Active noise reduction method, system, device and storage medium for audio

Abstract

The application relates to the field of active noise reduction, and discloses an active noise reduction method, system, device and storage medium for audio. The method comprises the following steps: a microphone receiving system receives external noise, generates an audio input value, and generates a noise estimation value based on an input parameter; a loudspeaker playing system superimposes an audio playing value and an active noise reduction playing value to obtain a loudspeaker playing value, and performs audio playing processing on the loudspeaker playing value to generate a playing audio; the microphone receiving system receives feedback audio of the playing audio through a feedback path to obtain a feedback audio value, generates a residual audio value according to a feedback path estimation value and a residual algorithm; the loudspeaker playing system obtains a new feedback path estimation value by using convolution of the feedback audio value and the audio playing value based on an LMS algorithm; and the loudspeaker playing system obtains a new noise reduction parameter by using convolution of the new feedback path estimation value and the residual audio value according to a preset fixed step parameter.

Description

Technical Field

[0001] This invention relates to the field of active noise cancellation, and more particularly to an active noise cancellation method, system, device, and storage medium for audio. Background Technology

[0002] With the development of technology, voice devices have become increasingly common in people's daily lives. When people use voice devices to listen to audio information such as voice and music, external environmental noise can also pass through the voice device and be heard by the ears. Especially when the external environmental noise is complex, it can seriously affect people's listening experience. Therefore, noise reduction processing is needed to reduce the impact of external noise on voice devices. In addition to reducing noise interference through the physical structure of the device, voice devices can also use active noise reduction to reduce the impact of noise information.

[0003] However, existing noise reduction methods involve changes in both the audio itself and external sounds. Traditional active noise cancellation takes a long time to correct to a suitable noise reduction range, and the noise reduction effect is not good. Most existing noise reduction algorithms use fixed filters (non-adaptive), while some are adaptive active noise reduction methods (offline adaptive, or online adaptive using white noise). However, the performance of fixed noise reduction filters is limited, and adaptive methods ignore the situation where the user needs to simultaneously enable noise cancellation and listen to music. A new technology is needed to solve the technical problems of current noise reduction speed being too slow and ignoring the need for users to simultaneously enable noise cancellation and listen to music. Summary of the Invention

[0004] The main objective of this invention is to address the technical problem that current noise cancellation is too slow and ignores the need for users to simultaneously enable noise cancellation and listen to music.

[0005] The first aspect of this invention provides an active noise reduction method for audio, wherein the active noise reduction method is applied to an active noise reduction system for audio, the active noise reduction system for audio comprising: a microphone receiving system and a speaker playback system, and the active noise reduction method for audio comprising:

[0006] The microphone receiving system receives external noise, generates audio input values, and generates noise estimates based on preset input parameters, wherein the input parameters are the attenuation ratio of the external noise being recorded.

[0007] The speaker playback system convolves the preset noise reduction parameters with the noise estimate to obtain the active noise reduction playback value, and generates the audio playback value based on the preset audio file;

[0008] The audio playback value and the active noise reduction playback value are superimposed to obtain the speaker playback value, and the speaker playback value is subjected to audio playback processing to generate playback audio.

[0009] The microphone receiving system receives the feedback audio of the played audio through the feedback path, obtains the feedback audio value, and performs residual calculation processing on the feedback audio value according to the preset feedback path estimate and residual algorithm to generate residual audio value. The feedback path is the acoustic path from the speaker playback system to the microphone receiving system, and the feedback path estimate is the attenuation coefficient of the feedback path for sound propagation.

[0010] The speaker playback system is based on a preset LMS algorithm. It uses the convolution of the feedback audio value and the audio playback value to iteratively increase or decrease the feedback path estimate to obtain a new feedback path estimate.

[0011] Based on the preset fixed step size parameter, the noise reduction parameter is increased or decreased by a multiple step size using the convolution of the new feedback path estimate and the residual audio value to obtain a new noise reduction parameter. The noise reduction parameter is used to adjust the active noise reduction playback value in order to achieve active noise reduction.

[0012] Optionally, in a first implementation of the first aspect of the present invention, the step of increasing or decreasing the noise reduction parameters by a multiple of the step size using the convolution of the new feedback path estimate and the residual audio value according to a preset fixed step size parameter to obtain new noise reduction parameters includes:

[0013] Set the preset fixed step size parameter to be less than the reciprocal of the current audio playback value;

[0014] The fixed step size parameter, the new feedback path estimate, and the residual audio value are convolved to obtain the dynamic step size value.

[0015] Based on the dynamic step size value, the noise reduction parameters are increased or decreased to obtain new noise reduction parameters.

[0016] Optionally, in a second implementation of the first aspect of the present invention, the step of increasing or decreasing the noise reduction parameters according to the dynamic step size value to obtain new noise reduction parameters includes:

[0017] Based on the time-domain change, determine whether the absolute value of the residual audio value decreases;

[0018] If the absolute value decreases, the noise reduction parameter is reduced by the dynamic step size value to obtain a new noise reduction parameter;

[0019] If the absolute value does not decrease, the noise reduction parameter is increased by the dynamic step size value to obtain a new noise reduction parameter.

[0020] Optionally, in a third implementation of the first aspect of the present invention, the step of convolving the fixed step size parameter, the new feedback path estimate, and the residual audio value to obtain the dynamic step size value includes:

[0021] The fixed step size parameter, the new feedback path estimate, and the residual audio value are double-convolved to obtain the dynamic step size value.

[0022] Optionally, in a fourth implementation of the first aspect of the present invention, the step of performing residual calculation processing on the feedback audio value based on a preset feedback path estimate and a residual algorithm to generate residual audio values ​​includes:

[0023] The estimated feedback path value is convolved with the speaker playback value to generate a prediction residual value;

[0024] The residual audio value is generated by subtracting the predicted residual value from the feedback audio value.

[0025] Optionally, in a fifth implementation of the first aspect of the present invention, the step of iteratively increasing or decreasing the feedback path estimate based on a preset LMS algorithm using the convolution of the feedback audio value and the audio playback value to obtain a new feedback path estimate includes:

[0026] Calculate the convolution of the feedback audio value and the audio playback value to obtain the iteration value;

[0027] Based on the time-domain change, determine whether the absolute value of the residual audio value decreases;

[0028] If the absolute value does not decrease, the feedback path estimate is increased by the iteration value to obtain a new feedback path estimate.

[0029] If the absolute value decreases, the feedback path estimate is reduced by the iteration value to obtain a new feedback path estimate.

[0030] Optionally, in a sixth implementation of the first aspect of the present invention, the step of superimposing the audio playback value and the active noise reduction playback value to obtain the speaker playback value includes:

[0031] The active noise reduction playback value is increased by adding the frame-by-frame audio energy of the audio playback value to obtain the speaker playback value.

[0032] A second aspect of the present invention provides an active noise reduction system for audio, the active noise reduction system for audio comprising:

[0033] Microphone receiving system, speaker playback system;

[0034] The microphone receiving system is used to receive external noise, generate audio input values, and generate noise estimates based on preset input parameters, wherein the input parameters are the attenuation ratio of the external noise being recorded.

[0035] The speaker playback system is used to convolve preset noise reduction parameters with the noise estimate to obtain active noise reduction playback values, and to generate audio playback values ​​based on preset audio files;

[0036] The audio playback value and the active noise reduction playback value are superimposed to obtain the speaker playback value, and the speaker playback value is subjected to audio playback processing to generate playback audio.

[0037] The microphone receiving system is used to receive the feedback audio of the played audio through the feedback path, obtain the feedback audio value, and perform residual calculation processing on the feedback audio value according to the preset feedback path estimate and residual algorithm to generate residual audio value. The feedback path is the acoustic path from the speaker playback system to the microphone receiving system, and the feedback path estimate is the attenuation coefficient of the feedback path for sound propagation.

[0038] The speaker playback system is used to perform iterative increment and decrement processing on the feedback path estimate based on a preset LMS algorithm, using the convolution of the feedback audio value and the audio playback value, to obtain a new feedback path estimate.

[0039] Based on the preset fixed step size parameter, the noise reduction parameter is increased or decreased by a multiple step size using the convolution of the new feedback path estimate and the residual audio value to obtain a new noise reduction parameter. The noise reduction parameter is used to adjust the active noise reduction playback value in order to achieve active noise reduction.

[0040] A third aspect of the present invention provides an active noise reduction device for audio, comprising: a memory and at least one processor, wherein the memory stores instructions, and the memory and the at least one processor are interconnected via a circuit; the at least one processor invokes the instructions in the memory to cause the active noise reduction device for audio to perform the above-described active noise reduction method for audio.

[0041] A fourth aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described active noise reduction method for audio.

[0042] In this embodiment of the invention, the LMS algorithm is used to calculate the weighting coefficient W of external noise information during the sound propagation process. Based on the weighting coefficient W, the amplitude information of the corresponding noise information after change is obtained according to the optimal weighting coefficient. The speaker plays an electrical signal with the same amplitude but opposite phase to cancel the noise information, thereby improving the audio noise reduction speed and more accurately locating the opposite electrical signal. Compared with active noise reduction technology using a fixed filter (non-adaptive), this invention can adaptively identify secondary paths and improve the noise reduction width. Compared with adaptive active noise reduction methods, it takes into account the need for users to compensate the noise reduction system when listening to music, and has both offline adaptive stability and online adaptive performance. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of the overall noise reduction method for active audio noise reduction in an embodiment of the present invention;

[0044] Figure 2 This is a schematic diagram of an embodiment of the active audio noise reduction method of the present invention.

[0045] Figure 3 This is a schematic diagram of the offline adaptive secondary path estimation method in an embodiment of the present invention;

[0046] Figure 4 This is a schematic diagram of an embodiment of the active audio noise reduction system of the present invention;

[0047] Figure 5 This is a schematic diagram of one embodiment of the active noise reduction device for audio in this invention. Detailed Implementation

[0048] This invention provides an active noise reduction method, system, device, and storage medium for audio.

[0049] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” or “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0050] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 The diagram below illustrates the overall noise reduction process of the active audio noise reduction method in this invention. Please refer to [link / reference]. Figure 2 This is a schematic diagram of an embodiment of the active noise reduction method for audio in this invention.

[0051] Active audio noise reduction methods are applied to active audio noise reduction systems, which include: microphone receiving systems and speaker playback systems. The active audio noise reduction methods include:

[0052] 101. The microphone receiving system receives external noise, generates audio input values, and generates noise estimates based on preset input parameters. The input parameters represent the attenuation ratio of the external noise. In this embodiment, the microphone receiving system receives external noise (the external reference microphone receives the external noise), and the external reference microphone receives the external noise. The circuit generates noise estimates based on the external noise. P is the attenuation ratio of the sound formed by the external noise passing through the earcups. x(n) (which can be an array, i.e., the total of multiple microphones) is the external noise. However, the sound received by the microphone is d(n) = x(n) * P because it has been attenuated by the earcups. Earcup attenuation is measurable in the laboratory and is a fixed parameter. After obtaining the audio input value d(n) (which can be an array, i.e., the total of multiple microphones), the estimated noise value x(n) is calculated in reverse. The estimated noise value x(n) is then considered the external noise intensity. Using preset Gaussian white noise, this noise is played. After the error microphone receives the signal, offline adaptive calculation begins to obtain the secondary path estimate. A preset Gaussian white noise (vn) is played through a speaker and received by an internal error microphone (y'). Inside the circuit, the Gaussian white noise (vn) is convolved with the secondary path estimate to generate a Gaussian white noise estimate (vn'). The difference between this estimate and the microphone-received (y') is used to obtain the error signal (en = y' - vn'). Using the Gaussian white noise (vn) and the error signal (en), the secondary path estimate is iteratively updated using the LMS algorithm: S_HAT(n) = S_HAT(n-1) + μ*e(n)*v(n); μ is a fixed coefficient: 0.0001; μ*e(n)*v(n) is the convolution.

[0053] 102. The speaker playback system convolves the preset noise reduction parameters with the noise estimate to obtain the active noise reduction playback value, and generates the audio playback value based on the preset audio file;

[0054] In this embodiment, the speaker playback system superimposes the inverse noise from the active noise cancellation filter onto a preset Gaussian white noise or audio signal, and convolves it with the secondary path to obtain the audio signal before playback. The noise reduction parameter W is a virtual filter for active noise cancellation, which can be a software-controlled parameter that increases or decreases the volume of the white noise. Convolving the noise reduction parameter W (which can be an array) with the estimated x(n) (which can also be an array) yields the active noise cancellation playback value y (which can also be an array), where the active noise cancellation playback value y = x(n) * W is the calculation process. The audio playback value mc is obtained based on the information recorded in the audio file.

[0055] Please refer to Figure 3 , Figure 3 In the offline adaptive secondary path estimation method of this invention, during audio playback without setting an audio playback value (mc), an external reference microphone receives external noise (xn). The external noise is convolved with an active noise reduction filter and then inverted (y = -xn*W). The speaker plays reverse noise (y'). After the reverse noise (y') cancels out the external noise (dn), it enters the error microphone as an error signal. The external noise received by the external microphone is convolved with the secondary path estimation value in step 101 to generate a noise estimation value (xn'). P is the main path, which is the attenuation ratio of the sound formed by the external noise passing through the earcups. x(n) is the external noise, which is also the noise received by the microphone. d(n) = x(n)*P is equivalent to the noise remaining after the external noise is isolated by the earcups, which is mostly low frequency.

[0056] 103. The audio playback value and the active noise reduction playback value are superimposed to obtain the speaker playback value, and the speaker playback value is processed to generate the playback audio.

[0057] In this embodiment, the inverse noise is superimposed on the audio signal or Gaussian white noise and played to obtain the played audio. The audio playback value mc and the active noise cancellation playback value y are added together to obtain the speaker playback value y+mc. S is the attenuation parameter of the sound change from the speaker to the microphone, which is the actual secondary path. The sound propagated through the secondary path, y', and the external noise received by the microphone, d(n), are superimposed in the microphone.

[0058] Furthermore, in step 103, the following operations can be performed:

[0059] 1031. The frame-by-frame audio energy of the audio playback value is increased by the active noise reduction playback value to obtain the speaker playback value.

[0060] In step 1031, the active noise reduction playback value y is the speaker playback value y+mc generated by superimposing the energy of each frame of the audio playback value mc.

[0061] 104. The microphone receiving system receives the playback audio through the feedback path and obtains the feedback audio value. Based on the preset feedback path estimate and the residual algorithm, the feedback audio value is processed by residual calculation to generate the residual audio value. Here, the feedback path is the acoustic path from the speaker playback system to the microphone receiving system, and the feedback path estimate is the attenuation coefficient of the feedback path for sound propagation.

[0062] In this embodiment, the microphone receiving system (internally an error microphone; in offline adaptive mode, it receives the Gaussian active noise cancellation playback value; in active noise cancellation mode, it receives the error audio after canceling out external noise) receives the playback audio. Internally, the circuit convolves the audio signal or preset Gaussian white noise with the secondary path estimate, subtracts it from the error audio, and generates a residual audio value. The secondary path is the acoustic path from the speaker playback system to the microphone receiving system, and the secondary path estimate is the attenuation coefficient of the secondary path for sound propagation. The sound propagating through the secondary path, y', and the external noise received by the microphone, d(n), are superimposed in the microphone to generate a feedback audio value e(n) (which can be an array), where e(n) = d(n) - y' = x(n) * P - (x(n) * W + mc) * S. The feedback path is the acoustic path from the speaker playback system to the microphone receiving system, and the feedback path estimate S-HAT (which can be an array) is the attenuation coefficient of the feedback path for sound propagation. The convolution mc` between the feedback path estimate S-HAT and the audio playback value mc is mc*S-HAT, and the residual audio value is e(n)-(mc*S-HAT).

[0063] Furthermore, in step 104, "based on the preset feedback path estimate and residual algorithm, perform residual calculation processing on the feedback audio value to generate residual audio value," the following steps can be performed:

[0064] 1041. Convolve the feedback path estimate with the speaker playback value to generate the prediction residual value;

[0065] 1042. Subtract the predicted residual value from the feedback audio value to generate the residual audio value.

[0066] In steps 1041-1042, the feedback path estimate S-HAT is adjusted to approximate the acoustic path S, thus completing the precise noise reduction process. First, the feedback path estimate S-HAT is convolved with the speaker playback value mc to obtain the predicted residual value S-HAT*mc. Then, the feedback audio value e(n) is subtracted from the predicted residual value S-HAT*mc, resulting in the residual audio value e(n) - (mc*S-HAT).

[0067] 105. The speaker playback system is based on the preset LMS algorithm. It uses the convolution of the feedback audio value and the audio playback value to iteratively increase or decrease the feedback path estimate to obtain a new feedback path estimate.

[0068] In this embodiment, the speaker playback system is based on the offline adaptive LMS algorithm. It uses the convolution of the previously estimated secondary path estimate and the audio signal to iteratively increment and decrement the secondary path estimate, obtaining a better secondary path estimate. The LMS algorithm iterates the feedback path estimate S-HAT based on the convolution between the feedback audio value e(n)-(mc*S-HAT) and the audio playback value mc. That is, (e(n)-(mc*S-HAT))*mc is used as the increment / decrement value to obtain a new feedback path estimate new-S-HAT. The new feedback path estimate is obtained by either increasing (e(n)-(mc*S-HAT))*mc or decreasing (e(n)-(mc*S-HAT))*mc from the feedback path estimate S-HAT.

[0069] Furthermore, the following steps can be performed in step 105:

[0070] 1051. Calculate the convolution of the feedback audio value and the audio playback value to obtain the iterative value;

[0071] 1052. Based on the time-domain change, determine whether the absolute value of the residual audio value decreases;

[0072] 1053. If the absolute value does not decrease, increase the feedback path estimate by the iteration value to obtain a new feedback path estimate;

[0073] 1054. If the absolute value decreases, the feedback path estimate is reduced by the iteration value to obtain a new feedback path estimate.

[0074] In steps 1051-1054, the iteration value is (e(n)-(mc*S-HAT))*mc, which is the convolution between the feedback audio value e(n)-(mc*S-HAT) and the audio playback value mc. We determine if the absolute value of e(n)-(mc*S-HAT) decreases. If the absolute value does not decrease, it means the feedback path estimate S-HAT is too small, and the feedback path estimate is increased by one iteration value (e(n)-(mc*S-HAT))*mc. If the absolute value decreases, it means the feedback path estimate S-HAT is too large, and the feedback path estimate is decreased by one iteration value (e(n)-(mc*S-HAT))*mc.

[0075] 106. Based on the preset fixed step size parameters, the noise reduction parameters are increased or decreased by multiple steps using the convolution of the new feedback path estimate and the residual audio value to obtain new noise reduction parameters. The noise reduction parameters are used to adjust the active noise reduction playback value in order to achieve active noise reduction.

[0076] In this embodiment, based on a preset fixed step size parameter, the noise estimate obtained by convolving the external noise with the secondary path estimate, and the error audio, are used to perform LMS calculations on the active noise reduction filter, continuously optimizing and updating it. The fixed step size parameter u is a parameter that is adjusted according to the magnitude of the audio playback value. The fixed step size parameter u, the new feedback path estimate new-S-HAT, and the residual audio value e(n)-(mc*S-HAT) are convolved to obtain the adjusted value, which is used to adjust the noise reduction parameter W. The adjustment amount at one time is u*new-S-HAT*(e(n)-(mc*S-HAT)). As the noise reduction parameter changes, the active noise reduction playback value also changes, dynamically updating the active noise reduction playback value to achieve the cancellation of external noise x(n).

[0077] Furthermore, the following steps can be performed in step 106:

[0078] 1061. Set the preset fixed step size parameter to be less than the reciprocal of the current audio playback value;

[0079] 1062. Convolve the fixed step size parameter, the new feedback path estimate, and the residual audio value to obtain the dynamic step size value;

[0080] 1063. Based on the dynamic step size value, the noise reduction parameters are increased or decreased to obtain new noise reduction parameters.

[0081] In steps 1061-1063, the fixed step size parameter u is set to be less than the reciprocal of the audio playback value 1 / mc. The fixed step size parameter u, the new feedback path estimate new-S-HAT, and the residual audio value e(n)-(mc*S-HAT) are convolved to obtain the dynamic step size value (e(n)-(mc*S-HAT))*new-S-HAT*u. Then, the dynamic step size value (e(n)-(mc*S-HAT))*new-S-HAT*u is used as a value to gradually increase and decrease the noise reduction parameter W.

[0082] Furthermore, the following steps can be performed in step 1063:

[0083] 10631. Based on the time-domain change, determine whether the absolute value of the residual audio value decreases;

[0084] 10632. If the absolute value decreases, the noise reduction parameter is reduced by the dynamic step size value to obtain a new noise reduction parameter;

[0085] 10633. If the absolute value does not decrease, increase the dynamic step size value of the noise reduction parameter to obtain a new noise reduction parameter.

[0086] In steps 10631-10633, the time domain change changes with time. For example, every 0.1 seconds, it is determined whether the absolute value of the residual audio value e(n)-(mc*S-HAT) decreases. If the absolute value decreases, it means that the noise reduction parameter W is too large and needs to be reduced. The reduction value is a dynamic step size value (e(n)-(mc*S-HAT))*new-S-HAT*u.

[0087] If the absolute value does not decrease, it means that the noise reduction parameter W is too small, resulting in an excessively large residual. The noise reduction parameter W is increased by adding a dynamic step size value (e(n)-(mc*S-HAT))*new-S-HAT*u.

[0088] Furthermore, the following steps can be performed in step 1062:

[0089] 10621. Double-convolve the fixed step size parameter, the new feedback path estimate, and the residual audio value to obtain the dynamic step size value.

[0090] In step 10621, the dynamic step size value is adjusted to 2*(e(n)-(mc*S-HAT))*new-S-HAT*u. This is because the noise fluctuation function of the target is a quadratic function during the modification process, and the coefficient with a prefix of 2 will appear after the derivative operation. In order to make more precise adjustments, the noise reduction effect is faster at a step size of twice the convolution.

[0091] In this embodiment of the invention, the LMS algorithm is used to calculate the weighting coefficient W of external noise information during the sound propagation process. Based on the weighting coefficient W, the amplitude information of the corresponding noise information after change is obtained according to the optimal weighting coefficient. The speaker plays an electrical signal with the same amplitude but opposite phase to cancel the noise information, thereby improving the audio noise reduction speed and more accurately locating the opposite electrical signal. Compared with active noise reduction technology using a fixed filter (non-adaptive), this invention can adaptively identify secondary paths and improve the noise reduction width. Compared with adaptive active noise reduction methods, it takes into account the need for users to compensate the noise reduction system when listening to music, and has both offline adaptive stability and online adaptive performance.

[0092] The above describes the active audio noise reduction method in the embodiments of the present invention. The following describes the active audio noise reduction system in the embodiments of the present invention. Please refer to [link / reference]. Figure 4 One embodiment of the active audio noise reduction system in this invention includes:

[0093] Microphone receiving system 301, speaker playback system 302;

[0094] The microphone receiving system 301 is used to receive external noise, generate audio input values, and generate noise estimation values ​​based on preset input parameters, wherein the input parameters are the attenuation ratio of the external noise being recorded.

[0095] The speaker playback system 302 is used to convolve preset noise reduction parameters with the noise estimate to obtain active noise reduction playback values, and to generate audio playback values ​​based on preset audio files;

[0096] The audio playback value and the active noise reduction playback value are superimposed to obtain the speaker playback value, and the speaker playback value is subjected to audio playback processing to generate playback audio.

[0097] The microphone receiving system 301 is used to receive the feedback audio of the played audio through the feedback path, obtain the feedback audio value, and perform residual calculation processing on the feedback audio value according to the preset feedback path estimation value and residual algorithm to generate residual audio value. The feedback path is the acoustic path from the speaker playback system to the microphone receiving system, and the feedback path estimation value is the attenuation coefficient of the feedback path for sound propagation.

[0098] The speaker playback system 302 is used to perform iterative increment and decrement processing on the feedback path estimate based on a preset LMS algorithm, using the convolution of the feedback audio value and the audio playback value, to obtain a new feedback path estimate.

[0099] Based on the preset fixed step size parameter, the noise reduction parameter is increased or decreased by a multiple step size using the convolution of the new feedback path estimate and the residual audio value to obtain a new noise reduction parameter. The noise reduction parameter is used to adjust the active noise reduction playback value in order to achieve active noise reduction.

[0100] Specifically, the speaker playback system 302 is used for:

[0101] Set the preset fixed step size parameter to be less than the reciprocal of the current audio playback value;

[0102] The fixed step size parameter, the new feedback path estimate, and the residual audio value are convolved to obtain the dynamic step size value.

[0103] Based on the dynamic step size value, the noise reduction parameters are increased or decreased to obtain new noise reduction parameters.

[0104] The speaker playback system 302 can also be specifically used for:

[0105] Based on the time-domain change, determine whether the absolute value of the residual audio value decreases;

[0106] If the absolute value decreases, the noise reduction parameter is reduced by the dynamic step size value to obtain a new noise reduction parameter;

[0107] If the absolute value does not decrease, the noise reduction parameter is increased by the dynamic step size value to obtain a new noise reduction parameter.

[0108] The speaker playback system 302 can also be specifically used for:

[0109] The fixed step size parameter, the new feedback path estimate, and the residual audio value are double-convolved to obtain the dynamic step size value.

[0110] Specifically, the microphone receiving system 301 is used for:

[0111] The estimated feedback path value is convolved with the speaker playback value to generate a prediction residual value;

[0112] The residual audio value is generated by subtracting the predicted residual value from the feedback audio value.

[0113] The speaker playback system 302 can also be specifically used for:

[0114] Calculate the convolution of the feedback audio value and the audio playback value to obtain the iteration value;

[0115] Based on the time-domain change, determine whether the absolute value of the residual audio value decreases;

[0116] If the absolute value does not decrease, the feedback path estimate is increased by the iteration value to obtain a new feedback path estimate.

[0117] If the absolute value decreases, the feedback path estimate is reduced by the iteration value to obtain a new feedback path estimate.

[0118] The speaker playback system 302 can also be specifically used for:

[0119] The active noise reduction playback value is increased by adding the frame-by-frame audio energy of the audio playback value to obtain the speaker playback value.

[0120] In this embodiment of the invention, the LMS algorithm is used to calculate the weighting coefficient W of external noise information during the sound propagation process. Based on the weighting coefficient W, the amplitude information of the corresponding noise information after change is obtained according to the optimal weighting coefficient. The speaker plays an electrical signal with the same amplitude but opposite phase to cancel the noise information, thereby improving the audio noise reduction speed and more accurately locating the opposite electrical signal. Compared with active noise reduction technology using a fixed filter (non-adaptive), this invention can adaptively identify secondary paths and improve the noise reduction width. Compared with adaptive active noise reduction methods, it takes into account the need for users to compensate the noise reduction system when listening to music, and has both offline adaptive stability and online adaptive performance.

[0121] above Figure 4 The active audio noise reduction system in this embodiment of the invention is described in detail from the perspective of modular functional entities. The active audio noise reduction device in this embodiment of the invention is described in detail from the perspective of hardware processing.

[0122] Figure 5 This is a schematic diagram of the structure of an active audio noise reduction device 400 provided in an embodiment of the present invention. The active audio noise reduction device 400 can vary significantly due to different configurations or performance characteristics. It may include one or more central processing units (CPUs) 410 (e.g., one or more processors) and a memory 420, and one or more storage media 430 (e.g., one or more mass storage devices) storing application programs 433 or data 432. The memory 420 and storage media 430 can be temporary or persistent storage. The program stored in the storage media 430 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the active audio noise reduction device 400. Furthermore, the processor 410 may be configured to communicate with the storage media 430 and execute the series of instruction operations in the storage media 430 on the active audio noise reduction device 400.

[0123] The audio-based active noise cancellation device 400 may also include one or more power supplies 440, one or more wired or wireless network interfaces 450, one or more input / output interfaces 460, and / or one or more operating systems 431, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 5 The illustrated active noise cancellation device structure does not constitute a limitation on audio-based active noise cancellation devices, which may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0124] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the active noise reduction method for the audio.

[0125] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the system or system / unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0126] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0127] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An active noise reduction method for audio, characterized in that, The active noise reduction method for audio is applied to the active noise reduction system for audio, which includes a microphone receiving system and a speaker playback system. The active noise reduction method for audio includes: The microphone receiving system receives external noise, generates audio input values, and generates noise estimates based on preset input parameters, wherein the input parameters are the attenuation ratio of the external noise being recorded. The speaker playback system convolves the preset noise reduction parameters with the noise estimate to obtain the active noise reduction playback value, and generates the audio playback value based on the preset audio file; The audio playback value and the active noise reduction playback value are superimposed to obtain the speaker playback value, and the speaker playback value is subjected to audio playback processing to generate playback audio. The microphone receiving system receives the feedback audio of the played audio through the feedback path, obtains the feedback audio value, and performs residual calculation processing on the feedback audio value according to the preset feedback path estimate and residual algorithm to generate residual audio value. The feedback path is the acoustic path from the speaker playback system to the microphone receiving system, and the feedback path estimate is the attenuation coefficient of the feedback path for sound propagation. The speaker playback system is based on a preset LMS algorithm. It uses the convolution of the feedback audio value and the audio playback value to iteratively increase or decrease the feedback path estimate to obtain a new feedback path estimate. Based on the preset fixed step size parameter, the noise reduction parameter is increased or decreased by a multiple step size using the convolution of the new feedback path estimate and the residual audio value to obtain a new noise reduction parameter. The noise reduction parameter is used to adjust the active noise reduction playback value in order to achieve active noise reduction.

2. The active noise reduction method for audio according to claim 1, characterized in that, The step of increasing or decreasing the noise reduction parameters by a factor of 1 based on a preset fixed step size parameter, using the convolution of the new feedback path estimate and the residual audio value, to obtain new noise reduction parameters includes: Set the preset fixed step size parameter to be less than the reciprocal of the current audio playback value; The fixed step size parameter, the new feedback path estimate, and the residual audio value are convolved to obtain the dynamic step size value. Based on the dynamic step size value, the noise reduction parameters are increased or decreased to obtain new noise reduction parameters.

3. The active noise reduction method for audio according to claim 2, characterized in that, The step of increasing or decreasing the noise reduction parameters according to the dynamic step size value to obtain new noise reduction parameters includes: Based on the time-domain change, determine whether the absolute value of the residual audio value decreases; If the absolute value decreases, the noise reduction parameter is reduced by the dynamic step size value to obtain a new noise reduction parameter; If the absolute value does not decrease, the noise reduction parameter is increased by the dynamic step size value to obtain a new noise reduction parameter.

4. The active noise reduction method for audio according to claim 2 or 3, characterized in that, The step of convolving the fixed step size parameter, the new feedback path estimate, and the residual audio value to obtain the dynamic step size value includes: The fixed step size parameter, the new feedback path estimate, and the residual audio value are double-convolved to obtain the dynamic step size value.

5. The active noise reduction method for audio according to claim 1, characterized in that, The step of performing residual calculation processing on the feedback audio value based on the preset feedback path estimate and residual algorithm to generate residual audio value includes: The estimated feedback path value is convolved with the speaker playback value to generate a predicted residual value. The residual audio value is generated by subtracting the predicted residual value from the feedback audio value.

6. The active noise reduction method for audio according to claim 5, characterized in that, The method based on the preset LMS algorithm, which uses the convolution of the feedback audio value and the audio playback value to iteratively increment and decrement the feedback path estimate to obtain a new feedback path estimate, includes: Calculate the convolution between the feedback audio value and the audio playback value to obtain the iteration value; Based on the time-domain change, determine whether the absolute value of the residual audio value decreases; If the absolute value does not decrease, the feedback path estimate is increased by the iteration value to obtain a new feedback path estimate. If the absolute value decreases, the feedback path estimate is reduced by the iteration value to obtain a new feedback path estimate.

7. The active noise reduction method for audio according to claim 1, characterized in that, The step of superimposing the audio playback value and the active noise cancellation playback value to obtain the speaker playback value includes: The active noise reduction playback value is obtained by adding the frame-by-frame audio energy of the audio playback value.

8. An active noise reduction system for audio, characterized in that, The active noise reduction system for the audio includes: Microphone receiving system, speaker playback system; The microphone receiving system is used to receive external noise, generate audio input values, and generate noise estimates based on preset input parameters, wherein the input parameters are the attenuation ratio of the external noise being recorded. The speaker playback system is used to convolve preset noise reduction parameters with the noise estimate to obtain active noise reduction playback values, and to generate audio playback values ​​based on preset audio files; The audio playback value and the active noise reduction playback value are superimposed to obtain the speaker playback value, and the speaker playback value is subjected to audio playback processing to generate playback audio. The microphone receiving system is used to receive the feedback audio of the played audio through the feedback path, obtain the feedback audio value, and perform residual calculation processing on the feedback audio value according to the preset feedback path estimate and residual algorithm to generate residual audio value. The feedback path is the acoustic path from the speaker playback system to the microphone receiving system, and the feedback path estimate is the attenuation coefficient of the feedback path for sound propagation. The speaker playback system is used to perform iterative increment and decrement processing on the feedback path estimate based on a preset LMS algorithm, using the convolution of the feedback audio value and the audio playback value, to obtain a new feedback path estimate. Based on the preset fixed step size parameter, the noise reduction parameter is increased or decreased by a multiple step size using the convolution of the new feedback path estimate and the residual audio value to obtain a new noise reduction parameter. The noise reduction parameter is used to adjust the active noise reduction playback value in order to achieve active noise reduction.

9. An active noise reduction device for audio, characterized in that, The active noise cancellation device for audio includes: a memory and at least one processor, wherein the memory stores instructions, and the memory and the at least one processor are interconnected via a line; The at least one processor invokes the instructions in the memory to cause the active noise cancellation device for audio to perform the active noise cancellation method for audio as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program thereon, characterized in that, When the computer program is executed by the processor, it implements the active noise reduction method for audio as described in any one of claims 1-7.

Images