computer program product, wherein the computer program product stores computer instructions, and when the computer instructions are executed by a processor, the steps of the method described above are implemented.
  • the information including but not limited to user device information, user personal information, etc.
  • data including but not limited to data used for analysis, stored data, displayed data, etc.
  • signals involved in the embodiments of the present application are all authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards of relevant countries and regions.

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Acoustics & Sound (AREA)
  • Signal Processing (AREA)
  • Soundproofing, Sound Blocking, And Sound Damping (AREA)

Abstract

本申请公开了一种降噪方法、耳机、装置、存储介质及计算机程序产品，属于音频处理技术领域。所述耳机包括至少一个参考麦克风、一个误差麦克风和多个第一扬声器，所述方法包括：确定与所述多个第一扬声器一一对应的多组目标降噪参数；基于所述多组目标降噪参数，生成与所述多个第一扬声器一一对应的多组目标反相噪声，所述多组目标反相噪声中每个目标反相噪声的频段均覆盖所述多个第一扬声器的发声频段；利用所述多组目标反相噪声，通过所述多个第一扬声器进行降噪。本方案能够通过多个降噪通道的全频段反相噪声提升耳机降噪效果。

Description

降噪方法、耳机、装置、存储介质及计算机程序产品
本申请要求于2022年11月28日提交的申请号为202211506453.2、发明名称为“降噪方法、耳机、装置、存储介质及计算机程序产品”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及音频处理技术领域，特别涉及一种降噪方法、耳机、装置、存储介质及计算机程序产品。
背景技术
用户佩戴耳机收听音乐或语音等音频信号时，若外界存在环境噪声，用户听到的音频信号的清晰度将会受到影响。当环境噪声比较严重时，用户甚至无法听清耳机内的音频信号。因此，需要实现耳机主动降噪，以尽可能消除耳机佩戴者听到的环境噪声。
耳机主动降噪存在很多挑战：一方面，环境噪声多变且无规律，另一方面，环境噪声泄漏到耳道内部的程度同耳机与人耳的贴合程度相关。然而，不同人的耳道大小和形状存在差别，不同人佩戴同一款耳机时耳机与人耳的贴合程度不同，导致噪声的泄漏程度也不同。同一用户多次佩戴同一款耳机时耳机与人耳的贴合程度也会有所差异。因此，如何提升耳机主动降噪的效果，以尽可能避免环境噪声对耳机佩戴者的影响，是当前的一个研究热点。
发明内容
本申请提供了一种降噪方法、耳机、装置、存储介质及计算机程序产品，能够提升耳机的主动降噪效果。所述技术方案如下：
第一方面，提供了一种降噪方法，应用于耳机，所述耳机包括至少一个参考麦克风、一个误差麦克风和多个第一扬声器；所述方法包括：确定与所述多个第一扬声器一一对应的多组目标降噪参数；基于所述多组目标降噪参数，生成与所述多个第一扬声器一一对应的多组目标反相噪声，所述多组目标反相噪声中每个目标反相噪声的频段均覆盖所述多个第一扬声器的发声频段；利用所述多组目标反相噪声，通过所述多个第一扬声器进行降噪。
由于该多组目标反相噪声与该多个第一扬声器一一对应，且该多组目标反相噪声中每个目标反相噪声的频段均覆盖该多个第一扬声器的发声频段，即每个目标反相噪声均为全频段的反相噪声，所以，无论第一扬声器为高频扬声器、低频扬声器还是全频扬声器，在利用该多组目标反相噪声进行降噪时，均能够充分发挥每个第一扬声器的降噪能力。换种方式来讲，在多个降噪通道、多个扬声器的耳机架构下，本方案能够通过多个降噪通道的全频段反相噪声提升耳机降噪效果。
本申请提供的降噪方法能够以帧为单位来确定该多组目标降噪参数，即，每帧都确定出与该多个第一扬声器一一对应的多组目标降噪参数。当然还能够以其他时间单位来确定目标降噪参数，比如，每两帧确定出与该多个第一扬声器一一对应的多组目标降噪参数。接下来以帧为单位进行介绍。
该耳机还包括与该多个第一扬声器一一对应的多个前馈(feed forward，FF)滤波器，此时，该多组目标降噪参数包括该多个FF滤波器的第k帧滤波系数，k为大于等于1的整数。在某些情况下，该耳机还包括与该多个第一扬声器一一对应的多个反馈(feed back，FB)滤波器，也即是，该多个FB滤波器与该多个FF滤波器一一对应。此时，该多组目标降噪参数还包括该多个FB滤波器的第k帧滤波系数。而且，在该耳机还包括下行补偿滤波器的情况下，该多组目标降噪参数还包括下行补偿滤波器的第k帧滤波系数。另外，在k大于1的情况下，还可以确定目标降噪档位。因此接下来将对这四部分内容分别进行介绍。
(1)确定该多个FF滤波器的第k帧滤波系数。
在k等于1的情况下，将该多个FF滤波器的初始滤波系数确定为该多个FF滤波器的第k帧滤波系数，即，该多个FF滤波器的第1帧滤波系数为相应FF滤波器的初始滤波系数，或者，基于初始降噪档位、以及降噪档位与FF滤波系数的映射关系，确定该多个FF滤波器的第k帧滤波系数。在k大于1的情 况下，基于该至少一个参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号以及目标降噪档位，确定该多个FF滤波器的第k帧滤波系数。即，通过自适应方法来确定该多个FF滤波器的第k帧滤波系数，该确定过程是一个自适应过程，也可以称为迭代过程。
需要说明的是，该多个FF滤波器的初始滤波系数可以相同，也可以不同，该初始滤波系数可以为0，也可以不为0，本申请实施例对此不做限定。该初始降噪档位可以是事先设置的档位，该档位是指对应的降噪系数能够正常进行降噪同时不至于引入稳定性问题的档位。当然，该初始降噪档位也可以是在降噪开始时，通过用户终端发送的“降噪开”、“叮咚”等提示音确定的一个档位，该档位对应的降噪系数与当前的人耳及佩戴姿态能更好适配，而且在该档位对应的降噪系数的基础上进行自适应迭代能够更快达到收敛状态，本申请实施例同样对此不做限定。
其中，基于该至少一个参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号以及目标降噪档位，确定该多个FF滤波器的第k帧滤波系数的实现过程包括：基于目标降噪档位、以及降噪档位与次级路径(secondary path，SP)的滤波系数的映射关系，确定多个SP的第k-1帧滤波系数，该多个SP是指该多个第一扬声器到误差麦克风的路径。基于该至少一个参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号和该多个SP的第k-1帧滤波系数，确定该多个FF滤波器的第k帧滤波系数。
在确定该多个FF滤波器的第k帧滤波系数时，可以通过多通道联动的方式来确定。而且，在该耳机包括多个FF滤波器的情况下，可能还包括与该多个第一扬声器一一对应的多个FB滤波器，也可能不包括该多个FB滤波器。在不同的情况下，确定该多个FF滤波器的第k帧滤波系数的方式不同。接下来将分别进行介绍。
由于基于该至少一个参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号和该多个SP的第k-1帧滤波系数，确定每个FF滤波器的第k帧滤波系数的确定过程相同，因此接下来以其中一个为例进行介绍。即，将该多个FF滤波器中的一个FF滤波器作为目标FF滤波器，按照如下方式确定目标FF滤波器的第k帧滤波系数，该多个FF滤波器中其他FF滤波器的第k帧滤波系数的确定过程可以参考目标FF滤波器的第k帧滤波系数的确定过程。
第一种情况，该耳机不包括该多个FB滤波器。如果目标FF滤波器是首个FF滤波器，则基于目标参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号、以及目标SP的第k-1帧滤波系数，确定目标FF滤波器的第k帧滤波系数，目标参考麦克风为目标FF滤波器对应的参考麦克风，目标SP是指目标FF滤波器对应的第一扬声器到误差麦克风的路径。如果目标FF滤波器是非首个FF滤波器，则基于目标参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号、该多个SP的第k-1帧滤波系数、以及位于目标FF滤波器之前的各个FF滤波器的第k帧频响信息和第k-1帧频响信息，确定目标FF滤波器的第k帧滤波系数。
在目标FF滤波器是首个FF滤波器的情况下，基于目标参考麦克风采集的第k-1帧参考信号和误差麦克风采集的第k-1帧误差信号，确定残留误差，基于目标FF滤波器的第k-1帧频响信息、目标SP的第k-1帧滤波系数和该残留误差，确定目标FF滤波器的第k帧频响信息。基于目标FF滤波器的第k帧频响信息，确定目标FF滤波器的第k帧滤波系数。
该多个FF滤波器中的一个FF滤波器对应一个参考麦克风。即，目标参考麦克风包括一个参考麦克风。此时，基于目标参考麦克风采集的第k-1帧参考信号和误差麦克风采集的第k-1帧误差信号，确定残留误差。
该多个FF滤波器中的一个FF滤波器对应至少两个参考麦克风。即，目标参考麦克风包括至少两个参考麦克风。此时，将目标参考麦克风包括的至少两个参考麦克风采集的第k-1帧参考信号进行混音，以得到第k-1帧混音参考信号。基于第k-1帧混音参考信号和误差麦克风采集的第k-1帧误差信号，确定残留误差。这样，可以提升参考信号的信噪比。
在目标FF滤波器为非首个FF滤波器的情况下，基于目标参考麦克风采集的第k-1帧参考信号和误差麦克风采集的第k-1帧误差信号，确定残留误差。基于目标FF滤波器的第k-1帧频响信息、该残留误差、该多个SP的第k-1帧滤波系数、以及位于目标FF滤波器之前的各个FF滤波器的第k帧频响信息和第k-1帧频响信息，确定目标FF滤波器的第k帧频响信息。基于目标FF滤波器的第k帧频响信息，确定目标FF滤波器的第k帧滤波系数。
确定目标FF滤波器的第k帧频响信息时，可以基于目标FF滤波器的第k-1帧频响信息、该残留误差、目标SP的第k-1帧滤波系数、位于目标FF滤波器之前的各个FF滤波器的第k帧频响信息和第k-1帧频响信息、以及位于目标FF滤波器之前的各个FF滤波器对应的SP的第k-1帧滤波系数，确定目标FF滤波器的第k帧频响信息。
其中，基于目标FF滤波器的第k帧频响信息，确定目标FF滤波器的第k帧滤波系数的实现过程包括：建立目标FF滤波器的滤波系数变量与目标FF滤波器的第k帧频响信息之间的损失函数。基于该损失函数，通过梯度下降法确定滤波系数变量的值，基于该滤波系数变量的值确定目标FF滤波器的第k帧滤波系数。也即是，建立目标FF滤波器的滤波系数变量与目标FF滤波器的第k帧频响信息之间的损失函数。通过梯度下降法确定该变量的最优值，从而通过该变量的最优值确定目标FF滤波器的第k帧滤波系数。
目标FF滤波器的每帧滤波系数都是按照梯度下降法确定的，在确定目标FF滤波器的每帧滤波系数时都会确定出一个损失函数的值，在该损失函数的值达到最小门限值时，确定目标FF滤波器的滤波系数达到收敛稳定条件。比如，对于目标FF滤波器的第k帧滤波系数来说，在滤波系数变量与目标FF滤波器的第k帧频响信息之间的损失函数的值达到最小门限值时，确定目标FF滤波器的第k帧滤波系数达到收敛稳定条件。在该损失函数的值未达到最小门限值时，确定目标FF滤波器的第k帧滤波系数未达到收敛稳定条件。该最小门限值是事先设置的，在不同的情况下，可以按照不同的需求进行调整。
可选地，每个FF滤波器的滤波系数包括至少一个双二阶滤波系数和一个增益。该双二阶滤波系数对应的变量包括滤波器类型、截止频率和品质因数。当然，实际应用中，每个FF滤波器的滤波系数还可以包括其他更多或更少的参数，本申请对此不做限定。
在某些情况下，安静环境具有底噪问题，即背景噪声，比如对于半开放耳机来说，该耳机在安静环境下相比入耳式耳机更易出现底噪问题。而且安静环境不需要强降噪，部分人在安静环境下强降噪会有不适感。并且降噪力度越大的情况下，人的负压感越强，所以，通过梯度下降法确定滤波系数变量的值时，可以基于环境音量动态调节目标降噪幅度，从而按照目标降噪幅度来确定目标FF滤波器的第k帧滤波系数，提高自适应降噪的主观体验效果。即，根据第k-1帧的环境音量以及位于第k-1帧之前的t帧环境音量确定目标降噪幅度，t大于等于1且小于k-1。基于目标降噪幅度和该损失函数，通过梯度下降法确定滤波系数变量的值，基于该滤波系数变量的值确定目标FF滤波器的第k帧滤波系数。
第二种情况，该耳机还包括该多个FB滤波器。如果目标FF滤波器是首个FF滤波器，则基于目标参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号、该多个SP的第k-1帧滤波系数和该多个FB滤波器的第k-1帧滤波系数，确定目标FF滤波器的第k帧滤波系数。如果目标FF滤波器是非首个FF滤波器，则基于目标参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号、该多个SP的第k-1帧滤波系数、该多个FB滤波器的第k-1帧滤波系数、以及位于目标FF滤波器之前的各个FF滤波器的第k帧频响信息和第k-1帧频响信息，确定目标FF滤波器的第k帧滤波系数。
在目标FF滤波器为首个FF滤波器的情况下，可以基于目标参考麦克风采集的第k-1帧参考信号和误差麦克风采集的第k-1帧误差信号，确定残留误差。基于目标FF滤波器的第k-1帧频响信息、该残留误差、该多个FB滤波器的第k-1帧滤波系数、以及该多个SP的第k-1帧滤波系数，确定目标FF滤波器的第k帧频响信息。基于目标FF滤波器的第k帧频响信息，确定目标FF滤波器的第k帧滤波系数。
在目标FF滤波器为非首个FF滤波器的情况下，可以基于目标参考麦克风采集的第k-1帧参考信号和误差麦克风采集的第k-1帧误差信号，确定残留误差。基于目标FF滤波器的第k-1帧频响信息、该残留误差、该多个SP的第k-1帧滤波系数、该多个FB滤波器的第k-1帧滤波系数、以及位于目标FF滤波器之前的各个FF滤波器的第k帧频响信息和第k-1帧频响信息，确定目标FF滤波器的第k帧频响信息。基于目标FF滤波器的第k帧频响信息，确定目标FF滤波器的第k帧滤波系数。
在上述确定目标FF滤波器的第k帧频响信息的过程中，不管该耳机是否包括目标FB滤波器，目标FF滤波器的第k帧频响信息都是基于目标SP的第k-1帧滤波系数来确定的，而目标SP的第k-1帧滤波系数是基于目标降噪档位通过查询降噪档位与SP的滤波系数的映射关系来确定的，也即是，目标SP的第k-1帧滤波系数是一个估计值，通过这个估计值来确定目标FF滤波器的第k帧频响信息，可以摆脱对目标SP真实值的依赖，从而在没有下行信号的情况下，也能够实现FF滤波器的滤波系数的自适应。
(2)确定该多个FB滤波器的第k帧滤波系数。
在k等于1的情况下，将该多个FB滤波器的初始滤波系数确定为该多个FB滤波器的第k帧滤波系数，即，该多个FB滤波器的第1帧滤波系数为相应FB滤波器的初始滤波系数，或者，基于初始降噪档位、以及降噪档位与FB滤波系数的映射关系，确定该多个FB滤波器的第k帧滤波系数。在k大于1的情况下，可以基于目标降噪档位确定该多个FB滤波器的第k帧滤波系数。
需要说明的是，该多个FB滤波器的初始滤波系数可以相同，也可以不同，该初始滤波系数可以为0，也可以不为0，本申请实施例对此不做限定。
由于基于目标降噪档位确定每个FB滤波器的第k帧滤波系数的过程相同，因此，接下来以其中一个为例进行介绍。即，将该多个FB滤波器中的一个FB滤波器作为目标FB滤波器，按照如下两种方式确定目标FB滤波器的第k帧滤波系数，该多个FB滤波器中其他FB滤波器的第k帧滤波系数的确定过程可以参考目标FB滤波器的第k帧滤波系数的确定过程。也就是说，在k大于1的情况下，可以按照如下两种方式确定目标FB滤波器的第k帧滤波系数。
第一种方式，基于目标降噪档位，以及降噪档位与FB滤波系数的映射关系，确定目标FB滤波器的第k帧滤波系数。
由于降噪档位与FB滤波系数的映射关系是事先存储的，所以通过第一种方式确定目标FB滤波器的第k帧滤波系数的稳定性较好，操作也较为简单，效率较高。
第二种方式，如果目标FB滤波器属于第一类FB滤波器，则基于目标降噪档位，以及降噪档位与FB滤波系数的映射关系，确定目标FB滤波器的第k帧滤波系数。如果目标FB滤波器属于第二类FB滤波器，则基于误差麦克风采集的第k-1帧误差信号、目标FB滤波器的第k-1帧滤波系数以及目标降噪档位，确定目标FB滤波器的第k帧滤波系数。
由于目标FB滤波器的第一帧滤波系数可以通过初始降噪档位查询降噪档位与FB滤波系数的映射关系确定得到，所以，在k大于等于1的情况下，相当于可以通过三种方式来确定目标FB滤波器的第k帧滤波系数。即，(1)通过查询降噪档位与FB滤波系数的映射关系来确定目标FB滤波器的第k帧滤波系数。(2)如果目标FB滤波器属于第一类FB滤波器，则通过查询降噪档位与FB滤波系数的映射关系来确定目标FB滤波器的第k帧滤波系数，如果目标FB滤波器属于第二类FB滤波器，则基于误差麦克风采集的第k-1帧误差信号、目标FB滤波器的第k-1帧滤波系数以及目标降噪档位，确定目标FB滤波器的第k帧滤波系数。(3)如果目标FB滤波器属于第一类FB滤波器，或者，目标FB滤波器属于第二类FB滤波器且k等于1，则通过查询降噪档位与FB滤波系数的映射关系来确定目标FB滤波器的第k帧滤波系数。如果目标FB滤波器属于第二类FB滤波器且k大于1，则基于误差麦克风采集的第k-1帧误差信号、目标FB滤波器的第k-1帧滤波系数以及目标降噪档位，确定目标FB滤波器的第k帧滤波系数。
基于误差麦克风采集的第k-1帧误差信号、目标FB滤波器的第k-1帧滤波系数以及目标降噪档位，确定目标FB滤波器的第k帧滤波系数的实现过程包括：基于目标降噪档位，以及降噪档位与SP的滤波系数的映射关系，确定目标SP的第k-1帧滤波系数，目标SP是指目标FB滤波器对应的第一扬声器到误差麦克风的路径；基于误差麦克风采集的第k-1帧误差信号、目标FB滤波器的第k-1帧滤波系数以及目标SP的第k-1帧滤波系数，确定目标FB滤波器的第k帧滤波系数。
第一类FB滤波器对应的第一扬声器的发声频段高于第二类FB滤波器对应的第一扬声器的发声频段。也即是，第一类FB滤波器对应的第一扬声器为高频扬声器，第二类FB滤波器对应的第一扬声器为低频扬声器。当然，第一类FB滤波器和第二类FB滤波器也可以不用按照发声频段来区分，而是按照其他的方式来区分，本申请同样对此不做限定。
上述第二种方式和第三方式，将查询降噪档位与FB滤波系数的映射关系的方式，与自适应的方式结合，不仅能够提高降噪效果，复杂性相对来说也不是很高，而且稳定性也相对可控。
需要说明的是，不仅可以按照上述三种方式来确定目标FB滤波器的第k帧滤波系数，还可以按照其他方式来确定目标FB滤波器的第k帧滤波系数。比如，不管目标FB滤波器属于第一类FB滤波器还是第二类FB滤波器，都基于误差麦克风采集的第k-1帧误差信号、目标FB滤波器的第k-1帧滤波系数以及目标降噪档位，确定目标FB滤波器的第k帧滤波系数。本申请实施例对此不做限定。
(3)确定下行补偿滤波器的第k帧滤波系数。
在k等于1的情况下，将下行补偿滤波器的初始滤波系数确定为下行补偿滤波器的第k帧滤波系数，或者，基于初始降噪档位、以及降噪档位与下行补偿滤波系数的映射关系，确定下行补偿滤波器的第k帧 滤波系数。在k大于1的情况下，基于目标降噪档位，以及降噪档位与下行补偿滤波系数的映射关系，确定下行补偿滤波器的第k帧滤波系数。
降噪档位与下行补偿滤波系数的映射关系包括多个降噪档位，每个降噪档位均与该下行补偿滤波器的滤波系数存在映射关系，不同的降噪档位与该下行补偿滤波器的滤波系数的映射关系可能不同，所以，在确定出目标降噪档位之后，基于目标降噪档位，从该降噪档位与下行补偿滤波系数的映射关系中，能够获取到对应的下行补偿滤波系数，将获取的下行补偿滤波系数作为该下行补偿滤波器的第k帧滤波系数。
(4)确定目标降噪档位。
确定第k-1帧降噪档位，获取位于第k-1帧之前的m帧降噪档位，m大于等于1且小于k-1。基于第k-1帧降噪档位和该m帧降噪档位，确定目标降噪档位。
由于第k-1帧可能存在有效下行信号，也可能不存在有效下行信号，而且有可能处于安静环境，也可能处于非安静环境，当然，还可能存在异常信号。在不同的情况下，确定第k-1帧降噪档位的方式不同，接下来将分别进行介绍。
第一种情况，第k-1帧不存在有效下行信号且处于非安静环境。在这种情况下，根据该多个FF滤波器的参考滤波系数，以及降噪档位与FF滤波器的频响信息的映射关系，确定第k-1帧降噪档位。其中，在k等于2的情况下，参考滤波系数为相应FF滤波器的初始滤波系数，在k大于2的情况下，参考滤波系数为相应FF滤波器在第k帧之前最近一次达到收敛稳定条件的滤波系数，或者为相应FF滤波器的第k-1帧滤波系数。
基于该多个FF滤波器的参考滤波系数，确定该多个FF滤波器的参考频响信息。基于降噪档位与FF滤波器的频响信息的映射关系，确定与该多个FF滤波器的参考频响信息分别匹配的降噪档位，以得到多个参考降噪档位。基于该多个参考降噪档位，确定第k-1帧降噪档位。
其中，基于该多个参考降噪档位确定第k-1帧降噪档位的方式包括多种，比如，按照该多个参考降噪档位的平均值确定第k-1帧降噪档位。或者，按照该多个参考降噪档位中数量最多的参考降噪档位确定第k-1帧降噪档位。
按照该多个参考降噪档位的平均值确定第k-1帧降噪档位时，可以直接将该多个参考降噪档位的平均值确定为第k-1帧降噪档位，也可以对该多个参考降噪档位的平均值进行调整，以得到第k-1帧降噪档位。同理，按照该多个参考降噪档位中数量最多的参考降噪档位确定第k-1帧降噪档位时，可以直接将该多个参考降噪档位中数量最多的参考降噪档位确定为第k-1帧降噪档位，也可以对该多个参考降噪档位中数量最多的参考降噪档位进行调整，以得到第k-1帧降噪档位。
第二种情况，第k-1帧存在有效下行信号。在这种情况下，基于第k-1帧的有效下行信号、该至少一个参考麦克风采集的第k-1帧参考信号，以及误差麦克风采集的第k-1帧误差信号，确定第k-1帧降噪档位。
基于上文描述，在耳机处于下行使能状态且未处于下行间歇期的情况下，确定第k-1帧存在有效下行信号。此时，可以基于第k-1帧的有效下行信号、该至少一个参考麦克风采集的第k-1帧参考信号，以及误差麦克风采集的第k-1帧误差信号，从误差麦克风采集的第k-1帧误差信号中提取有效下行信号，从而基于提取的有效下行信号确定第k-1帧降噪档位。
第三种情况，第k-1帧不存在有效下行信号且处于安静环境，或者，第k-1帧存在异常噪声信号。在这种情况下，将第k-3帧降噪档位确定为第k-1帧降噪档位。即，维持降噪档位不变。
在第k-1帧不存在有效下行信号且处于安静环境的情况下，噪声基本没有发生变化，此时，可以维持降噪档位不变。在第k-1帧存在异常噪声信号的情况下，通过维持降噪档位不变，以进行鲁棒性控制，从而避免降噪档位的发散。
通过上述三种情况确定出第k-1帧降噪档位之后，可以将第k-1帧降噪档位和位于第k-1帧之前的m帧降噪档位进行综合，以确定目标降噪档位。
其中，该m帧降噪档位可以为位于第k-1帧之前的任意m帧降噪档位，也可以为位于第k-1帧之前且距离第k-1帧最近的m帧降噪档位。另外，基于第k-1帧降噪档位和位于第k-1帧之前的m帧降噪档位确定目标降噪档位的实现方式包括多种，比如，按照相关算法进行降噪效果的评优，以确定第k-1帧降噪档位对应的降噪概率以及该m帧降噪档位分别对应的降噪概率，将降噪概率最大的降噪档位确定为目标降噪档位。或者，确定第k-1帧降噪档位与该m帧降噪档位的算术平均值或者加权平均值，以得到目标降噪档 位。又或者，将第k-1帧降噪档位和该m帧降噪档位中出现数量最多的降噪档位确定为目标降噪档位等等。
基于上文描述，该多组目标降噪参数可以称为多个降噪通道的降噪参数，这样，生成的该多组目标反相噪声也可以称为多个降噪通道的反相噪声。由于每个降噪通道的反相噪声的生成过程相同，因此接下来以其中一个降噪通道为例进行说明。
将该多个降噪通道中的一个降噪通道作为目标降噪通道，目标降噪通道包括目标FF滤波器和目标第一扬声器，目标FF滤波器对应的参考麦克风称为目标参考麦克风。此时，目标反相噪声包括前馈反相噪声。即，基于目标FF滤波器的第k帧滤波系数，对目标参考麦克风采集的第k帧参考信号进行处理，以得到前馈反相噪声。
基于上文描述，目标参考麦克风可能包括一个参考麦克风，也可能包括至少两个参考麦克风。在目标参考麦克风包括一个参考麦克风的情况下，可以直接基于目标FF滤波器的第k帧滤波系数，对目标参考麦克风采集的第k帧参考信号进行处理，以得到前馈反相噪声。在目标参考麦克风包括至少两个参考麦克风的情况下，将该至少两个参考麦克风采集的第k帧参考信号进行混音，以得到第k帧混音参考信号，然后，基于目标FF滤波器的第k帧滤波系数，对第k帧混音参考信号进行处理，以得到前馈反相噪声。
在该耳机还包括FB滤波器的情况下，目标降噪通道还包括目标FB滤波器。此时，目标反相噪声还包括反馈反相噪声。即，基于下行补偿滤波器的第k帧滤波系数，对用户终端发送的第k帧下行信号进行下行补偿。然后，将下行补偿后的第k帧下行信号取反后与误差麦克风采集的第k帧误差信号进行混音，以得到误差麦克风采集的第k帧噪声信号。基于目标FB滤波器的第k帧滤波系数，对误差麦克风采集的第k帧噪声信号进行处理，以得到反馈反相噪声。
通过下行补偿能够将误差麦克风采集的误差信号中的所有下行信号去掉，从而只将残留的噪声信号通过FB滤波器做降噪，以避免对下行信号造成音质损伤。并且，通过对用户终端发送的第k帧下行信号进行下行补偿，能够将所有扬声器在误差麦克风处的下行信号都去掉，从而避免对全频的下行信号造成音质损伤。
基于上文描述，在该多组目标降噪参数以帧为单位来确定的情况下，由于一帧可能包括一个样点，也可能包括多个样点，所以，在生成目标反相噪声时，可以在每个样点处都生成一组目标反相噪声，也可以一帧生成一组目标反相噪声。
由于在确定多组目标降噪参数时，对下行信号并未进行分频，也即是，通过全频的下行信号来确定该多组目标降噪参数。这样，基于该多组目标降噪参数，生成与该多个第一扬声器一一对应的多组目标反相噪声之后，该多组目标反相噪声中每个目标反相噪声的频段均覆盖该多个第一扬声器的发声频段，即每个目标反相噪声的频段均为全频段。
在生成多组目标反相噪声之后，将该多组目标反相噪声分别与该多个第一扬声器待播放的第k帧下行信号进行混音后，通过对应的第一扬声器进行播放，从而实现降噪的目的。
该多个第一扬声器可能一部分为高频扬声器，另一部分为低频扬声器。或者，一部分为全频扬声器，另一部分为非全频扬声器。也即是，该多个第一扬声器的发声频段可能不同。或者，该多个第一扬声器全部为全频扬声器。又或者，该多个第一扬声器全部为非全频扬声器。在该多个第一扬声器全部为全频扬声器的情况下，该多个第一扬声器待播放的第k帧下行信号均为用户终端发送的第k帧下行信号。在该多个第一扬声器不全为全频扬声器的情况下，需要按照每个第一扬声器的发声频段，对用户终端发送的第k帧下行信号进行分频，从而得到每个第一扬声器待播放的第k帧下行信号。
该多个第一扬声器中的两个第一扬声器可以包括由一个双振膜(或者称为双动圈)喇叭形成的两个第一扬声器。或者，该多个第一扬声器包括分体喇叭的多个扬声器。
可选地，该耳机还可以包括至少一个第二扬声器，该至少一个第二扬声器不参与降噪，此时，第二扬声器可以参与下行补偿(即，将用户终端发送的下行信号进行下行补偿，该下行信号为全频段的音频信号，包括第二扬声器的发声频段的音频信号)。在这种情况下，第一扬声器可以为中低频的扬声器，也可以为全频的扬声器，第二扬声器可以为高频的扬声器，也可以为中频的扬声器或者低频的扬声器。可选地，第二扬声器也可以不参与下行补偿，此时，第一扬声器可以为中低频的扬声器，也可以为全频的扬声器，第二扬声器为高频的扬声器。
由于上述通过自适应的方法确定该多组目标降噪参数的过程需要一定的时间，在一帧包括多个样点且 一帧的时长较长的情况下，该多组目标降噪参数的确定时长会小于一帧的时长，所以，通过第k-1帧的相关数据，可以在第k帧开始的部分时间段内进行计算，从而得到第k帧的多组目标降噪参数，进而按照第k帧的多组目标降噪参数，在第k帧后续的部分时间段内进行主动降噪。但是，在一帧包括一个样点，或者，一帧包括多个样点且一帧的时长较短的情况下，该多组目标降噪参数的确定时长可能等于一帧的时长，此时，通过第k-1帧的相关数据，可能需要在第k帧的整个时间段内进行计算，从而得到多组目标降噪参数。在这种情况下，可以将该多组目标降噪参数确定为第k+1帧的多组目标降噪参数，进而按照第k+1帧的多组目标降噪参数，在第k+1帧的时间段内进行主动降噪。上述内容都是以前者为例进行介绍的。
第二方面，提供了一种耳机，所述耳机包括至少一个参考麦克风、一个误差麦克风、多个第一扬声器和一个降噪处理器，所述降噪处理器用于实现上述第一方面所述方法的步骤。
可选地，所述多个第一扬声器包括由一个双振膜喇叭形成的两个第一扬声器；或者，所述多个第一扬声器包括分体喇叭的多个扬声器。
可选地，所述耳机还包括至少一个第二扬声器，所述至少一个第二扬声器不参与降噪。
第三方面，提供了一种降噪装置，所述降噪装置具有实现上述第一方面中降噪方法行为的功能。所述降噪装置包括一个或多个模块，该一个或多个模块用于实现上述第一方面所提供的降噪方法。
第四方面，提供了一种计算机可读存储介质，所述计算机可读存储介质中存储有指令，当其在计算机上运行时，使得计算机执行上述第一方面所述的降噪方法。
第五方面，提供了一种包含指令的计算机程序产品，当其在计算机上运行时，使得计算机执行上述第一方面所述的降噪方法。
上述第二方面至第五方面所获得的技术效果与第一方面中对应的技术手段获得的技术效果近似，在这里不再赘述。
附图说明
图1是本申请实施例提供的一种降噪方法所涉及的系统架构图；
图2是本申请实施例提供的一种降噪方法的流程图；
图3是本申请实施例提供的一种确定目标降噪幅度的流程图；
图4是本申请实施例提供的一种16个降噪档位下FF滤波器的频响曲线的示意图；
图5是本申请实施例提供的一种确定第k-1帧降噪档位的流程图；
图6是本申请实施例提供的一种确定多组目标降噪参数的流程图；
图7是本申请实施例提供的一种耳机的结构示意图；
图8是本申请实施例提供的另一种耳机的结构示意图；
图9是本申请实施例提供的另一种耳机的结构示意图；
图10是本申请实施例提供的另一种耳机的结构示意图；
图11是本申请实施例提供的另一种耳机的结构示意图；
图12是本申请实施例提供的一种降噪装置的结构示意图；
图13是本申请实施例提供的另一种耳机的结构示意图。
具体实施方式
为使本申请的目的、技术方案和优点更加清楚，下面将结合附图对本申请实施方式作进一步地详细描述。
主动降噪耳机近年来很流行，传统的降噪耳机一般都是入耳式或者头戴式，原因在于这两种形态下耳机与耳道的密闭条件好，不同的人佩戴声学泄漏也相对稳定，从技术上比较容易实现主动降噪，效果更有保证，因而一般采用固定系数的降噪模式即可。但这两种形态的耳机也存在一些不足，比如耳机与耳道之 间的密闭性过好，影响人的主观舒适度，典型表现在异物感以及走路等条件下的闭塞感，很难长时间佩戴。
半开放耳机由于其较好的舒适性为用户所广泛接受。但半开放形态下由于耳机与人耳的密封性不好，环境噪声更容易被人感受到。半开放形态下实现耳机主动降噪更具挑战，原因就在于不同的人佩戴时、哪怕是同一个人不同次佩戴时，佩戴姿态差异也会很大，用技术语言表述就是耳机与耳道之间的响应函数、声学泄漏程度的差别大。因而如何实现自适应降噪是半开放形态下的迫切诉求，以应对耳道响应的差异化问题，从而实现耳机与耳道之间的最佳匹配。而且，即便是入耳式或头戴式形态下，耳道响应也不是绝对一致，还是存在或大或小的差别，目前业界也在探索耳机自适应降噪的可行性。
前面也提到，耳机的形态有入耳、头戴、半开放、开放式等多种，扬声器(即喇叭)在整机下的音频性能表现又与耳机的具体形态强相关，尤其是低频的表现。在入耳、头戴等相对密封的形态下，高中低频的音频表现一般是有保证的；在半开放或开放式形态下，由于声学泄漏严重，低频频响的跌落相当厉害，不仅关系到低频音质的表现力，也严重影响主动降噪的效果(不足以产生足够能量的反相噪声)。
基于以上提到的一些问题，本申请实施例提供了一种降噪方法，以实现耳机自适应主动降噪(active noise cancellation，ANC)。请参考图1，图1是本申请实施例提供的一种降噪方法所涉及的系统架构图。该系统可称为耳机降噪系统。该系统包括耳机101和用户终端102。耳机101和用户终端102之间通过有线或无线的方式连接以进行通信。例如，耳机101与用户终端102通过蓝牙进行通信，或者通过其他无线网络进行通信。
耳机101与用户终端102之间能够传输音频信号以及控制信号。例如，用户终端102将音乐或语音等音频信号发送至耳机101进行播放，又例如，用户终端102向耳机101发送控制信号，以控制耳机101的主动降噪功能是否开启等等。
用户终端102可以为手机、电脑(如笔记本电脑、台式电脑、手持平板电脑、车载平板电脑)等电子设备，该用户终端102也可以为其他电子设备，例如，智能音箱、车载音箱等。本申请实施例对用户终端102的类型和结构等不作限定。
可选地，本申请实施例提供的耳机101可以是有线的，也可以是无线的。而且，从佩戴方式来说，本申请实施例提供的耳机101可以为颈戴式、耳挂/耳夹式、真无线立体声(true wireless stereo，TWS)等等，从外观形态来说，本申请实施例提供的耳机101可以为入耳式、半开放式、开放式、头戴式等等。本申请实施例对耳机的通信方式、佩戴方式和外观形态不作限定。下面结合耳机在人耳中的佩戴形态介绍本申请实施例提供的耳机的硬件结构。
如图1所示，耳机101包括多个扬声器(即喇叭)、多个麦克风、微控制器(micro control unit，MCU)、ANC芯片以及存储器。该多个扬声器包括多个第一扬声器，如喇叭1和喇叭2。该多个第一扬声器需要参与降噪，例如，第一扬声器为中低频扬声器，中低频扬声器需要参与降噪。可选地，该多个扬声器还包括至少一个第二扬声器，该至少一个第二扬声器不参与降噪，例如，第二扬声器为高频扬声器，高频扬声器无需参与降噪。当然，对于任一扬声器来说，无论该扬声器是高频还是中低频扬声器，都可以参与降噪或不参与降噪。换种方式来讲，本申请实施例不限定参与降噪的第一扬声器的发声频段，也不限定不参与降噪的第二扬声器的发声频段。该多个麦克风包括至少一个参考麦克风和一个误差麦克风。图1以一个参考麦克风为例介绍。
其中，扬声器用于播放下行信号(如音乐、语音等音频信号)，每个扬声器使用独立的数字模拟转换器(digital to analog converter，DAC)和功率放大器(power amplifier，PA)进行驱动，即一个扬声器对应一个DAC和一个PA，不同的扬声器对应的DAC和PA不同。其中，在降噪过程中，第一扬声器还用于播放反相噪声，该反相噪声用于减弱用户耳道内的噪声信号，从而达到主动降低噪声的效果。
参考麦克风部署在耳机外侧，在耳机被佩戴至人耳之后，参考麦克风位于人耳外侧。参考麦克风用于采集外界环境的噪声信号。在本申请实施例中，将参考麦克风采集的噪声信号称为参考信号。
误差麦克风部署在耳机内侧，在耳机被佩戴至人耳之后，误差麦克风位于人耳内侧。误差麦克风用于采集耳道内的噪声信号。在本申请实施例中，将误差麦克风采集的噪声信号称为误差信号。
微控制器用于处理参考麦克风采集的参考信号、误差麦克风采集的误差信号、下行信号等等，从而确定多个第一扬声器中每个第一扬声器对应的一组目标降噪参数，并将每个第一扬声器对应的一组目标降噪参数写入ANC芯片。
ANC芯片用于基于每个第一扬声器对应的一组目标降噪参数，对参考麦克风采集的参考信号和误差 麦克风采集的误差信号进行处理，以生成反相噪声，进而将生成的反相噪声与第一扬声器待播放的下行信号进行混音后输出到对应的第一扬声器，以减弱耳道内的噪声信号。
存储器用于存储确定每个第一扬声器对应的目标降噪参数时所涉及的初始参数，以及映射关系等等。
需要说明的是，微控制器、ANC芯片和存储器可以被集成在同一块电路板上，也可以被部署在不同的电路板上，本申请实施例对此不做限定。另外，微控制器和ANC芯片只是逻辑功能上的表述区分，在实际物理形态上，微控制器和ANC芯片可以集成在一个芯片上，也可以分别部署在多个芯片上，比如，将微控制器和ANC芯片部署在两个芯片上。
可选地，耳机101还可以包括其他元件，例如接近光传感器，用于检测耳机101是否在耳。若耳机101是无线耳机，耳机101还可以包括无线通信模块，该无线通信模块可以为无线局域网模块或者蓝牙模块。该无线通信模块用于耳机101与其他设备通过通信。
可以理解的是，本申请实施例示意的结构并不构成对耳机的限定，在另一些实施例中，耳机101可以包括比图示更多或更少的部件，或者组合某些部件，或者拆分某些部件，或者不同的部件布置。图示的部件可以以硬件、软件、或软硬结合的方式实现。
本申请实施例描述的系统架构以及业务场景是为了更加清楚的说明本申请实施例的技术方案，并不构成对于本申请实施例提供的技术方案的限定，本领域普通技术人员可知，随着系统架构的演变和新业务场景的出现，本申请实施例提供的技术方案对于类似的技术问题，同样适用。
图2是本申请实施例提供的一种降噪方法的流程图，该方法应用于耳机，该耳机包括至少一个参考麦克风、一个误差麦克风和多个第一扬声器。请参考图2，该方法包括如下步骤。
步骤201：确定与该多个第一扬声器一一对应的多组目标降噪参数。
本申请实施例提供的降噪方法能够以帧为单位来确定该多组目标降噪参数，即，每帧都确定出与该多个第一扬声器一一对应的多组目标降噪参数。当然还能够以其他时间单位来确定目标降噪参数，比如，每两帧确定出与该多个第一扬声器一一对应的多组目标降噪参数。接下来以帧为单位进行介绍。
在一些实施例中，该耳机还包括与该多个第一扬声器一一对应的多个FF滤波器，此时，该多组目标降噪参数包括该多个FF滤波器的第k帧滤波系数，k为大于等于1的整数。在某些情况下，该耳机还包括与该多个第一扬声器一一对应的多个FB滤波器，也即是，该多个FB滤波器与该多个FF滤波器一一对应。此时，该多组目标降噪参数还包括该多个FB滤波器的第k帧滤波系数。而且，在该耳机还包括下行补偿滤波器的情况下，该多组目标降噪参数还包括下行补偿滤波器的第k帧滤波系数。另外，在k大于1的情况下，还可以确定目标降噪档位。因此接下来将对这四部分内容分别进行介绍。
需要说明的是，该多组目标降噪参数也可以称为多个降噪通道的降噪参数，一个降噪通道包括一个FF滤波器以及一个第一扬声器。在该耳机还包括与该多个第一扬声器一一对应的多个FB滤波器的情况下，一个降噪通道还包括一个FB滤波器。
(1)确定该多个FF滤波器的第k帧滤波系数。
在k等于1的情况下，将该多个FF滤波器的初始滤波系数确定为该多个FF滤波器的第k帧滤波系数，即，该多个FF滤波器的第1帧滤波系数为相应FF滤波器的初始滤波系数，或者，基于初始降噪档位、以及降噪档位与FF滤波系数的映射关系，确定该多个FF滤波器的第k帧滤波系数。在k大于1的情况下，基于该至少一个参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号以及目标降噪档位，确定该多个FF滤波器的第k帧滤波系数。即，通过自适应方法来确定该多个FF滤波器的第k帧滤波系数，该确定过程是一个自适应过程，也可以称为迭代过程。
需要说明的是，该多个FF滤波器的初始滤波系数可以相同，也可以不同，该初始滤波系数可以为0，也可以不为0，本申请实施例对此不做限定。该初始降噪档位可以是事先设置的档位，该档位是指对应的降噪系数能够正常进行降噪同时不至于引入稳定性问题的档位。当然，该初始降噪档位也可以是在降噪开始时，通过用户终端发送的“降噪开”、“叮咚”等提示音确定的一个档位，该档位对应的降噪系数与当前的人耳及佩戴姿态能更好适配，而且在该档位对应的降噪系数的基础上进行自适应迭代能够更快达到收敛状态，本申请实施例同样对此不做限定。
其中，基于该至少一个参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号以及目标降噪档位，确定该多个FF滤波器的第k帧滤波系数的实现过程包括：基于目标降噪档位、以及降 噪档位与SP的滤波系数的映射关系，确定多个SP的第k-1帧滤波系数，该多个SP是指该多个第一扬声器到误差麦克风的路径。基于该至少一个参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号和该多个SP的第k-1帧滤波系数，确定该多个FF滤波器的第k帧滤波系数。
该多个SP也可以称为多个降噪通道的SP，降噪档位与SP的滤波系数的映射关系包括多个降噪档位，每个降噪档位均与该多个SP的滤波系数存在映射关系，不同的降噪档位与该多个SP的滤波系数的映射关系可能不同，所以，在确定出目标降噪档位之后，基于目标降噪档位，从该降噪档位与SP的滤波系数的映射关系中，能够获取到该多个SP对应的滤波系数，将获取的滤波系数作为该多个SP的第k-1帧滤波系数。初始降噪档位同理。
在确定该多个FF滤波器的第k帧滤波系数时，可以通过多通道联动的方式来确定。而且，在该耳机包括多个FF滤波器的情况下，可能还包括与该多个第一扬声器一一对应的多个FB滤波器，也可能不包括该多个FB滤波器。在不同的情况下，确定该多个FF滤波器的第k帧滤波系数的方式不同。接下来将分别进行介绍。
由于基于该至少一个参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号和该多个SP的第k-1帧滤波系数，确定每个FF滤波器的第k帧滤波系数的确定过程相同，因此接下来以其中一个为例进行介绍。即，将该多个FF滤波器中的一个FF滤波器作为目标FF滤波器，按照如下方式确定目标FF滤波器的第k帧滤波系数，该多个FF滤波器中其他FF滤波器的第k帧滤波系数的确定过程可以参考目标FF滤波器的第k帧滤波系数的确定过程。
第一种情况，该耳机不包括该多个FB滤波器。如果目标FF滤波器是首个FF滤波器，则基于目标参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号、以及目标SP的第k-1帧滤波系数，确定目标FF滤波器的第k帧滤波系数，目标参考麦克风为目标FF滤波器对应的参考麦克风，目标SP是指目标FF滤波器对应的第一扬声器到误差麦克风的路径。如果目标FF滤波器是非首个FF滤波器，则基于目标参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号、该多个SP的第k-1帧滤波系数、以及位于目标FF滤波器之前的各个FF滤波器的第k帧频响信息和第k-1帧频响信息，确定目标FF滤波器的第k帧滤波系数。
在目标FF滤波器是首个FF滤波器的情况下，基于目标参考麦克风采集的第k-1帧参考信号和误差麦克风采集的第k-1帧误差信号，确定残留误差，基于目标FF滤波器的第k-1帧频响信息、目标SP的第k-1帧滤波系数和该残留误差，确定目标FF滤波器的第k帧频响信息。基于目标FF滤波器的第k帧频响信息，确定目标FF滤波器的第k帧滤波系数。
在一些实施例中，该多个FF滤波器中的一个FF滤波器对应一个参考麦克风。即，目标参考麦克风包括一个参考麦克风。此时，基于目标参考麦克风采集的第k-1帧参考信号和误差麦克风采集的第k-1帧误差信号，按照如下公式(1)确定残留误差。
其中，在上述公式(1)中，Resk-1是指残留误差，Refk-1是指目标参考麦克风采集的第k-1帧参考信号，Errk-1是指误差麦克风采集的第k-1帧误差信号。
在另一些实施例中，该多个FF滤波器中的一个FF滤波器对应至少两个参考麦克风。即，目标参考麦克风包括至少两个参考麦克风。此时，将目标参考麦克风包括的至少两个参考麦克风采集的第k-1帧参考信号进行混音，以得到第k-1帧混音参考信号。基于第k-1帧混音参考信号和误差麦克风采集的第k-1帧误差信号，确定残留误差。这样，可以提升参考信号的信噪比。
其中，基于第k-1帧混音参考信号和误差麦克风采集的第k-1帧误差信号确定残留误差的方式，与上述通过公式(1)确定残留误差的方式类似，即，将误差麦克风采集的第k-1帧误差信号除以第k-1帧混音参考信号，以得到残留误差。
在一些实施例中，可以确定目标SP的第k-1帧滤波系数的频响信息，然后，基于目标FF滤波器的第k-1帧频响信息、目标SP的第k-1帧滤波系数的频响信息和该残留误差，按照如下公式(2)确定目标FF滤波器的第k帧频响信息。
在上述公式(2)中，FFk是指目标FF滤波器的第k帧频响信息，FFk-1是指目标FF滤波器的第k-1帧频响信息，μ是指步长，事先设置的，SPk-1是指目标SP的第k-1帧滤波系数的频响信息。
在目标FF滤波器为非首个FF滤波器的情况下，基于目标参考麦克风采集的第k-1帧参考信号和误差麦克风采集的第k-1帧误差信号，确定残留误差。基于目标FF滤波器的第k-1帧频响信息、该残留误差、该多个SP的第k-1帧滤波系数、以及位于目标FF滤波器之前的各个FF滤波器的第k帧频响信息和第k-1帧频响信息，确定目标FF滤波器的第k帧频响信息。基于目标FF滤波器的第k帧频响信息，确定目标FF滤波器的第k帧滤波系数。
其中，确定残留误差的方式与上文相同，详细实现过程请参考上述内容，此处不再赘述。
确定目标FF滤波器的第k帧频响信息时，可以基于目标FF滤波器的第k-1帧频响信息、该残留误差、目标SP的第k-1帧滤波系数、位于目标FF滤波器之前的各个FF滤波器的第k帧频响信息和第k-1帧频响信息、以及位于目标FF滤波器之前的各个FF滤波器对应的SP的第k-1帧滤波系数，确定目标FF滤波器的第k帧频响信息。
作为一种示例，可以确定目标SP的第k-1帧滤波系数的频响信息、以及位于目标FF滤波器之前的各个FF滤波器对应的SP的第k-1帧滤波系数的频响信息，然后，基于目标FF滤波器的第k-1帧频响信息、该残留误差、目标SP的第k-1帧滤波系数的频响信息、位于目标FF滤波器之前的各个FF滤波器的第k帧频响信息和第k-1帧频响信息、以及位于目标FF滤波器之前的各个FF滤波器对应的SP的第k-1帧滤波系数的频响信息，按照如下公式(3)确定目标FF滤波器的第k帧频响信息。
其中，在上述公式(3)中，FFi,k是指目标FF滤波器的第k帧频响信息，即目标FF滤波器为该多个FF滤波器中的第i个FF滤波器，FFi,k-1是指目标FF滤波器的第k-1帧频响信息，Resi,k-1是指该残留误差，SPi,k-1是指目标SP的第k-1帧滤波系数的频响信息，FFj,k是指位于目标FF滤波器之前的第j个FF滤波器的第k帧频响信息，FFj,k-1是指位于目标FF滤波器之前的第j个FF滤波器的第k-1帧频响信息，SPj,k-1是指位于目标FF滤波器之前的第j个FF滤波器对应的SP的第k-1帧滤波系数的频响信息。
其中，基于目标FF滤波器的第k帧频响信息，确定目标FF滤波器的第k帧滤波系数的实现过程包括：建立目标FF滤波器的滤波系数变量与目标FF滤波器的第k帧频响信息之间的损失函数。基于该损失函数，通过梯度下降法确定滤波系数变量的值，基于该滤波系数变量的值确定目标FF滤波器的第k帧滤波系数。也即是，建立目标FF滤波器的滤波系数变量与目标FF滤波器的第k帧频响信息之间的损失函数。通过梯度下降法确定该变量的最优值，从而通过该变量的最优值确定目标FF滤波器的第k帧滤波系数。
目标FF滤波器的每帧滤波系数都是按照梯度下降法确定的，在确定目标FF滤波器的每帧滤波系数时都会确定出一个损失函数的值，在该损失函数的值达到最小门限值时，确定目标FF滤波器的滤波系数达到收敛稳定条件。比如，对于目标FF滤波器的第k帧滤波系数来说，在滤波系数变量与目标FF滤波器的第k帧频响信息之间的损失函数的值达到最小门限值时，确定目标FF滤波器的第k帧滤波系数达到收敛稳定条件。在该损失函数的值未达到最小门限值时，确定目标FF滤波器的第k帧滤波系数未达到收敛稳定条件。该最小门限值是事先设置的，在不同的情况下，可以按照不同的需求进行调整。
可选地，每个FF滤波器的滤波系数包括至少一个双二阶滤波系数和一个增益。该双二阶滤波系数对应的变量包括滤波器类型、截止频率和品质因数。当然，实际应用中，每个FF滤波器的滤波系数还可以包括其他更多或更少的参数，本申请实施例对此不做限定。
其中，可以基于该滤波系数变量的值，按照相关算法确定第一FF滤波器的第k帧滤波系数，本申请实施例对该算法不做限定。
在某些情况下，安静环境具有底噪问题，即背景噪声，比如对于半开放耳机来说，该耳机在安静环境下相比入耳式耳机更易出现底噪问题。而且安静环境不需要强降噪，部分人在安静环境下强降噪会有不适感。并且降噪力度越大的情况下，人的负压感越强，所以，通过梯度下降法确定滤波系数变量的值时，可以基于环境音量动态调节目标降噪幅度，从而按照目标降噪幅度来确定目标FF滤波器的第k帧滤波系数，提高自适应降噪的主观体验效果。即，根据第k-1帧的环境音量以及位于第k-1帧之前的t帧环境音量确定目标降噪幅度，t大于等于1且小于k-1。基于目标降噪幅度和该损失函数，通过梯度下降法确定滤波系数变量的值，基于该滤波系数变量的值确定目标FF滤波器的第k帧滤波系数。
根据第k-1帧的环境音量以及位于第k-1帧之前的t帧环境音量确定目标环境音量。如果目标环境音量小于等于第一音量阈值，则将第一降噪幅度确定为目标降噪幅度。如果目标环境音量大于第一音量阈值，则确定目标环境音量是否显著增大或显著减小，如果目标环境音量显著增大，则增大第k-1帧的降噪幅度，以得到目标降噪幅度。如果环境音量显著减小，则减小第k-1帧的降噪幅度，以得到目标降噪幅度。如果目标环境音量没有显著增大且没有显著减小，则将第k-1帧的降噪幅度确定为目标降噪幅度，即维持降噪幅度不变。
根据第k-1帧的环境音量以及位于第k-1帧之前的t帧环境音量确定目标环境音量的方式包括多种，比如，取算术平均值、加权平均值等等，本申请实施例对此不做限定。其中，该t帧可以为位于第k-1帧之前的任意t帧，也可以为位于第k-1帧之前且距离第k-1帧最近的t帧，本申请实施例对此不做限定。
需要说明的是，第一音量阈值是事先设置的，第一音量阈值用于表征当前是否处于安静环境。即，如果目标环境音量小于等于第一音量阈值，则表明处于安静环境，如果目标环境音量大于第一音量阈值，则表明处于非安静环境。第一降噪幅度是事先为安静环境设置的，用于进行弱降噪，避免过于放大底噪或引入主观舒适度问题。实际应用中，可以按照不同的需求，对第一音量阈值和第一降噪幅度进行调整。
比如，请参考图3，通过目标环境音量确定是否处于安静环境，在处于安静环境的情况下，将第一降噪幅度确定为目标降噪幅度。在非安静环境的情况下，如果目标环境音量显著增大，则增大第k-1帧的降噪幅度，以得到目标降噪幅度。如果目标环境音量显著减小，则减小第k-1帧的降噪幅度，以得到目标降噪幅度。如果目标环境音量没有显著增大且没有显著减小，则将第k-1帧的降噪幅度确定为目标降噪幅度，即维持降噪幅度不变。
目标环境音量是否显著增大或显著减小的确定方式包括多种，比如，如果本次确定的目标环境音量大于上次确定的目标环境音量，且本次确定的目标环境音量与上次确定的目标环境音量的差值大于第二音量阈值，确定本次确定的目标环境音量显著增大。同理，如果本次确定的目标环境音量小于上次确定的目标环境音量，且本次确定的目标环境音量与上次确定的目标环境音量的差值大于第二音量阈值，确定本次确定的目标环境音量显著减小。
第二音量阈值也是事先设置的，比如3dB。实际应用中，还可以按照不同的需求，对第第二音量阈值进行调整。
第二种情况，该耳机还包括该多个FB滤波器。如果目标FF滤波器是首个FF滤波器，则基于目标参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号、该多个SP的第k-1帧滤波系数和该多个FB滤波器的第k-1帧滤波系数，确定目标FF滤波器的第k帧滤波系数。如果目标FF滤波器是非首个FF滤波器，则基于目标参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号、该多个SP的第k-1帧滤波系数、该多个FB滤波器的第k-1帧滤波系数、以及位于目标FF滤波器之前的各个FF滤波器的第k帧频响信息和第k-1帧频响信息，确定目标FF滤波器的第k帧滤波系数。
在目标FF滤波器为首个FF滤波器的情况下，可以基于目标参考麦克风采集的第k-1帧参考信号和误差麦克风采集的第k-1帧误差信号，确定残留误差。基于目标FF滤波器的第k-1帧频响信息、该残留误差、该多个FB滤波器的第k-1帧滤波系数、以及该多个SP的第k-1帧滤波系数，确定目标FF滤波器的第k帧频响信息。基于目标FF滤波器的第k帧频响信息，确定目标FF滤波器的第k帧滤波系数。
作为一种示例，可以确定该多个FB滤波器的第k-1帧滤波系数的频响信息、以及该多个SP的第k-1帧滤波系数的频响信息，然后，基于目标FF滤波器的第k-1帧频响信息、该残留误差、该多个FB滤波器的第k-1帧滤波系数的频响信息、以及该多个SP的第k-1帧滤波系数的频响信息，按照如下公式(4)确定目标FF滤波器的第k帧频响信息。
其中，在上述公式(4)中，FF1,k是指目标FF滤波器的第k帧频响信息，FF1,k-1是指目标FF滤波器的第k-1帧频响信息，Res1,k-1是指该残留误差，SP1,k-1是指目标SP的第k-1帧滤波系数的频响信息，FBj,k-1是指该多个FB滤波器中第j个FB滤波器的第k-1帧滤波系数的频响信息，SPj,k-1是指第j个FB滤波器对应的SP的第k-1帧滤波系数的频响信息，n是指该多个FB滤波器的总数，即该多个降噪通道的总数。
在目标FF滤波器为非首个FF滤波器的情况下，可以基于目标参考麦克风采集的第k-1帧参考信号和误差麦克风采集的第k-1帧误差信号，确定残留误差。基于目标FF滤波器的第k-1帧频响信息、该残留误差、该多个SP的第k-1帧滤波系数、该多个FB滤波器的第k-1帧滤波系数、以及位于目标FF滤波器之前的各个FF滤波器的第k帧频响信息和第k-1帧频响信息，确定目标FF滤波器的第k帧频响信息。基于目标FF滤波器的第k帧频响信息，确定目标FF滤波器的第k帧滤波系数。
作为一种示例，可以确定该多个SP的第k-1帧滤波系数的频响信息、该多个FB滤波器的第k-1帧滤波系数的频响信息，然后，基于目标FF滤波器的第k-1帧频响信息、该残留误差、该多个SP的第k-1帧滤波系数的频响信息、该多个FB滤波器的第k-1帧滤波系数的频响信息、以及位于目标FF滤波器之前的各个FF滤波器的第k帧频响信息和第k-1帧频响信息，按照如下公式(5)确定目标FF滤波器的第k帧频响信息。
其中，在上述公式(5)中，FBj,k-1是指该多个FB滤波器中第j个FB滤波器的第k-1帧滤波系数的频响信息，n是指该多个FB滤波器的总数，即该多个降噪通道的总数，其他字母代表的含义与上述公式(3)中的含义相同。
基于目标FF滤波器的第k帧频响信息，确定目标FF滤波器的第k帧滤波系数的实现过程，与上述第一种情况相同，详细内容请参考上文描述，此次不再赘述。而且，上述SP的滤波系数的频响信息可以基于SP的滤波系数，按照相关算法确定得到，FB滤波系数的频响信息也可以基于FB滤波器的滤波系数，按照相关算法确定得到，本申请实施例对该算法不做限定。
在上述确定目标FF滤波器的第k帧频响信息的过程中，不管该耳机是否包括目标FB滤波器，目标FF滤波器的第k帧频响信息都是基于目标SP的第k-1帧滤波系数来确定的，而目标SP的第k-1帧滤波系数是基于目标降噪档位通过查询降噪档位与SP的滤波系数的映射关系来确定的，也即是，目标SP的第k-1帧滤波系数是一个估计值，通过这个估计值来确定目标FF滤波器的第k帧频响信息，可以摆脱对目标SP真实值的依赖，从而在没有下行信号的情况下，也能够实现FF滤波器的滤波系数的自适应。
(2)确定该多个FB滤波器的第k帧滤波系数。
在k等于1的情况下，将该多个FB滤波器的初始滤波系数确定为该多个FB滤波器的第k帧滤波系数，即，该多个FB滤波器的第1帧滤波系数为相应FB滤波器的初始滤波系数，或者，基于初始降噪档位、以及降噪档位与FB滤波系数的映射关系，确定该多个FB滤波器的第k帧滤波系数。在k大于1的情况下，可以基于目标降噪档位确定该多个FB滤波器的第k帧滤波系数。
需要说明的是，该多个FB滤波器的初始滤波系数可以相同，也可以不同，该初始滤波系数可以为0，也可以不为0，本申请实施例对此不做限定。
由于基于目标降噪档位确定每个FB滤波器的第k帧滤波系数的过程相同，因此，接下来以其中一个为例进行介绍。即，将该多个FB滤波器中的一个FB滤波器作为目标FB滤波器，按照如下两种方式确定目标FB滤波器的第k帧滤波系数，该多个FB滤波器中其他FB滤波器的第k帧滤波系数的确定过程可以参考目标FB滤波器的第k帧滤波系数的确定过程。也就是说，在k大于1的情况下，可以按照如下两种方式确定目标FB滤波器的第k帧滤波系数。
第一种方式，基于目标降噪档位，以及降噪档位与FB滤波系数的映射关系，确定目标FB滤波器的第 k帧滤波系数。
降噪档位与FB滤波系数的映射关系中包括多个降噪档位，每个降噪档位均与该多个FB滤波器的滤波系数存在映射关系，不同的降噪档位与该多个FB滤波器的滤波系数的映射关系可能不同，所以，基于目标降噪档位，从该降噪档位与FB滤波系数的映射关系中，能够获取到目标FB滤波器对应的滤波系数，将获取的滤波系数作为目标FB滤波器的第k帧滤波系数。
由于降噪档位与FB滤波系数的映射关系是事先存储的，所以通过第一种方式确定目标FB滤波器的第k帧滤波系数的稳定性较好，操作也较为简单，效率较高。
第二种方式，如果目标FB滤波器属于第一类FB滤波器，则基于目标降噪档位，以及降噪档位与FB滤波系数的映射关系，确定目标FB滤波器的第k帧滤波系数，如果目标FB滤波器属于第二类FB滤波器，则基于误差麦克风采集的第k-1帧误差信号、目标FB滤波器的第k-1帧滤波系数以及目标降噪档位，确定目标FB滤波器的第k帧滤波系数。
与上文同理，在目标FB滤波器属于第二类FB滤波器的情况下，可以通过自适应方法来确定目标FB滤波器的第k帧滤波系数。目标FB滤波器的第k帧滤波系数的确定过程是一个自适应过程，也可以称为迭代过程。
基于误差麦克风采集的第k-1帧误差信号、目标FB滤波器的第k-1帧滤波系数以及目标降噪档位，确定目标FB滤波器的第k帧滤波系数的实现过程包括：基于目标降噪档位，以及降噪档位与SP的滤波系数的映射关系，确定目标SP的第k-1帧滤波系数，目标SP是指目标FB滤波器对应的第一扬声器到误差麦克风的路径；基于误差麦克风采集的第k-1帧误差信号、目标FB滤波器的第k-1帧滤波系数以及目标SP的第k-1帧滤波系数，确定目标FB滤波器的第k帧滤波系数。
其中，可以按照相关算法，基于误差麦克风采集的第k-1帧误差信号、目标FB滤波器的第k-1帧滤波系数以及目标SP的第k-1帧滤波系数，确定目标FB滤波器的第k帧滤波系数，本申请实施例对该算法不做限定。
由于目标FB滤波器的第一帧滤波系数可以为初始滤波系数，也可以通过初始降噪档位查询降噪档位与FB滤波系数的映射关系确定得到，所以，在k大于等于1的情况下，相当于可以通过三种方式来确定目标FB滤波器的第k帧滤波系数。即，(1)通过查询降噪档位与FB滤波系数的映射关系来确定目标FB滤波器的第k帧滤波系数。(2)如果目标FB滤波器属于第一类FB滤波器，则通过查询降噪档位与FB滤波系数的映射关系来确定目标FB滤波器的第k帧滤波系数，如果目标FB滤波器属于第二类FB滤波器，则基于误差麦克风采集的第k-1帧误差信号、目标FB滤波器的第k-1帧滤波系数以及目标降噪档位，确定目标FB滤波器的第k帧滤波系数。(3)如果目标FB滤波器属于第一类FB滤波器，或者，目标FB滤波器属于第二类FB滤波器且k等于1，则通过查询降噪档位与FB滤波系数的映射关系来确定目标FB滤波器的第k帧滤波系数。如果目标FB滤波器属于第二类FB滤波器且k大于1，则基于误差麦克风采集的第k-1帧误差信号、目标FB滤波器的第k-1帧滤波系数以及目标降噪档位，确定目标FB滤波器的第k帧滤波系数。
第一类FB滤波器对应的第一扬声器的发声频段高于第二类FB滤波器对应的第一扬声器的发声频段。也即是，第一类FB滤波器对应的第一扬声器为高频扬声器，第二类FB滤波器对应的第一扬声器为低频扬声器。当然，第一类FB滤波器和第二类FB滤波器也可以不用按照发声频段来区分，而是按照其他的方式来区分，本申请实施例同样对此不做限定。
上述第二种方式和第三方式，将查询降噪档位与FB滤波系数的映射关系的方式，与自适应的方式结合，不仅能够提高降噪效果，复杂性相对来说也不是很高，而且稳定性也相对可控。
需要说明的是，在本申请实施例中，不仅可以按照上述三种方式来确定目标FB滤波器的第k帧滤波系数，还可以按照其他方式来确定目标FB滤波器的第k帧滤波系数。比如，不管目标FB滤波器属于第一类FB滤波器还是第二类FB滤波器，都基于误差麦克风采集的第k-1帧误差信号、目标FB滤波器的第k-1帧滤波系数以及目标降噪档位，确定目标FB滤波器的第k帧滤波系数。本申请实施例对此不做限定。
(3)确定下行补偿滤波器的第k帧滤波系数。
在k等于1的情况下，将初始下行补偿滤波系数确定为下行补偿滤波器的第k帧滤波系数，或者，基于初始降噪档位、以及降噪档位与下行补偿滤波系数的映射关系，确定下行补偿滤波器的第k帧滤波系数。在k大于1的情况下，基于目标降噪档位，以及降噪档位与下行补偿滤波系数的映射关系，确定下行补偿 滤波器的第k帧滤波系数。
降噪档位与下行补偿滤波系数的映射关系包括多个降噪档位，每个降噪档位均与该下行补偿滤波器的滤波系数存在映射关系，不同的降噪档位与该下行补偿滤波器的滤波系数的映射关系可能不同，所以，在确定出目标降噪档位之后，基于目标降噪档位，从该降噪档位与下行补偿滤波系数的映射关系中，能够获取到对应的下行补偿滤波系数，将获取的下行补偿滤波系数作为该下行补偿滤波器的第k帧滤波系数。
(4)确定目标降噪档位。
确定第k-1帧降噪档位，获取位于第k-1帧之前的m帧降噪档位，m大于等于1且小于k-1。基于第k-1帧降噪档位和该m帧降噪档位，确定目标降噪档位。
由于第k-1帧可能存在有效下行信号，也可能不存在有效下行信号，而且有可能处于安静环境，也可能处于非安静环境，当然，还可能存在异常信号。在不同的情况下，确定第k-1帧降噪档位的方式不同，接下来将分别进行介绍。
第一种情况，第k-1帧不存在有效下行信号且处于非安静环境。在这种情况下，根据该多个FF滤波器的参考滤波系数，以及降噪档位与FF滤波器的频响信息的映射关系，确定第k-1帧降噪档位。其中，在k等于2的情况下，参考滤波系数为相应FF滤波器的初始滤波系数，在k大于2的情况下，参考滤波系数为相应FF滤波器在第k帧之前最近一次达到收敛稳定条件的滤波系数，或者为相应FF滤波器的第k-1帧滤波系数。
在通过耳机播放音频信号时，比如播放音乐或通话，用户终端会向耳机下发播放音频信号的控制信令，所以，可以通过耳机是否接收到该控制信令来确定当前是否处于下行使能状态。在耳机未处于下行使能状态的情况下，确定第k-1帧不存在有效下行信号。在耳机处于下行使能状态的情况下，第k-1帧并一定真正有声音持续输出，比如，在讲话的停顿期、音乐换曲的过渡期等不会有声音输出，而且这个时间往往还不一定短，所以，在耳机处于下行使能状态的情况下，还可以确定第k-1帧是否处于下行间歇期，如果第k-1帧处于下行间歇期，那么确定第k-1帧不存在有效下行信号。如果第k-1帧不处于下行间歇期，那么确定第k-1帧存在有效下行信号。
在第k-1帧不存在有效下行信号但处于非安静环境的情况下，第k-1帧降噪档位可能会随环境噪声的不同而不同，所以，需要基于该多个FF滤波器的参考滤波系数，以及降噪档位与FF滤波器的频响信息的映射关系，确定第k-1帧降噪档位。
在一些实施例中，基于该多个FF滤波器的参考滤波系数，确定该多个FF滤波器的参考频响信息。基于降噪档位与FF滤波器的频响信息的映射关系，确定与该多个FF滤波器的参考频响信息分别匹配的降噪档位，以得到多个参考降噪档位。基于该多个参考降噪档位，确定第k-1帧降噪档位。
其中，可以基于该多个FF滤波器的参考滤波系数，按照相关算法确定该多个FF滤波器的参考频响信息，本申请实施例对该算法不做限定。
在降噪档位不同的情况下，FF滤波器的频响信息也可能不同，所以，可以事先存储降噪档位与FF滤波器的频响信息的映射关系。这样，在确定出该多个FF滤波器的参考频响信息之后，对于该多个FF滤波器中的任一FF滤波器来说，将该FF滤波器的参考频响信息与该映射关系中不同降噪档位下FF滤波器的频响信息进行匹配，以从该映射关系中确定与该FF滤波器的参考频响信息匹配的频响信息，进而将匹配的频响信息对应的降噪档位作为一个参考降噪档位。对于该多个FF滤波器中的其他FF滤波器，按照相同的方式进行处理，从而能够得到多个参考降噪档位。
FF滤波器的频响信息可以通过频响曲线的方式来表征，所以，在确定出该多个FF滤波器的参考频响曲线之后，对于该多个FF滤波器中的任一FF滤波器来说，将该FF滤波器的参考频响曲线与该映射关系中不同降噪档位下FF滤波器的频响曲线进行匹配。
在实际应用中，可以将该FF滤波器的完整的参考频响曲线与该映射关系中不同降噪档位下FF滤波器的完整的频响曲线进行匹配。还可以将该FF滤波器的参考频响曲线中位于目标频段内的曲线与该映射关系中不同降噪档位下FF滤波器的频响曲线中位于目标频段内的曲线进行匹配，本申请实施例对此不做限定。
需要说明的是，目标频段是指频响曲线中特征区分明显的频段，目标频段是事先设置的，比如，目标频段为100至200赫兹(Hz)的频段。当然，在耳机的声学状况不同的情况下，目标频段的取值也可能不同。
比如，降噪档位与FF滤波器的频响信息的映射关系包括16个降噪档位下FF滤波器的频响曲线，且该16个降噪档位下FF滤波器的频响曲线如图4所示。由于图4中100至200Hz的频段内的特征区分比较明显，所以，将100至200Hz作为目标频段。然后，将该FF滤波器的参考频响曲线中位于100至200Hz内的曲线与这16个降噪档位下FF滤波器的频响曲线中位于100至200Hz内的曲线进行匹配。
其中，基于该多个参考降噪档位确定第k-1帧降噪档位的方式包括多种，比如，按照该多个参考降噪档位的平均值确定第k-1帧降噪档位。或者，按照该多个参考降噪档位中数量最多的参考降噪档位确定第k-1帧降噪档位。
按照该多个参考降噪档位的平均值确定第k-1帧降噪档位时，可以直接将该多个参考降噪档位的平均值确定为第k-1帧降噪档位，也可以对该多个参考降噪档位的平均值进行调整，以得到第k-1帧降噪档位。同理，按照该多个参考降噪档位中数量最多的参考降噪档位确定第k-1帧降噪档位时，可以直接将该多个参考降噪档位中数量最多的参考降噪档位确定为第k-1帧降噪档位，也可以对该多个参考降噪档位中数量最多的参考降噪档位进行调整，以得到第k-1帧降噪档位。本申请实施例对此不做限定。
由于FF滤波器的滤波系数的确定过程是一个迭代过程，也可以称为自适应过程，所以，上述收敛稳定条件是指FF滤波器的滤波系数收敛至基本不发生变化。而且，由于整个降噪过程中FF滤波器的滤波系数可能会进行多次自适应调整，所以，在确定FF滤波器的第k帧滤波系数时，可以将该FF滤波器在第k帧之前最近一次达到收敛稳定条件的滤波系数作为参考滤波系数，也可以将该FF滤波器的第k-1帧滤波系数作为参考滤波系数。
第二种情况，第k-1帧存在有效下行信号。在这种情况下，基于第k-1帧的有效下行信号、该至少一个参考麦克风采集的第k-1帧参考信号，以及误差麦克风采集的第k-1帧误差信号，确定第k-1帧降噪档位。
基于上文描述，在耳机处于下行使能状态且未处于下行间歇期的情况下，确定第k-1帧存在有效下行信号。此时，可以基于第k-1帧的有效下行信号、该至少一个参考麦克风采集的第k-1帧参考信号，以及误差麦克风采集的第k-1帧误差信号，从误差麦克风采集的第k-1帧误差信号中提取有效下行信号，从而基于提取的有效下行信号确定第k-1帧降噪档位。
其中，可以按照相关算法，基于第k-1帧的有效下行信号、该至少一个参考麦克风采集的第k-1帧参考信号，以及误差麦克风采集的第k-1帧误差信号，从误差麦克风采集的第k-1帧误差信号中提取有效下行信号，从而基于提取的有效下行信号确定第k-1帧降噪档位，本申请实施例对该算法不做限定。
第三种情况，第k-1帧不存在有效下行信号且处于安静环境，或者，第k-1帧存在异常噪声信号。在这种情况下，将第k-3帧降噪档位确定为第k-1帧降噪档位。即，维持降噪档位不变。
在第k-1帧不存在有效下行信号且处于安静环境的情况下，噪声基本没有发生变化，此时，可以维持降噪档位不变。在第k-1帧存在异常噪声信号的情况下，通过维持降噪档位不变，以进行鲁棒性控制，从而避免降噪档位的发散。
异常噪声信号是指对用户的听音感受带来比较严重影响的信号，比如啸叫、削波、底噪、风噪等等。其中，啸叫是单频声音信号的幅度或能量由小突然增大的现象，通常是耳机在受到挤压、或者用户快速改变耳机的佩戴姿态等动作而导致的，将啸叫时发出的声音信号称为啸叫噪声，啸叫会引起用户不适，并且干扰下行信号的播放，严重影响音频的播放效果。削波是低频信号溢出产生噼啪杂音的现象，将产生的噼啪声称为削波噪声。通常，在环境中突发低频大噪声时，将发生削波，例如车辆大颠簸、飞机着陆过程中会时产生低频大噪声。底噪即本底噪声，底噪也可以称为背景噪声，底噪是由于设备的硬件(例如耳机中的电路或其他元器件)性能限制而导致的噪声，例如电视声中除节目声音外的沙沙声等。在嘈杂的环境中，底噪一般是无法被用户感知或听到的，当环境安静时，用户可以感知底噪。过强的本底噪声，不仅会使人烦躁，还会淹没声音中较弱的细节部分。风噪是环境中有风时，而产生的呼呼声，风噪影响用户正常使用耳机。并且由于风噪的方向的随机性比较大，风噪对用户双耳的影响是不同的，即左耳和右耳在风噪的影响下，听感不一致。
接下来通过图5对上述三种情况进行简单概括。请参考图5，在第k-1帧存在异常噪声信号的情况下，维持降噪档位不变。在第k-1帧不存在异常噪声信号的情况下，确定第k-1帧是否下行使能。在第k-1帧未下行使能的情况下，确定第k-1帧是否处于安静环境。在第k-1帧处于安静环境的情况下，维持降噪档位不变。在第k-1帧处于非安静环境的情况下，根据FF滤波器的参考滤波系数确定对应的参考降噪档位， 在所有通道都轮询完成后，基于该多个参考降噪档位确定第k-1帧降噪档位。在第k-1帧下行使能的情况下，确定第k-1帧是否处于下行间歇期，在第k-1帧处于下行间歇期的情况下，按照上述相同的方式确定多个参考降噪档位，进而基于该多个参考降噪档位确定第k-1帧降噪档位。在第k-1帧未处于下行间歇期的情况下，基于第k-1帧的有效下行信号、该至少一个参考麦克风采集的第k-1帧参考信号，以及误差麦克风采集的第k-1帧误差信号，确定第k-1帧降噪档位。
通过上述三种情况确定出第k-1帧降噪档位之后，可以将第k-1帧降噪档位和位于第k-1帧之前的m帧降噪档位进行综合，以确定目标降噪档位。
其中，该m帧降噪档位可以为位于第k-1帧之前的任意m帧降噪档位，也可以为位于第k-1帧之前且距离第k-1帧最近的m帧降噪档位，本申请实施例对此不做限定。另外，基于第k-1帧降噪档位和位于第k-1帧之前的m帧降噪档位确定目标降噪档位的实现方式包括多种，比如，按照相关算法进行降噪效果的评优，以确定第k-1帧降噪档位对应的降噪概率以及该m帧降噪档位分别对应的降噪概率，将降噪概率最大的降噪档位确定为目标降噪档位。或者，确定第k-1帧降噪档位与该m帧降噪档位的算术平均值或者加权平均值，以得到目标降噪档位。又或者，将第k-1帧降噪档位和该m帧降噪档位中出现数量最多的降噪档位确定为目标降噪档位等等。
上述涉及的各种映射关系都是事先确定得到的，比如，在该多个第一扬声器中的一个第一扬声器工作，其他的第一扬声器不工作的情况下，根据多种泄漏状态中的每种泄漏状态下该至少一个参考麦克风采集的参考信号和误差麦克风采集的误差信号确定的，该多种泄漏状态由耳机与多种不同的耳道环境形成，该多种泄漏状态与多个降噪档位一一对应。
至此，已完成该多组目标降噪参数的确定。接下来通过图6对该多组目标降噪参数的一种确定过程进行简单地示例性概括。请参考图6，可以通过离线设置初值，包括上述的初始降噪档位、初始滤波系数以及各种映射关系。然后确定第k-1帧是否存在有效下行信号、是否处于安静环境、是否存在异常噪声信号，从而基于不同的情况确定第k-1帧降噪档位。通过第k-1帧降噪档位以及之前的m帧降噪档位确定目标降噪档位。之后，基于第k-1帧的环境音量确定目标降噪幅度，并基于目标降噪档位进行FB滤波系数的自适应，以确定该多个FB滤波器的第k帧滤波系数。最后，基于目标降噪档位和目标降噪幅度进行FF滤波系数的自适应，以确定该多个FF滤波器的第k帧滤波系数。
步骤202：基于该多组目标降噪参数，生成与该多个第一扬声器一一对应的多组目标反相噪声，该多组目标反相噪声中每个目标反相噪声的频段均覆盖该多个第一扬声器的发声频段。
基于上文描述，该多组目标降噪参数可以称为多个降噪通道的降噪参数，这样，生成的该多组目标反相噪声也可以称为多个降噪通道的反相噪声。由于每个降噪通道的反相噪声的生成过程相同，因此接下来以其中一个降噪通道为例进行说明。
将该多个降噪通道中的一个降噪通道作为目标降噪通道，目标降噪通道包括目标FF滤波器和目标第一扬声器，目标FF滤波器对应的参考麦克风称为目标参考麦克风。此时，目标反相噪声包括前馈反相噪声。即，基于目标FF滤波器的第k帧滤波系数，对目标参考麦克风采集的第k帧参考信号进行处理，以得到前馈反相噪声。
基于上文描述，目标参考麦克风可能包括一个参考麦克风，也可能包括至少两个参考麦克风。在目标参考麦克风包括一个参考麦克风的情况下，可以直接基于目标FF滤波器的第k帧滤波系数，对目标参考麦克风采集的第k帧参考信号进行处理，以得到前馈反相噪声。在目标参考麦克风包括至少两个参考麦克风的情况下，将该至少两个参考麦克风采集的第k帧参考信号进行混音，以得到第k帧混音参考信号，然后，基于目标FF滤波器的第k帧滤波系数，对第k帧混音参考信号进行处理，以得到前馈反相噪声。
在该耳机还包括FB滤波器的情况下，目标降噪通道还包括目标FB滤波器。此时，目标反相噪声还包括反馈反相噪声。即，基于下行补偿滤波器的第k帧滤波系数，对用户终端发送的第k帧下行信号进行下行补偿。然后，将下行补偿后的第k帧下行信号取反后与误差麦克风采集的第k帧误差信号进行混音，以得到误差麦克风采集的第k帧噪声信号。基于目标FB滤波器的第k帧滤波系数，对误差麦克风采集的第k帧噪声信号进行处理，以得到反馈反相噪声。
通过下行补偿能够将误差麦克风采集的误差信号中的所有下行信号去掉，从而只将残留的噪声信号通过FB滤波器做降噪，以避免对下行信号造成音质损伤。并且，通过对用户终端发送的第k帧下行信号进行下行补偿，能够将所有扬声器在误差麦克风处的下行信号都去掉，从而避免对全频的下行信号造成音质 损伤。
基于上文描述，在该多组目标降噪参数以帧为单位来确定的情况下，由于一帧可能包括一个样点，也可能包括多个样点，所以，在生成目标反相噪声时，可以在每个样点处都生成一组目标反相噪声，也可以一帧生成一组目标反相噪声。
由于本申请实施例在确定多组目标降噪参数时，对下行信号并未进行分频，也即是，通过全频的下行信号来确定该多组目标降噪参数。这样，基于该多组目标降噪参数，生成与该多个第一扬声器一一对应的多组目标反相噪声之后，该多组目标反相噪声中每个目标反相噪声的频段均覆盖该多个第一扬声器的发声频段，即每个目标反相噪声的频段均为全频段。
步骤203：利用该多组目标反相噪声，通过该多个第一扬声器进行降噪。
在生成多组目标反相噪声之后，将该多组目标反相噪声分别与该多个第一扬声器待播放的第k帧下行信号进行混音后，通过对应的第一扬声器进行播放，从而实现降噪的目的。
该多个第一扬声器可能一部分为高频扬声器，另一部分为低频扬声器。或者，一部分为全频扬声器，另一部分为非全频扬声器。也即是，该多个第一扬声器的发声频段可能不同。或者，该多个第一扬声器全部为全频扬声器。又或者，该多个第一扬声器全部为非全频扬声器。在该多个第一扬声器全部为全频扬声器的情况下，该多个第一扬声器待播放的第k帧下行信号均为用户终端发送的第k帧下行信号。在该多个第一扬声器不全为全频扬声器的情况下，需要按照每个第一扬声器的发声频段，对用户终端发送的第k帧下行信号进行分频，从而得到每个第一扬声器待播放的第k帧下行信号。
该多个第一扬声器中的两个第一扬声器可以包括由一个双振膜(或者称为双动圈)喇叭形成的两个第一扬声器。或者，该多个第一扬声器包括分体喇叭的多个扬声器。
可选地，该耳机还可以包括至少一个第二扬声器，该至少一个第二扬声器不参与降噪，此时，第二扬声器可以参与下行补偿(即，将用户终端发送的下行信号进行下行补偿，该下行信号为全频段的音频信号，包括第二扬声器的发声频段的音频信号)。在这种情况下，第一扬声器可以为中低频的扬声器，也可以为全频的扬声器，第二扬声器可以为高频的扬声器，也可以为中频的扬声器或者低频的扬声器。可选地，第二扬声器也可以不参与下行补偿，此时，第一扬声器可以为中低频的扬声器，也可以为全频的扬声器，第二扬声器为高频的扬声器。
需要说明的是，该至少一个第二扬声器的发声频段高于该至少一个第一扬声器的发声频段。当然，该至少一个第二扬声器的发声频段也可能低于该至少一个第一扬声器的发声频段，本申请实施例对此不做限定。
另外，由于上述通过自适应的方法确定该多组目标降噪参数的过程需要一定的时间，在一帧包括多个样点且一帧的时长较长的情况下，该多组目标降噪参数的确定时长会小于一帧的时长，所以，通过第k-1帧的相关数据，可以在第k帧开始的部分时间段内进行计算，从而得到第k帧的多组目标降噪参数，进而按照第k帧的多组目标降噪参数，在第k帧后续的部分时间段内进行主动降噪。但是，在一帧包括一个样点，或者，一帧包括多个样点且一帧的时长较短的情况下，该多组目标降噪参数的确定时长可能等于一帧的时长，此时，通过第k-1帧的相关数据，可能需要在第k帧的整个时间段内进行计算，从而得到多组目标降噪参数。在这种情况下，可以将该多组目标降噪参数确定为第k+1帧的多组目标降噪参数，进而按照第k+1帧的多组目标降噪参数，在第k+1帧的时间段内进行主动降噪。上述内容都是以前者为例进行介绍的。
综上所述，在本申请实施例中，由于该多组目标反相噪声与该多个第一扬声器一一对应，且该多组目标反相噪声中每个目标反相噪声的频段均覆盖该多个第一扬声器的发声频段，即每个目标反相噪声均为全频段的反相噪声，所以，无论第一扬声器为高频扬声器、低频扬声器还是全频扬声器，在利用该多组目标反相噪声进行降噪时，均能够充分发挥每个第一扬声器的降噪能力。换种方式来讲，在多个降噪通道、多个扬声器的耳机架构下，本方案能够通过多个降噪通道的全频段反相噪声提升耳机降噪效果。
接下来对本申请实施例涉及的几种可能的耳机架构进行示例性说明。
图7是本申请实施例提供的一种耳机的结构示意图。请参考图7，该耳机包括f个参考麦克风、一个误差麦克风、n个FF滤波器、与该n个FF滤波器一一对应的n个FF自适应引擎、n个FB滤波器、与该n个FB滤波器一一对应的n个FB自适应引擎、n个第一扬声器(即扬声器1至扬声器n)、下行补偿滤波 器、下行补偿自适应引擎(图中未示出)、数字分频器、n个均衡(equalizer，EQ)校准器。其中，f和n均为大于等于1的整数，f和n可以相等，也可以不相等。
该f个参考麦克风用于采集外界环境的噪声信号，即参考信号。误差麦克风用于采集耳道内的噪声信号，即误差信号。该n个FF自适应引擎用于确定各自对应的FF滤波器的第k帧滤波系数，并将确定的第k帧滤波系数刷新至对应的FF滤波器中。该n个FB自适应引擎用于确定各自对应的FB滤波器的第k帧滤波系数，并将确定的第k帧滤波系数刷新至对应的FB滤波器中。该下行补偿自适应引擎用于确定下行补偿滤波器的第k帧滤波系数，并将确定的第k帧滤波系数刷新至下行补偿滤波器中。
数字分频器用于按照该n个第一扬声器的发声频段，将用户终端发送的第k帧下行信号进行分频，以得到各个第一扬声器对应的第k帧下行信号。该n个EQ校准器用于校准对应的第一扬声器的量产参数，以使该n个第一扬声器的量产参数的公差一致性对齐。
在进行降噪时，该n个FF滤波器用于基于各自的第k帧滤波系数，对各自对应的参考麦克风采集的第k帧参考信号进行处理，以得到前馈反相噪声。下行补偿滤波器用于基于自身的第k帧滤波系数，对用户终端发送的第k帧下行信号进行下行补偿。然后，将下行补偿后的第k帧下行信号取反后与误差麦克风采集的第k帧误差信号进行混音，以得到误差麦克风采集的第k帧噪声信号。该n个FB滤波器用于基于各自的第k帧滤波系数，对误差麦克风采集的第k帧噪声信号进行处理，以得到反馈反相噪声。之后，将各个降噪通道的前馈反相噪声、反馈反相噪声、第一扬声器的第k帧下行信号进行混音后，通过对应的第一扬声器进行播放，从而实现降噪。
图8是本申请实施例提供的另一种耳机的结构示意图。请参考图8，该耳机包括一个参考麦克风、一个误差麦克风、两个FF滤波器、与该两个FF滤波器一一对应的两个FF自适应引擎、两个FB滤波器、与该两个FB滤波器一一对应的两个FB自适应引擎、两个第一扬声器(即扬声器1和扬声器2)、下行补偿滤波器、下行补偿自适应引擎(图中未示出)、两个EQ校准器。该两个FF滤波器均对应该参考麦克风，该两个第一扬声器为双振膜(或者称为双动圈)喇叭形成的扬声器。此时，该耳机可以不包括数字分频器。而且，这两个第一扬声器可以认为是两个动圈喇叭的组合但共磁路，物理实体上看是一个扬声器。由于两个动圈喇叭在全频段的发声能力都好，因而可以看做是两个全频ANC降噪模块的叠加，这样，可以充分发挥扬声器的降噪能力。
图9是本申请实施例提供的另一种耳机的结构示意图。请参考图9，该耳机包括一个参考麦克风、一个误差麦克风、两个FF滤波器、与该两个FF滤波器一一对应的两个FF自适应引擎、两个FB滤波器、与该两个FB滤波器一一对应的两个FB自适应引擎、两个第一扬声器(即扬声器1和扬声器2)、下行补偿滤波器、下行补偿自适应引擎(图中未示出)、数字分频器、两个EQ校准器。与图8的不同点在于第一扬声器的物理实体是两个，且两个分体的第一扬声器可以不一样，也可以一样。尽管两个第一扬声器不完全一样，但独立来看，每个第一扬声器都能够按照全频降噪单元去设计，这样也就充分发挥每个扬声器的最大降噪能力。
图10是本申请实施例提供的另一种耳机的结构示意图。请参考图10，该耳机包括两个参考麦克风，一个误差麦克风、两个FF滤波器、与该两个FF滤波器一一对应的两个FF自适应引擎、两个FB滤波器、与该两个FB滤波器一一对应的两个FB自适应引擎、两个第一扬声器(即扬声器1和扬声器2)、一个第二扬声器(即扬声器3)、下行补偿滤波器、下行补偿自适应引擎(图中未示出)、数字分频器、三个EQ校准器。图10采用三个扬声器以实现高清音质需求。三个扬声器的频响分别侧重在低、中、高音频段。这三个扬声器中的两个扬声器(即扬声器1和扬声器2)作为第一扬声器参与降噪，另一个扬声器(即扬声器3)作为第二扬声器不参与降噪，但是第二扬声器参与下行补偿(即，将用户终端发送的下行信号进行下行补偿，该下行信号为全频段的音频信号，包括第二扬声器的发声频段的音频信号)。其中，第一扬声器可以为中低频的扬声器，也可以为全频的扬声器，第二扬声器可以为高频的扬声器，也可以为中频的扬声器或者低频的扬声器。
可选地，鉴于当前主流的ANC芯片可能获取不到高频扬声器的信号，所以，请参考图11，第二扬声器也可以不参与下行补偿(即，将用户终端发送的下行信号进行数字分频之后，得到两个第一扬声器对应的下行信号，将两个第一扬声器对应的下行信号进行下行补偿，不包括第二扬声器对应的下行信号)，此时，第一扬声器可以为中低频的扬声器，也可以为全频的扬声器，第二扬声器为高频的扬声器。为了减小ANC对下行音质的损伤，高频扬声器的分频点可以在6kHz以上，即对6kHz以上的音频信号并未实现补 偿。当然，本申请实施例对该6kHz的分频点也不做限制，也可以为其他的高频分频点。
需要说明的是，上述提及的FF自适应引擎、FB自适应引擎、以及下行补偿自适应引擎可以部署在微控制器上。FF滤波器、FB滤波器、以及下行补偿滤波器可以部署在ANC芯片上。该微控制器和该ANC芯片可以统称为降噪处理器。其中，微控制器和ANC芯片可以集成在一个芯片上，也可以部署在两个芯片上。
图12是本申请实施例提供的一种降噪装置的结构示意图，该降噪装置可以由软件、硬件或者两者的结合实现成为耳机的部分或者全部，该耳机可以为图1所示的耳机。参见图12，该装置包括：降噪参数确定模块1201、反相噪声生成模块1202和降噪模块1203。
降噪参数确定模块1201，用于确定与多个第一扬声器一一对应的多组目标降噪参数；
反相噪声生成模块1202，用于基于多组目标降噪参数，生成与多个第一扬声器一一对应的多组目标反相噪声，该多组目标反相噪声中每个目标反相噪声的频段均覆盖该多个第一扬声器的发声频段；
降噪模块1203，用于利用多组目标反相噪声，通过多个第一扬声器进行降噪。
可选地，耳机还包括与多个第一扬声器一一对应的多个FF滤波器，多组目标降噪参数包括多个FF滤波器的第k帧滤波系数，k为大于等于1的整数；
降噪参数确定模块1201包括：
第一FF滤波系数确定子模块，用于在k等于1的情况下，将多个FF滤波器的初始滤波系数确定为多个FF滤波器的第k帧滤波系数，或者，基于初始降噪档位、以及降噪档位与FF滤波系数的映射关系，确定所述多个FF滤波器的第k帧滤波系数；
第二FF滤波系数确定子模块，用于在k大于1的情况下，基于至少一个参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号以及目标降噪档位，确定多个FF滤波器的第k帧滤波系数。
可选地，第二FF滤波系数确定子模块具体用于：
基于目标降噪档位、以及降噪档位与次级路径SP的滤波系数的映射关系，确定多个SP的第k-1帧滤波系数，多个SP是指多个第一扬声器到误差麦克风的路径；
基于至少一个参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号和多个SP的第k-1帧滤波系数，确定多个FF滤波器的第k帧滤波系数。
可选地，第二FF滤波系数确定子模块还具体用于：
将多个FF滤波器中的一个FF滤波器作为目标FF滤波器，按照如下操作确定目标FF滤波器的第k帧滤波系数，直至确定出每个FF滤波器的第k帧滤波系数为止：
如果目标FF滤波器是首个FF滤波器，则基于目标参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号、以及目标SP的第k-1帧滤波系数，确定目标FF滤波器的第k帧滤波系数，目标参考麦克风为目标FF滤波器对应的参考麦克风，目标SP是指目标FF滤波器对应的第一扬声器到误差麦克风的路径；
如果目标FF滤波器是非首个FF滤波器，则基于目标参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号、多个SP的第k-1帧滤波系数、以及位于目标FF滤波器之前的各个FF滤波器的第k帧频响信息和第k-1帧频响信息，确定目标FF滤波器的第k帧滤波系数。
可选地，耳机还包括与多个第一扬声器一一对应的多个FB滤波器；
第二FF滤波系数确定子模块还具体用于：
基于至少一个参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号、多个SP的第k-1帧滤波系数和多个FB滤波器的第k-1帧滤波系数，确定多个FF滤波器的第k帧滤波系数。
可选地，第二FF滤波系数确定子模块还具体用于：
将多个FF滤波器中的一个FF滤波器作为目标FF滤波器，按照如下操作确定目标FF滤波器的第k帧滤波系数，直至确定出每个FF滤波器的第k帧滤波系数为止：
如果目标FF滤波器是首个FF滤波器，则基于目标参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号、多个SP的第k-1帧滤波系数和多个FB滤波器的第k-1帧滤波系数，确定目标FF滤波器的第k帧滤波系数，目标参考麦克风为目标FF滤波器对应的参考麦克风；
如果目标FF滤波器是非首个FF滤波器，则基于目标参考麦克风采集的第k-1帧参考信号、误差麦克风采集的第k-1帧误差信号、多个SP的第k-1帧滤波系数、多个FB滤波器的第k-1帧滤波系数、以及位于目标FF滤波器之前的各个FF滤波器的第k帧频响信息和第k-1帧频响信息，确定目标FF滤波器的第k帧滤波系数。
可选地，耳机还包括与多个第一扬声器一一对应的多个反馈FB滤波器，多组目标降噪参数还包括多个FB滤波器的第k帧滤波系数，k为大于等于1的整数；
降噪参数确定模块还包括：
第一FB滤波系数确定子模块，用于在k等于1的情况下，将多个FB滤波器的初始滤波系数确定为多个FB滤波器的第k帧滤波系数，或者，基于初始降噪档位、以及降噪档位与FB滤波系数的映射关系，确定所述多个FB滤波器的第k帧滤波系数；
第二FB滤波系数确定子模块，用于在k大于1的情况下，基于目标降噪档位，确定多个FB滤波器的第k帧滤波系数。
可选地，第二FB滤波系数确定子模块具体用于：
将多个FB滤波器中的一个FB滤波器作为目标FB滤波器，按照如下操作确定目标FB滤波器的第k帧滤波系数，直至确定出每个FB滤波器的第k帧滤波系数为止：
基于目标降噪档位，以及降噪档位与FB滤波系数的映射关系，确定目标FB滤波器的第k帧滤波系数；或者，
如果目标FB滤波器属于第一类FB滤波器，则基于目标降噪档位，以及降噪档位与FB滤波系数的映射关系，确定目标FB滤波器的第k帧滤波系数，如果目标FB滤波器属于第二类FB滤波器，则基于误差麦克风采集的第k-1帧误差信号、目标FB滤波器的第k-1帧滤波系数以及目标降噪档位，确定目标FB滤波器的第k帧滤波系数。
可选地，第二FB滤波系数确定子模块还具体用于：
基于目标降噪档位，以及降噪档位与次级路径SP的滤波系数的映射关系，确定目标SP的第k-1帧滤波系数，目标SP是指目标FB滤波器对应的第一扬声器到误差麦克风的路径；
基于误差麦克风采集的第k-1帧误差信号、目标FB滤波器的第k-1帧滤波系数以及目标SP的第k-1帧滤波系数，确定目标FB滤波器的第k帧滤波系数。
可选地，第一类FB滤波器对应的第一扬声器的发声频段高于第二类FB滤波器对应的第一扬声器的发声频段。
可选地，该装置还包括：
第一降噪档位确定模块，用于确定第k-1帧降噪档位；
降噪档位获取模块，用于获取位于第k-1帧之前的m帧降噪档位，m大于等于1且小于k-1；
第二降噪档位确定模块，用于基于第k-1帧降噪档位和m帧降噪档位，确定目标降噪档位。
可选地，第一降噪档位确定模块具体用于：
在第k-1帧不存在有效下行信号且处于非安静环境的情况下，根据多个FF滤波器的参考滤波系数，以及降噪档位与FF滤波器的频响信息的映射关系，确定第k-1帧降噪档位；
其中，在k等于2的情况下，参考滤波系数为相应FF滤波器的初始滤波系数，在k大于2的情况下，参考滤波系数为相应FF滤波器在第k帧之前最近一次达到收敛稳定条件的滤波系数，或者为相应FF滤波器的第k-1帧滤波系数。
可选地，第一降噪档位确定模块具体用于：
根据多个FF滤波器的参考滤波系数，确定多个FF滤波器的参考频响信息；
基于降噪档位与FF滤波器的频响信息的映射关系，确定与多个FF滤波器的参考频响信息分别匹配的降噪档位，以得到多个参考降噪档位；
基于多个参考降噪档位，确定第k-1帧降噪档位。
可选地，第一降噪档位确定模块还具体用于：
按照多个参考降噪档位的平均值确定第k-1帧降噪档位；或者，
按照多个参考降噪档位中数量最多的参考降噪档位确定第k-1帧降噪档位。
可选地，第一降噪档位确定模块具体用于：
在第k-1帧存在有效下行信号的情况下，基于第k-1帧的有效下行信号、至少一个参考麦克风采集的第k-1帧参考信号，以及误差麦克风采集的第k-1帧误差信号，确定第k-1帧降噪档位。
可选地，每个FF滤波器的滤波系数包括至少一个双二阶滤波系数和一个增益。
综上所述，在本申请实施例中，由于该多组目标反相噪声与该多个第一扬声器一一对应，且该多组目标反相噪声中每个目标反相噪声的频段均覆盖该多个第一扬声器的发声频段，即目标反相噪声为全频段的反相噪声，所以，无论第一扬声器为高频扬声器、低频扬声器还是全频扬声器，在利用目标反相噪声进行降噪时，均能够充分发挥每个第一扬声器的降噪能力。换种方式来讲，在多个降噪通道、多个扬声器的耳机架构下，本方案能够通过多个降噪通道的全频段反相噪声提升耳机降噪效果。
需要说明的是：上述实施例提供的降噪装置在进行降噪时，仅以上述各功能模块的划分进行举例说明，实际应用中，可以根据需要而将上述功能分配由不同的功能模块完成，即将装置的内部结构划分成不同的功能模块，以完成以上描述的全部或者部分功能。另外，上述实施例提供的降噪装置与降噪方法实施例属于同一构思，其具体实现过程详见方法实施例，这里不再赘述。
请参考图13，图13是根据本申请实施例示出的另一种耳机的结构示意图。该耳机包括一个或多个处理器1301、通信总线1302、存储器1303以及一个或多个通信接口1304。
处理器1301为一个通用中央处理器(central processing unit，CPU)、网络处理器(network processing，NP)、微处理器、或者为一个或多个用于实现本申请方案的集成电路，例如，专用集成电路(application-specific integrated circuit，ASIC)，可编程逻辑器件(programmable logic device，PLD)或其组合。可选地，上述PLD为复杂可编程逻辑器件(complex programmable logic device，CPLD)，现场可编程逻辑门阵列(field-programmable gate array，FPGA)，通用阵列逻辑(generic array logic，GAL)或其任意组合。
通信总线1302用于在上述组件之间传送信息。可选地，通信总线1302分为地址总线、数据总线、控制总线等。为便于表示，图中仅用一条粗线表示，但并不表示仅有一根总线或一种类型的总线。
可选地，存储器1303为只读存储器(read-only memory，ROM)、随机存取存储器(random access memory，RAM)、电可擦可编程只读存储器(electrically erasable programmable read-only memory，EEPROM)、光盘(包括只读光盘(compact disc read-only memory，CD-ROM)、压缩光盘、激光盘、数字通用光盘、蓝光光盘等)、磁盘存储介质或者其它磁存储设备，或者是能够用于携带或存储具有指令或数据结构形式的期望的程序代码并能够由计算机存取的任何其它介质，但不限于此。存储器1303独立存在，并通过通信总线1302与处理器1301相连接，或者，存储器1303与处理器1301集成在一起。
通信接口1304使用任何收发器一类的装置，用于与其它设备或通信网络通信。通信接口1304包括有线通信接口，可选地，还包括无线通信接口。其中，有线通信接口例如以太网接口等。可选地，以太网接口为光接口、电接口或其组合。无线通信接口为无线局域网(wireless local area networks，WLAN)接口、蜂窝网络通信接口或其组合等。
在一些实施例中，存储器1303用于存储执行本申请方案的程序代码1305，处理器1301能够执行存储器1303中存储的程序代码1305。该程序代码中包括一个或多个软件模块，该耳机能够通过处理器1301以及存储器1303中的程序代码1305，来实现上文图2实施例提供的降噪方法。
在上述实施例中，可以全部或部分地通过软件、硬件、固件或者其任意结合来实现。当使用软件实现时，可以全部或部分地以计算机程序产品的形式实现。所述计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行所述计算机指令时，全部或部分地产生按照本申请实施例所述的流程或功能。所述计算机可以是通用计算机、专用计算机、计算机网络或其他可编程装置。所述计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输，例如，所述计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如：同轴电缆、光纤、数据用户线(digital subscriber line，DSL))或无线(例如：红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。所述计算机可读存储介质可以是计算机能够存取的任何可用介质，或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。所述可用介质可以是磁性介质(例如：软盘、硬盘、磁带)、光介质(例如：数字通用光盘(digital versatile disc，DVD))或半导体介质(例如：固态硬盘(solid state disk，SSD))等。值得注意的是，本申请实施例提到的计算机可读存储介质可以为非 易失性存储介质，换句话说，可以是非瞬时性存储介质。
本申请实施例还提供了一种计算机可读存储介质，所述存储介质内存储有计算机程序，所述计算机程序被处理器执行时实现上述所述的方法的步骤。
本申请实施例还提供了一种计算机程序产品，所述计算机程序产品内存储有计算机指令，所述计算机指令被处理器执行时实现上述所述的方法的步骤。
应当理解的是，本文提及的“至少一个”是指一个或多个，“多个”是指两个或两个以上。在本申请实施例的描述中，除非另有说明，“/”表示或的意思，例如，A/B可以表示A或B；本文中的“和/或”仅仅是一种描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。另外，为了便于清楚描述本申请实施例的技术方案，在本申请的实施例中，采用了“第一”、“第二”等字样对功能和作用基本相同的相同项或相似项进行区分。本领域技术人员可以理解“第一”、“第二”等字样并不对数量和执行次序进行限定，并且“第一”、“第二”等字样也并不限定一定不同。
需要说明的是，本申请实施例所涉及的信息(包括但不限于用户设备信息、用户个人信息等)、数据(包括但不限于用于分析的数据、存储的数据、展示的数据等)以及信号，均为经用户授权或者经过各方充分授权的，且相关数据的收集、使用和处理需要遵守相关国家和地区的相关法律法规和标准。
以上所述为本申请提供的实施例，并不用以限制本申请，凡在本申请的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本申请的保护范围之内。

Claims (22)

  1. 一种降噪方法，其特征在于，应用于耳机，所述耳机包括至少一个参考麦克风、一个误差麦克风和多个第一扬声器；所述方法包括：
    确定与所述多个第一扬声器一一对应的多组目标降噪参数；
    基于所述多组目标降噪参数，生成与所述多个第一扬声器一一对应的多组目标反相噪声，所述多组目标反相噪声中每个目标反相噪声的频段均覆盖所述多个第一扬声器的发声频段；
    利用所述多组目标反相噪声，通过所述多个第一扬声器进行降噪。
  2. 如权利要求1所述的方法，其特征在于，所述耳机还包括与所述多个第一扬声器一一对应的多个前馈FF滤波器，所述多组目标降噪参数包括所述多个FF滤波器的第k帧滤波系数，k为大于等于1的整数；
    所述确定与所述多个第一扬声器一一对应的多组目标降噪参数，包括：
    在k等于1的情况下，将所述多个FF滤波器的初始滤波系数确定为所述多个FF滤波器的第k帧滤波系数，或者，基于初始降噪档位、以及降噪档位与FF滤波系数的映射关系，确定所述多个FF滤波器的第k帧滤波系数；
    在k大于1的情况下，基于所述至少一个参考麦克风采集的第k-1帧参考信号、所述误差麦克风采集的第k-1帧误差信号以及目标降噪档位，确定所述多个FF滤波器的第k帧滤波系数。
  3. 如权利要求2所述的方法，其特征在于，所述基于所述至少一个参考麦克风采集的第k-1帧参考信号、所述误差麦克风采集的第k-1帧误差信号以及目标降噪档位，确定所述多个FF滤波器的第k帧滤波系数，包括：
    基于所述目标降噪档位、以及降噪档位与次级路径SP的滤波系数的映射关系，确定多个SP的第k-1帧滤波系数，所述多个SP是指所述多个第一扬声器到所述误差麦克风的路径；
    基于所述至少一个参考麦克风采集的第k-1帧参考信号、所述误差麦克风采集的第k-1帧误差信号和所述多个SP的第k-1帧滤波系数，确定所述多个FF滤波器的第k帧滤波系数。
  4. 如权利要求3所述的方法，其特征在于，所述基于所述至少一个参考麦克风采集的第k-1帧参考信号、所述误差麦克风采集的第k-1帧误差信号和所述多个SP的第k-1帧滤波系数，确定所述多个FF滤波器的第k帧滤波系数，包括：
    将所述多个FF滤波器中的一个FF滤波器作为目标FF滤波器，按照如下操作确定所述目标FF滤波器的第k帧滤波系数，直至确定出每个FF滤波器的第k帧滤波系数为止：
    如果所述目标FF滤波器是首个FF滤波器，则基于目标参考麦克风采集的第k-1帧参考信号、所述误差麦克风采集的第k-1帧误差信号、以及目标SP的第k-1帧滤波系数，确定所述目标FF滤波器的第k帧滤波系数，所述目标参考麦克风为所述目标FF滤波器对应的参考麦克风，所述目标SP是指所述目标FF滤波器对应的第一扬声器到所述误差麦克风的路径；
    如果所述目标FF滤波器是非首个FF滤波器，则基于所述目标参考麦克风采集的第k-1帧参考信号、所述误差麦克风采集的第k-1帧误差信号、所述多个SP的第k-1帧滤波系数、以及位于所述目标FF滤波器之前的各个FF滤波器的第k帧频响信息和第k-1帧频响信息，确定所述目标FF滤波器的第k帧滤波系数。
  5. 如权利要求3所述的方法，其特征在于，所述耳机还包括与所述多个第一扬声器一一对应的多个反馈FB滤波器；
    所述基于所述至少一个参考麦克风采集的第k-1帧参考信号、所述误差麦克风采集的第k-1帧误差信号和所述多个SP的第k-1帧滤波系数，确定所述多个FF滤波器的第k帧滤波系数，包括：
    基于所述至少一个参考麦克风采集的第k-1帧参考信号、所述误差麦克风采集的第k-1帧误差信号、 所述多个SP的第k-1帧滤波系数和所述多个FB滤波器的第k-1帧滤波系数，确定所述多个FF滤波器的第k帧滤波系数。
  6. 如权利要求5所述的方法，其特征在于，所述基于所述至少一个参考麦克风采集的第k-1帧参考信号、所述误差麦克风采集的第k-1帧误差信号、所述多个SP的第k-1帧滤波系数和所述多个FB滤波器的第k-1帧滤波系数，确定所述多个FF滤波器的第k帧滤波系数，包括：
    将所述多个FF滤波器中的一个FF滤波器作为目标FF滤波器，按照如下操作确定所述目标FF滤波器的第k帧滤波系数，直至确定出每个FF滤波器的第k帧滤波系数为止：
    如果所述目标FF滤波器是首个FF滤波器，则基于目标参考麦克风采集的第k-1帧参考信号、所述误差麦克风采集的第k-1帧误差信号、所述多个SP的第k-1帧滤波系数和所述多个FB滤波器的第k-1帧滤波系数，确定所述目标FF滤波器的第k帧滤波系数，所述目标参考麦克风为所述目标FF滤波器对应的参考麦克风；
    如果所述目标FF滤波器是非首个FF滤波器，则基于所述目标参考麦克风采集的第k-1帧参考信号、所述误差麦克风采集的第k-1帧误差信号、所述多个SP的第k-1帧滤波系数、所述多个FB滤波器的第k-1帧滤波系数、以及位于所述目标FF滤波器之前的各个FF滤波器的第k帧频响信息和第k-1帧频响信息，确定所述目标FF滤波器的第k帧滤波系数。
  7. 如权利要求1-6任一所述的方法，其特征在于，所述耳机还包括与所述多个第一扬声器一一对应的多个反馈FB滤波器，所述多组目标降噪参数还包括所述多个FB滤波器的第k帧滤波系数，k为大于等于1的整数；
    所述确定与所述多个第一扬声器一一对应的多组目标降噪参数，包括：
    在k等于1的情况下，将所述多个FB滤波器的初始滤波系数确定为所述多个FB滤波器的第k帧滤波系数，或者，基于初始降噪档位、以及降噪档位与FB滤波系数的映射关系，确定所述多个FB滤波器的第k帧滤波系数；
    在k大于1的情况下，基于目标降噪档位，确定所述多个FB滤波器的第k帧滤波系数。
  8. 如权利要求7所述的方法，其特征在于，所述基于目标降噪档位，确定所述多个FB滤波器的第k帧滤波系数，包括：
    将所述多个FB滤波器中的一个FB滤波器作为目标FB滤波器，按照如下操作确定所述目标FB滤波器的第k帧滤波系数，直至确定出每个FB滤波器的第k帧滤波系数为止：
    基于所述目标降噪档位，以及所述降噪档位与FB滤波系数的映射关系，确定所述目标FB滤波器的第k帧滤波系数；或者，
    如果所述目标FB滤波器属于第一类FB滤波器，则基于所述目标降噪档位，以及所述降噪档位与FB滤波系数的映射关系，确定所述目标FB滤波器的第k帧滤波系数，如果所述目标FB滤波器属于第二类FB滤波器，则基于所述误差麦克风采集的第k-1帧误差信号、所述目标FB滤波器的第k-1帧滤波系数以及所述目标降噪档位，确定所述目标FB滤波器的第k帧滤波系数。
  9. 如权利要求8所述的方法，其特征在于，所述基于所述误差麦克风采集的第k-1帧误差信号、所述目标FB滤波器的第k-1帧滤波系数以及所述目标降噪档位，确定所述目标FB滤波器的第k帧滤波系数，包括：
    基于所述目标降噪档位，以及降噪档位与次级路径SP的滤波系数的映射关系，确定目标SP的第k-1帧滤波系数，所述目标SP是指所述目标FB滤波器对应的第一扬声器到所述误差麦克风的路径；
    基于所述误差麦克风采集的第k-1帧误差信号、所述目标FB滤波器的第k-1帧滤波系数以及所述目标SP的第k-1帧滤波系数，确定所述目标FB滤波器的第k帧滤波系数。
  10. 如权利要求8或9所述的方法，其特征在于，所述第一类FB滤波器对应的第一扬声器的发声频段高于所述第二类FB滤波器对应的第一扬声器的发声频段。
  11. 如权利要求2-10任一所述的方法，其特征在于，所述方法还包括：
    确定第k-1帧降噪档位；
    获取位于第k-1帧之前的m帧降噪档位，m大于等于1且小于k-1；
    基于所述第k-1帧降噪档位和所述m帧降噪档位，确定所述目标降噪档位。
  12. 如权利要求11所述的方法，其特征在于，所述确定第k-1帧降噪档位，包括：
    在第k-1帧不存在有效下行信号且处于非安静环境的情况下，根据所述多个FF滤波器的参考滤波系数，以及降噪档位与FF滤波器的频响信息的映射关系，确定所述第k-1帧降噪档位；
    其中，在k等于2的情况下，所述参考滤波系数为相应FF滤波器的初始滤波系数，在k大于2的情况下，所述参考滤波系数为相应FF滤波器在第k帧之前最近一次达到收敛稳定条件的滤波系数，或者为相应FF滤波器的第k-1帧滤波系数。
  13. 如权利要求12所述的方法，其特征在于，所述根据所述多个FF滤波器的参考滤波系数，以及降噪档位与FF滤波器的频响信息的映射关系，确定所述第k-1帧降噪档位，包括：
    根据所述多个FF滤波器的参考滤波系数，确定所述多个FF滤波器的参考频响信息；
    基于所述降噪档位与FF滤波器的频响信息的映射关系，确定与所述多个FF滤波器的参考频响信息分别匹配的降噪档位，以得到多个参考降噪档位；
    基于所述多个参考降噪档位，确定所述第k-1帧降噪档位。
  14. 如权利要求13所述的方法，其特征在于，所述基于所述多个参考降噪档位，确定所述第k-1帧降噪档位，包括：
    按照所述多个参考降噪档位的平均值确定所述第k-1帧降噪档位；或者，
    按照所述多个参考降噪档位中数量最多的参考降噪档位确定所述第k-1帧降噪档位。
  15. 如权利要求11所述的方法，其特征在于，所述确定第k-1帧降噪档位，包括：
    在第k-1帧存在有效下行信号的情况下，基于所述第k-1帧的有效下行信号、所述至少一个参考麦克风采集的第k-1帧参考信号，以及所述误差麦克风采集的第k-1帧误差信号，确定所述第k-1帧降噪档位。
  16. 如权利要求2-15任一所述的方法，其特征在于，每个FF滤波器的滤波系数包括至少一个双二阶滤波系数和一个增益。
  17. 一种耳机，其特征在于，所述耳机包括至少一个参考麦克风、一个误差麦克风、多个第一扬声器和一个降噪处理器；
    所述降噪处理器用于实现权利要求1-16任一所述方法的步骤。
  18. 如权利要求17所述的耳机，其特征在于，所述多个第一扬声器包括由一个双振膜喇叭形成的两个第一扬声器；或者，所述多个第一扬声器包括分体喇叭的多个扬声器。
  19. 如权利要求17或18所述的耳机，其特征在于，所述耳机还包括至少一个第二扬声器，所述至少一个第二扬声器不参与降噪。
  20. 一种降噪装置，其特征在于，应用于耳机，所述耳机包括至少一个参考麦克风、一个误差麦克风和多个第一扬声器；所述装置包括：
    降噪参数确定模块，用于确定与所述多个第一扬声器一一对应的多组目标降噪参数；
    反相噪声生成模块，用于基于所述多组目标降噪参数，生成与所述多个第一扬声器一一对应的多组目标反相噪声，所述多组目标反相噪声中每个目标反相噪声的频段均覆盖所述多个第一扬声器的发声频段；
    降噪模块，用于利用所述多组目标反相噪声，通过所述多个第一扬声器进行降噪。
  21. 一种计算机可读存储介质，其特征在于，所述存储介质内存储有计算机程序，所述计算机程序被处理器执行时实现权利要求1-16任一所述的方法的步骤。
  22. 一种计算机程序产品，其特征在于，所述计算机程序产品内存储有计算机指令，所述计算机指令被处理器执行时实现权利要求1-16任一所述的方法的步骤。
PCT/CN2023/103264 2022-11-28 2023-06-28 降噪方法、耳机、装置、存储介质及计算机程序产品 WO2024113811A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211506453.2 2022-11-28
CN202211506453.2A CN118102159A (zh) 2022-11-28 2022-11-28 降噪方法、耳机、装置、存储介质及计算机程序产品

Publications (1)

Publication Number Publication Date
WO2024113811A1 true WO2024113811A1 (zh) 2024-06-06

Family

ID=91159013

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/103264 WO2024113811A1 (zh) 2022-11-28 2023-06-28 降噪方法、耳机、装置、存储介质及计算机程序产品

Country Status (2)

Country Link
CN (1) CN118102159A (zh)
WO (1) WO2024113811A1 (zh)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107005758A (zh) * 2014-12-15 2017-08-01 霍尼韦尔国际公司 具有扬声器阵列的主动降噪耳杯
CN113676803A (zh) * 2020-05-14 2021-11-19 华为技术有限公司 一种主动降噪方法及装置
CN113707121A (zh) * 2021-08-02 2021-11-26 杭州萤石软件有限公司 主动降噪系统、方法及装置
CN114040287A (zh) * 2021-11-05 2022-02-11 恒玄科技(上海)股份有限公司 对耳机主动降噪的方法、主动降噪系统以及耳机
CN115278438A (zh) * 2022-07-27 2022-11-01 北京爱德发科技有限公司 降噪耳机、降噪方法和装置、存储介质及处理器

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107005758A (zh) * 2014-12-15 2017-08-01 霍尼韦尔国际公司 具有扬声器阵列的主动降噪耳杯
CN113676803A (zh) * 2020-05-14 2021-11-19 华为技术有限公司 一种主动降噪方法及装置
CN113707121A (zh) * 2021-08-02 2021-11-26 杭州萤石软件有限公司 主动降噪系统、方法及装置
CN114040287A (zh) * 2021-11-05 2022-02-11 恒玄科技(上海)股份有限公司 对耳机主动降噪的方法、主动降噪系统以及耳机
CN115278438A (zh) * 2022-07-27 2022-11-01 北京爱德发科技有限公司 降噪耳机、降噪方法和装置、存储介质及处理器

Also Published As

Publication number Publication date
CN118102159A (zh) 2024-05-28

Similar Documents

Publication Publication Date Title
US9577596B2 (en) System and method for personalization of an audio equalizer
US10475434B2 (en) Electronic device and control method of earphone device
CN105307081A (zh) 一种主动降噪的语音信号处理系统及方法
JP6301508B2 (ja) 通信ヘッドセットにおける自己音声フィードバック
EP4429267A1 (en) Earphone having active noise reduction function and active noise reduction method
WO2021238458A1 (zh) 扬声设备音质的优化方法
US10097929B2 (en) Sound signal amplitude suppressing apparatus
US11842717B2 (en) Robust open-ear ambient sound control with leakage detection
CN104885360A (zh) 具有声压级和频率依赖性增益的助听器
CN211152190U (zh) 一种半入耳式主动降噪耳机
CN112804621B (zh) 音频均衡方法、耳机电路和耳机
Puder Hearing aids: an overview of the state-of-the-art, challenges, and future trends of an interesting audio signal processing application
CN109814833A (zh) 一种实时控制频响输出装置及其应用方法
WO2024113811A1 (zh) 降噪方法、耳机、装置、存储介质及计算机程序产品
TWI501657B (zh) 電子音訊裝置
WO2024113810A1 (zh) 降噪方法、耳机、装置、存储介质及计算机程序产品
US20230060353A1 (en) Context aware compressor for headphone audio feedback path
US20220141583A1 (en) Hearing assisting device and method for adjusting output sound thereof
WO2021129196A1 (zh) 一种语音信号处理方法及装置
US20240005903A1 (en) Headphone Speech Listening Based on Ambient Noise
JPH05145991A (ja) 低音域特性補正回路
TWI697238B (zh) 耳機
CN209231915U (zh) 一种实时控制频响输出装置
US20240323586A1 (en) Earphone and audio processing method and apparatus therefor, and storage medium
WO2023122407A1 (en) Communication device, hearing aid system and computer readable medium

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23895975

Country of ref document: EP

Kind code of ref document: A1