Voice compensation method and device for Bluetooth headset

Abstract

The present invention provides a voice compensation method for a Bluetooth headset, which is used for compensating uplink voice in at least one predetermined noise environment, comprising: pre-storing at least one voice compensation spectrum, each voice compensation spectrum corresponding to a predetermined noise environment, each voice compensation spectrum including voice amplitude compensation values ​​at a plurality of preset frequency points in the corresponding predetermined noise environment; receiving current environmental noise, judging whether the current noise environment belongs to one of the predetermined noise environments, and if so, selecting the voice compensation spectrum in the predetermined noise environment corresponding to the current noise environment as the current voice compensation spectrum; receiving a current uplink audio signal, performing noise reduction processing on the current uplink audio signal, and obtaining a current noise reduction voice signal; and superimposing the voice amplitude compensation value corresponding to each target frequency point in the current voice compensation spectrum on the basis of the current noise reduction voice signal, wherein the target frequency point is a frequency point included in a frequency band of the current noise reduction voice signal.

Description

Technical Field

[0001] The present invention relates to the field of audio processing, and in particular to a voice compensation method and device for a Bluetooth headset. Background Art

[0002] With the rapid adoption of Bluetooth headsets in the consumer market, they have quickly become an integral part of people's lives. People are no longer satisfied with simply using Bluetooth headsets for music; more and more people are choosing to use them for voice calls. Compared to using the phone's receiver directly for calls, using a Bluetooth headset frees your hands while significantly reducing the amount of radiation your body receives from the phone.

[0003] The quality of Bluetooth headset voice calls has gradually become a key consideration when choosing a Bluetooth headset. Currently, Bluetooth headsets are primarily used by office workers in major cities, who frequently make calls during their commutes. However, commutes are often subject to a high level of noise, such as sirens, tires on the road, and engine noise. Alternatively, when Bluetooth headset users are discussing project plans with clients in the office, ambient noise (such as other voices) can affect the clarity of the voice heard by the remote party (the recipient of the call).

[0004] Existing uplink noise reduction algorithms vary widely. If the noise reduction strength is too high, some voice signals may be lost, while if the noise reduction strength is too low, some ambient noise may remain. To address this problem, existing technologies have attempted to compensate for the voice signal through voice compensation. However, existing voice compensation algorithms are complex, involve multiple calculation steps, and suffer from large errors. Summary of the Invention

[0005] Based on the above situation, the main purpose of the present invention is to provide a voice compensation method for Bluetooth headsets. The algorithm is very simple and the voice compensation effect is very good, which greatly improves the call quality.

[0006] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] A speech compensation method for a Bluetooth headset is used to compensate for uplink speech of the Bluetooth headset in at least one predetermined noise environment, the method comprising:

[0008] At least one speech compensation spectrum is pre-stored, each speech compensation spectrum corresponds to a predetermined noise environment, and each speech compensation spectrum includes speech amplitude compensation values ​​at a plurality of preset frequency points in the corresponding predetermined noise environment, wherein the plurality of preset frequency points are frequency points distributed within the voice call bandwidth of the Bluetooth headset;

[0009] receiving current environmental noise, determining whether the current noise environment belongs to one of the predetermined noise environments, and if so, selecting a speech compensation spectrum under the predetermined noise environment corresponding to the current noise environment as the current speech compensation spectrum;

[0010] receiving a current uplink audio signal, and performing noise reduction processing on the current uplink audio signal to obtain a current noise-reduced voice signal;

[0011] Based on the current noise reduction speech signal, the speech amplitude compensation value corresponding to the frequency point in the current speech compensation spectrum is superimposed on each target frequency point, where the target frequency point is a frequency point included in the frequency band of the current noise reduction speech signal and belongs to a subset of the multiple preset frequency points.

[0012] Preferably, the speech amplitude compensation value is a predetermined speech amplitude difference between the first speech signal and the second speech signal at the plurality of preset frequency points.

[0013] The first speech signal is a signal of a pure speech signal after the noise reduction processing;

[0014] The second speech signal is a signal obtained by subjecting the clean speech signal to the noise reduction processing in the predetermined noise environment.

[0015] Preferably, the determining whether the current noise environment belongs to one of the predetermined noise environments includes:

[0016] At least two noise amplitude spectrums are pre-stored, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes noise amplitudes at a plurality of preset frequency points in the corresponding predetermined noise environment;

[0017] Calculating a current ambient noise amplitude spectrum, where the current ambient noise amplitude spectrum includes the amplitudes of the current ambient noise at the plurality of preset frequency points;

[0018] Calculating the amplitude differences between the current environmental noise amplitude spectrum and the noise amplitude spectrum of each predetermined noise environment at corresponding frequency points respectively;

[0019] Calculate the accumulated values ​​of the amplitude differences corresponding to each predetermined noise environment respectively;

[0020] The predetermined noise environment corresponding to the minimum accumulated value is determined as the predetermined noise environment to which the current noise environment belongs.

[0021] Preferably, the determining whether the current noise environment belongs to one of the predetermined noise environments includes:

[0022] At least one noise amplitude spectrum is pre-stored, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes noise amplitudes at a plurality of preset frequency points in the corresponding predetermined noise environment;

[0023] Calculating a current ambient noise amplitude spectrum, where the current ambient noise amplitude spectrum includes the amplitudes of the current ambient noise at the plurality of preset frequency points;

[0024] Calculating the amplitude differences between the current environmental noise amplitude spectrum and the noise amplitude spectrum of each predetermined noise environment at corresponding frequency points respectively;

[0025] Calculate the accumulated values ​​of the amplitude differences corresponding to each predetermined noise environment respectively;

[0026] If any accumulated value is less than or equal to the preset value, the predetermined noise environment corresponding to the accumulated value is determined as the predetermined noise environment to which the current noise environment belongs.

[0027] Preferably, the determining whether the current noise environment belongs to one of the predetermined noise environments includes:

[0028] At least two noise amplitude spectrums are pre-stored, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes noise amplitudes at a plurality of preset frequency points in the corresponding predetermined noise environment;

[0029] Divide the pre-stored noise amplitude spectrum into levels according to the noise intensity of the corresponding noise environment;

[0030] Calculating the noise intensity of the current ambient noise to obtain the current noise intensity;

[0031] If the current noise intensity is within the preset noise intensity level range, calculating the current environmental noise amplitude spectrum, where the current environmental noise amplitude spectrum includes the amplitudes of the current environmental noise at the plurality of preset frequency points;

[0032] Calculating the amplitude differences between the current ambient noise amplitude spectrum and the pre-stored noise amplitude spectrum within the level range of the current noise intensity at corresponding frequency points;

[0033] Calculating the accumulated values ​​of the amplitude differences corresponding to each pre-stored noise amplitude spectrum respectively;

[0034] The predetermined noise environment corresponding to the minimum accumulated value is determined as the predetermined noise environment to which the current noise environment belongs.

[0035] Preferably, the determining whether the current noise environment belongs to one of the predetermined noise environments includes:

[0036] At least one noise amplitude spectrum is pre-stored, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes noise amplitudes at a plurality of preset frequency points in the corresponding predetermined noise environment;

[0037] Divide the pre-stored noise amplitude spectrum into levels according to the noise intensity of the corresponding noise environment;

[0038] Calculating the noise intensity of the current ambient noise to obtain the current noise intensity;

[0039] If the current noise intensity is within the preset noise intensity level range, calculating the current environmental noise amplitude spectrum, where the current environmental noise amplitude spectrum includes the amplitudes of the current environmental noise at the plurality of preset frequency points;

[0040] Calculating the amplitude differences between the current ambient noise amplitude spectrum and the pre-stored noise amplitude spectrum within the level range of the current noise intensity at corresponding frequency points;

[0041] Calculating the accumulated values ​​of the amplitude differences corresponding to each pre-stored noise amplitude spectrum respectively;

[0042] If any accumulated value is less than or equal to the preset value, the predetermined noise environment corresponding to the accumulated value is determined as the predetermined noise environment to which the current noise environment belongs.

[0043] Preferably, the preset noise intensity levels include weak noise, medium noise and strong noise.

[0044] The weak noise refers to a noise intensity less than 60dB;

[0045] The medium noise refers to a noise intensity greater than or equal to 60dB and less than 80dB;

[0046] The strong noise refers to noise intensity greater than or equal to 80dB and less than 100dB.

[0047] Preferably, the noise reduction process uses a deep noise reduction algorithm (DNS).

[0048] Preferably, the voice compensation method for a Bluetooth headset further comprises:

[0049] When no uplink audio signal is detected, ambient noise is received to determine whether the noise environment belongs to one of the predetermined noise environments. If yes, the speech compensation spectrum under the predetermined noise environment corresponding to the noise environment is updated as the current speech compensation spectrum.

[0050] The present invention also provides a speech compensation device for a Bluetooth headset, which is used to compensate the uplink speech of the Bluetooth headset in at least one predetermined noise environment, including a speech compensation spectrum storage module, a speech compensation spectrum selection module, a noise reduction module and a compensation module, wherein:

[0051] The speech compensation spectrum storage module is used to pre-store at least one speech compensation spectrum, each speech compensation spectrum corresponding to a predetermined noise environment, and each speech compensation spectrum includes speech amplitude compensation values ​​at multiple preset frequency points in the corresponding predetermined noise environment, and the multiple preset frequency points are frequency points distributed within the voice call bandwidth of the Bluetooth headset;

[0052] The speech compensation spectrum selection module is used to determine whether the current noise environment belongs to one of the predetermined noise environments, and if so, select the speech compensation spectrum in the predetermined noise environment corresponding to the current noise environment as the current speech compensation spectrum in the speech compensation spectrum storage module;

[0053] The noise reduction module is used to perform noise reduction processing on the current uplink audio signal to obtain a current noise-reduced voice signal;

[0054] The compensation module is used to superimpose the speech amplitude compensation value corresponding to each target frequency point in the current speech compensation spectrum on the basis of the current noise reduction speech signal, where the target frequency point is a frequency point included in the frequency band of the current noise reduction speech signal and belongs to a subset of the multiple preset frequency points.

[0055] Preferably, the system further comprises a voice amplitude compensation value calculation module for calculating the predetermined voice amplitude difference between the first voice signal and the second voice signal at the plurality of preset frequency points.

[0056] The first speech signal is a signal of a pure speech signal after the noise reduction processing;

[0057] The second speech signal is a signal obtained by subjecting the pure speech signal to the noise reduction processing in the predetermined noise environment.

[0058] Preferably, the speech compensation spectrum selection module includes a noise amplitude spectrum storage module, a noise amplitude spectrum calculation unit, a noise amplitude difference calculation unit and a noise environment judgment unit, wherein:

[0059] The noise amplitude spectrum storage module is used to pre-store at least two noise amplitude spectra, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes noise amplitudes at multiple preset frequency points in the corresponding predetermined noise environment;

[0060] The noise amplitude spectrum calculation unit is used to calculate the current environmental noise amplitude spectrum, where the current environmental noise amplitude spectrum includes the amplitudes of the current environmental noise at the multiple preset frequency points;

[0061] The noise amplitude difference calculation unit is used to respectively calculate the amplitude differences between the current environmental noise amplitude spectrum and the noise amplitude spectrum of each predetermined noise environment at corresponding frequency points, and respectively calculate the cumulative value of each amplitude difference corresponding to each predetermined noise environment;

[0062] The noise environment judgment unit is configured to determine the predetermined noise environment corresponding to the minimum accumulated value as the predetermined noise environment to which the current noise environment belongs.

[0063] Preferably, the speech compensation spectrum selection module includes a noise amplitude spectrum storage module, a noise amplitude spectrum calculation unit, a noise amplitude difference calculation unit and a noise environment judgment unit, wherein:

[0064] The noise amplitude spectrum storage module is used to pre-store at least one noise amplitude spectrum, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes noise amplitudes at multiple preset frequency points in the corresponding predetermined noise environment;

[0065] The noise amplitude spectrum calculation unit is used to calculate the current environmental noise amplitude spectrum, where the current environmental noise amplitude spectrum includes the amplitudes of the current environmental noise at the multiple preset frequency points;

[0066] The noise amplitude difference calculation unit is used to respectively calculate the amplitude differences between the current environmental noise amplitude spectrum and the noise amplitude spectrum of each predetermined noise environment at corresponding frequency points, and respectively calculate the cumulative value of each amplitude difference corresponding to each predetermined noise environment;

[0067] The noise environment judgment unit is configured to determine, when any accumulated value is less than or equal to a preset value, the predetermined noise environment corresponding to the accumulated value as the predetermined noise environment to which the current noise environment belongs.

[0068] Preferably, the speech compensation spectrum selection module includes a noise amplitude spectrum storage module, a noise intensity calculation unit, a noise intensity level judgment unit, a noise amplitude spectrum calculation unit, a noise amplitude difference calculation unit and a noise environment judgment unit, wherein,

[0069] The noise amplitude spectrum storage module is used to pre-store at least two noise amplitude spectrums, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes noise amplitudes at multiple preset frequency points in the corresponding predetermined noise environment;

[0070] The noise intensity calculation unit is used to classify the pre-stored noise amplitude spectrum according to the noise intensity of the corresponding noise environment, and calculate the noise intensity of the current environmental noise to obtain the current noise intensity;

[0071] The noise intensity level determination unit is used to determine whether the current noise intensity is within a preset noise intensity level range;

[0072] The noise amplitude spectrum calculation unit is configured to calculate the current environmental noise amplitude spectrum when the noise intensity level judgment unit determines that the noise intensity level is yes, where the current environmental noise amplitude spectrum includes the amplitudes of the current environmental noise at the plurality of preset frequency points;

[0073] The noise amplitude difference calculation unit is used to respectively calculate the amplitude difference between the current ambient noise amplitude spectrum and each pre-stored noise amplitude spectrum within the level range of the current noise intensity at the corresponding frequency point, and respectively calculate the cumulative value of each amplitude difference corresponding to each pre-stored noise amplitude spectrum within the level range of the current noise intensity;

[0074] The noise environment judgment unit is configured to determine the predetermined noise environment corresponding to the minimum accumulated value as the predetermined noise environment to which the current noise environment belongs.

[0075] Preferably, the speech compensation spectrum selection module includes a noise amplitude spectrum storage module, a noise intensity calculation unit, a noise intensity level judgment unit, a noise amplitude spectrum calculation unit, a noise amplitude difference calculation unit and a noise environment judgment unit, wherein,

[0076] The noise amplitude spectrum storage module is further configured to pre-store at least one noise amplitude spectrum, each noise amplitude spectrum corresponding to a predetermined noise environment, and the noise amplitude spectrum including noise amplitudes at a plurality of preset frequency points in the corresponding predetermined noise environment;

[0077] The noise intensity calculation unit is used to classify the pre-stored noise amplitude spectrum according to the noise intensity of the corresponding noise environment, and calculate the noise intensity of the current environmental noise to obtain the current noise intensity;

[0078] The noise intensity level determination unit is used to determine whether the current noise intensity is within a preset noise intensity level range;

[0079] The noise amplitude spectrum calculation unit is configured to calculate the current environment noise amplitude spectrum when the noise environment judgment unit determines that the noise environment is positive, where the current environment noise amplitude spectrum includes the amplitudes of the current environment noise at the plurality of preset frequency points;

[0080] The noise amplitude difference calculation unit is used to respectively calculate the amplitude differences between the current ambient noise amplitude spectrum and the pre-stored noise amplitude spectrum within the level range of the current noise intensity at corresponding frequency points, and respectively calculate the cumulative value of each amplitude difference corresponding to each pre-stored noise amplitude spectrum within the level range of the current noise intensity;

[0081] The noise environment judgment unit is configured to determine, when any accumulated value is less than or equal to a preset value, the predetermined noise environment corresponding to the accumulated value as the predetermined noise environment to which the current noise environment belongs.

[0082] Preferably, the preset noise intensity levels include weak noise, medium noise and strong noise.

[0083] The weak noise refers to a noise intensity less than 60dB;

[0084] The medium noise refers to a noise intensity greater than or equal to 60dB and less than 80dB;

[0085] The strong noise refers to noise intensity greater than or equal to 80dB and less than 100dB.

[0086] Preferably, the noise reduction module includes a deep noise reduction unit, and the deep noise reduction unit is configured to perform noise reduction processing on the current uplink audio signal by adopting a deep noise reduction algorithm.

[0087] Preferably, the speech compensation spectrum selection module is also used to determine whether the noise environment belongs to one of the predetermined noise environments when there is no uplink audio signal. If so, update the speech compensation spectrum under the predetermined noise environment corresponding to the noise environment as the current speech compensation spectrum.

[0088] The present invention also provides a Bluetooth headset, which adopts the method described in the present invention to compensate for uplink voice, or includes the voice compensation device described in the present invention.

[0089] The present invention also provides a computer-readable storage medium having a computer program stored thereon, and the computer program can implement the method of the present invention when executed.

[0090] The present invention also provides an audio processing chip, comprising a processor and a memory, wherein the memory stores a computer program, and the processor is capable of executing the computer program to implement the method described in the present invention.

[0091] The voice compensation method for a Bluetooth headset of the present invention is designed for the situation where the headset operates in a predetermined noise environment. By pre-storing the voice compensation spectrum corresponding to the predetermined noise environment, if the current noise environment is one of the predetermined noise environments, the voice compensation spectrum of this predetermined noise environment is selected to perform accurate voice compensation on the current noise reduction voice signal at multiple preset frequency points. The algorithm is very simple, and the voice compensation effect is very good, which greatly improves the call quality.

[0092] The speech compensation device of the present invention, for the case of working in a predetermined noise environment, pre-stores the speech compensation spectrum corresponding to the predetermined noise environment through a speech compensation spectrum storage module. If the current noise environment is one of the predetermined noise environments, the speech compensation spectrum selection module selects the speech compensation spectrum of this predetermined noise environment. The compensation module uses the speech compensation spectrum of the predetermined noise environment selected by the speech compensation spectrum selection module to perform accurate speech compensation on the current noise reduction speech signal at multiple preset frequency points. The algorithm is simple, the speech compensation effect is good, and the call quality is greatly improved.

[0093] Other beneficial effects of the present invention will be explained through the introduction of specific technical features and technical solutions in the specific implementation methods. Those skilled in the art should be able to understand the beneficial technical effects brought about by the introduction of these technical features and technical solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0094] The following describes preferred embodiments of the adaptive speech compensation method and apparatus according to the present invention with reference to the accompanying drawings.

[0095] Figure 1 A flow chart of a voice compensation method for a Bluetooth headset according to a preferred embodiment of the present invention;

[0096] Figure 2 A flow chart of a method for determining whether a current noise environment belongs to one of predetermined noise environments according to a preferred embodiment of the present invention;

[0097] Figure 3 A speech compensation spectrum (compensation curve) corresponding to a weak noise environment according to a preferred embodiment of the present invention, and a corresponding speech compensation curve;

[0098] Figure 4 A speech compensation spectrum (compensation curve) corresponding to a medium noise environment according to a preferred embodiment of the present invention, and a corresponding speech compensation curve;

[0099] Figure 5 A speech compensation spectrum (compensation curve) corresponding to a strong noise environment according to a preferred embodiment of the present invention, and a corresponding speech compensation curve;

[0100] Figure 6 A flowchart of another method for determining whether a current noise environment belongs to a predetermined noise environment according to a preferred embodiment of the present invention;

[0101] Figure 7 A flowchart of another method for determining whether a current noise environment belongs to a predetermined noise environment according to a preferred embodiment of the present invention;

[0102] Figure 8 A flowchart of another method for determining whether a current noise environment belongs to a predetermined noise environment according to a preferred embodiment of the present invention;

[0103] Figure 9 A flowchart of another method for determining whether a current noise environment belongs to a predetermined noise environment according to a preferred embodiment of the present invention;

[0104] Figure 10 A schematic structural diagram of a voice compensation device for a Bluetooth headset according to a preferred embodiment of the present invention;

[0105] Figure 11 This is a structural diagram of a voice compensation device for a Bluetooth headset according to another preferred embodiment of the present invention;

[0106] Figure 12 A schematic structural diagram of a speech compensation spectrum selection module according to another preferred embodiment of the present invention;

[0107] Figure 13 Schematic diagram of the structure of a speech compensation spectrum selection module according to another preferred embodiment of the present invention. DETAILED DESCRIPTION

[0108] Figure 1 This is a flow chart of a voice compensation method for a Bluetooth headset according to a preferred embodiment of the present invention, which is used to compensate for the uplink voice of the Bluetooth headset in at least one predetermined noise environment. The predetermined noise environment can generally be an office noise environment, a subway noise environment, a noisy road noise environment, a noisy fan noise environment, etc.

[0109] Figure 1 The speech compensation methods in include:

[0110] Step 101: Prestore at least one speech compensation spectrum, each corresponding to a predetermined noise environment. Each speech compensation spectrum includes speech amplitude compensation values ​​at multiple preset frequency points within the corresponding predetermined noise environment. The multiple preset frequency points are frequencies distributed within the Bluetooth headset voice call bandwidth. The Bluetooth headset voice call bandwidth refers to the frequency range of voice signals that can be transmitted by the Bluetooth headset, typically 100Hz-8000Hz. Frequencies are typically selected within this range in 1Hz increments as the multiple preset frequency points. Within the technically permitted range, the increment can be greater than or less than 1Hz.

[0111] Step 102: Receive the current ambient noise and determine whether the current noise environment is one of the predetermined noise environments. If so, select the speech compensation spectrum for the predetermined noise environment corresponding to the current noise environment as the current speech compensation spectrum. In actual use, the current ambient noise can be received before the Bluetooth headset is activated but before a voice call is made. If the current noise environment is determined to be one of the predetermined noise environments, the speech compensation spectrum corresponding to the predetermined noise environment is selected as the current speech compensation spectrum.

[0112] Step 103: Receive the current uplink audio signal and perform noise reduction processing on the current uplink audio signal to obtain the current noise-reduced voice signal. When a voice call is initiated using a Bluetooth headset, noise reduction processing is performed on the current uplink audio signal to obtain a noise-reduced voice signal, where the uplink audio signal includes the current ambient noise signal and the current voice signal.

[0113] Step 104: Based on the current noise-reduced speech signal, the speech amplitude compensation value corresponding to each target frequency point in the current speech compensation spectrum is superimposed. The target frequency point is a frequency point included in the frequency band of the current noise-reduced speech signal and belongs to a subset of multiple preset frequencies. The speech amplitude compensation values ​​in the current speech compensation spectrum are superimposed on multiple frequencies within the frequency band of the current noise-reduced speech signal to compensate for the loss of the speech signal caused by the noise reduction process.

[0114] The speech compensation method of the present invention is designed for working in a predetermined noise environment. By pre-storing the speech compensation spectrum corresponding to the predetermined noise environment, if the current noise environment is one of the predetermined noise environments, the speech compensation spectrum of this predetermined noise environment is selected to perform accurate speech compensation on the current noise reduction speech signal at multiple preset frequency points. The algorithm is very simple, and the speech compensation effect is very good, greatly improving the call quality.

[0115] In an optional embodiment, the speech amplitude compensation values ​​at multiple preset frequency points in the pre-stored speech compensation spectrum in step 101 are obtained through prior testing. In a specific embodiment, the speech amplitude compensation values ​​may be predetermined speech amplitude differences between a first speech signal and a second speech signal at multiple preset frequency points, where the first speech signal is a pure speech signal subjected to noise reduction processing, and the second speech signal is the same pure speech signal subjected to noise reduction processing in a predetermined noise environment. The noise reduction processing here uses the same noise reduction algorithm as the noise reduction processing in step 103. The speech amplitude may be selected as the amplitude of speech intensity.

[0116] Typically, after a pure speech signal undergoes noise reduction processing, there is almost no speech loss. However, in a predetermined noise environment, the same pure speech signal will experience obvious speech loss after noise reduction processing. The above embodiment records the difference between the two at multiple preset frequency points to form a speech compensation spectrum for the predetermined noise environment. The speech compensation value is simple to calculate and is targeted, and can achieve accurate speech compensation.

[0117] In an optional embodiment, in step 102, determining whether the current noise environment belongs to one of the predetermined noise environments can be achieved by determining the noise intensity of the current environmental noise. Figure 2 As shown, it includes: step 201: calculating the noise intensity of the current environmental noise to obtain the current noise intensity; step 202: if the noise intensity is within the preset noise intensity level range, determining that the current noise environment belongs to one of the predetermined noise environments.

[0118] In a specific embodiment, the preset noise intensity levels can be divided into three levels: weak noise, medium noise, and strong noise. Specifically, weak noise refers to a noise intensity less than 60dB, medium noise refers to a noise intensity greater than or equal to 60dB and less than 80dB, and strong noise refers to a noise intensity greater than or equal to 80dB and less than 100dB. By classifying the noise intensity levels, pre-stored speech compensation spectra are corresponding to the noise levels. In other words, a speech compensation spectrum is pre-stored for each of weak noise, medium noise, and strong noise. If the current noise falls into one of these noise levels, the corresponding speech compensation spectrum is selected as the current speech compensation spectrum.

[0119] Figure 3-5 The corresponding speech compensation spectra (compensation curves) under weak noise, medium noise and strong noise conditions are respectively given as examples, as well as the speech compensation curves realized by adopting the technical solution of the present invention. Figure 3 The lower middle curve is the speech compensation spectrum corresponding to the weak noise environment. In the upper curve, the dotted curve is the speech signal obtained after noise reduction, and the solid curve is the call signal after speech compensation according to the technical solution of the present invention. Figure 4 The lower middle curve is the speech compensation spectrum corresponding to the medium noise environment. In the upper curve, the dotted curve is the speech signal obtained after noise reduction, and the solid curve is the call signal after speech compensation according to the technical solution of the present invention. Figure 5 The lower middle curve is the speech compensation spectrum corresponding to a strong noise environment. In the upper curve, the dotted curve is the speech signal obtained after noise reduction, and the solid curve is the call signal after speech compensation according to the technical solution of the present invention. Figure 3-5 It can be clearly seen that different speech compensation spectra are pre-stored for different noise environments, and speech compensation is carried out in a targeted manner, with very good compensation effects.

[0120] In an optional embodiment, in step 102, determining whether the current noise environment belongs to one of the predetermined noise environments can be achieved by determining the noise category of the current environmental noise. Figure 6 As shown, it includes: step 301: pre-storing at least two noise amplitude spectra, each noise amplitude spectrum corresponds to a predetermined noise environment, wherein the noise amplitude spectrum includes the noise amplitudes at multiple preset frequency points in the corresponding predetermined noise environment; step 302: calculating the current environment noise amplitude spectrum, the current environment noise amplitude spectrum includes the amplitudes of the current environment noise at the multiple preset frequency points; step 303: respectively calculating the amplitude differences between the current environment noise amplitude spectrum and the noise amplitude spectrum of each predetermined noise environment at the corresponding frequency points; step 304: respectively calculating the cumulative value of each amplitude difference corresponding to each predetermined noise environment; step 305: determining the predetermined noise environment corresponding to the minimum cumulative value as the predetermined noise environment to which the current noise environment belongs.

[0121] By pre-storing the noise amplitude spectra of at least two predetermined noise environments, after detecting the current environment noise, the amplitude spectrum of the current environment noise is calculated, and then calculating which of the pre-stored noise amplitude spectra the current environment noise amplitude spectrum is closest to. That is, in steps 303-305, the predetermined noise environment corresponding to the minimum cumulative value is selected to determine as the predetermined noise environment to which the current noise environment belongs, and the current noise environment is accurately located. Then, the speech compensation spectrum of the current noise environment can be accurately located to achieve accurate speech compensation.

[0122] Figure 7 Another method for determining whether a current noise environment belongs to one of the predetermined noise environments includes: step 401: pre-storing at least one noise amplitude spectrum, each noise amplitude spectrum corresponding to a predetermined noise environment, wherein the noise amplitude spectrum includes the noise amplitudes at multiple preset frequency points in the corresponding predetermined noise environment; step 402: calculating the current environment noise amplitude spectrum, the current environment noise amplitude spectrum including the amplitudes of the current environment noise at the multiple preset frequency points; step 403: respectively calculating the amplitude differences between the current environment noise amplitude spectrum and the noise amplitude spectrum of each predetermined noise environment at the corresponding frequency points; step 404: respectively calculating the cumulative value of each amplitude difference corresponding to each predetermined noise environment; step 405: if any cumulative value is less than or equal to the preset value, then determining the predetermined noise environment corresponding to the cumulative value as the predetermined noise environment to which the current noise environment belongs. In a specific embodiment, the preset value can be determined based on technical requirements and factors of the preset environment noise itself, and the present invention is not limited thereto.

[0123] If the minimum accumulated value is also very large, that is, if it is greater than the preset value, the predetermined noise environment cannot be found, and the predetermined compensation spectrum is not applicable. In this case, another speech compensation method can be selected. If both accumulated values ​​are less than the preset value, it means there are two relatively close predetermined environments, and the speech compensation spectrum of either one can be selected for calculation.

[0124] In an optional embodiment, in step 102, determining whether the current noise environment belongs to one of the predetermined noise environments can be achieved by first determining the noise intensity level of the current ambient noise, and then determining which noise category the current noise belongs to within the current noise level. Figure 8 As shown, the method includes: step 501, pre-storing at least two noise amplitude spectrums, each noise amplitude spectrum corresponding to a predetermined noise environment, and the noise amplitude spectrum including the noise amplitudes at multiple preset frequency points in the corresponding predetermined noise environment; step 502: dividing the pre-stored noise amplitude spectrum into levels according to the noise intensity of the corresponding noise environment; step 503, calculating the noise intensity of the current environment noise to obtain the current noise intensity; step 504, if the current noise intensity is within the preset noise intensity level range, calculating the current environment noise amplitude spectrum, which is the amplitude of the current environment noise at the multiple preset frequency points; step 505, respectively calculating the amplitude difference between the current environment noise amplitude spectrum and the pre-stored noise amplitude spectrum within the level range of the current noise intensity at the corresponding frequency points; step 506, respectively calculating the cumulative value of each amplitude difference corresponding to each pre-stored noise amplitude spectrum; step 507, determining the predetermined noise environment corresponding to the minimum cumulative value as the predetermined noise environment to which the current noise environment belongs.

[0125] By pre-storing the noise amplitude spectra of at least two predetermined noise environments and first determining the current noise level, the system calculates only which of the pre-stored noise amplitude spectra within the current noise level unit is closest to the current noise amplitude spectrum. This system then determines whether the current noise environment is the one with the closest noise amplitude spectrum. This allows the system to accurately locate the current noise environment, further narrowing the comparison range, reducing the computational effort, and improving speech compensation speed.

[0126] Figure 9Another method for determining whether the current noise environment belongs to one of the predetermined noise environments includes: step 601, pre-storing at least one noise amplitude spectrum, each noise amplitude spectrum corresponding to a predetermined noise environment, the noise amplitude spectrum including the noise amplitudes at a plurality of preset frequency points in the corresponding predetermined noise environment; step 602: classifying the pre-stored noise amplitude spectrum according to the noise intensity of the corresponding noise environment; step 603, calculating the noise intensity of the current environment noise to obtain the current noise intensity; step 604, if the current noise intensity is within the preset noise intensity level range, range, then calculate the current environmental noise amplitude spectrum, which is the amplitude of the current environmental noise at the multiple preset frequency points; step 605, respectively calculate the amplitude differences between the current environmental noise amplitude spectrum and the pre-stored noise amplitude spectrum within the level range of the current noise intensity at the corresponding frequency points; step 606, respectively calculate the cumulative values ​​of the amplitude differences corresponding to each pre-stored noise amplitude spectrum; step 607, if any cumulative value is less than or equal to the preset value, then determine the predetermined noise environment corresponding to the cumulative value as the predetermined noise environment to which the current noise environment belongs.

[0127] In a specific embodiment, the preset noise intensity levels may include weak noise, medium noise and strong noise, weak noise means noise intensity less than 60dB, medium noise means noise intensity greater than or equal to 60dB and less than 80dB, and strong noise means noise intensity greater than or equal to 80dB and less than 100dB.

[0128] In an optional embodiment, a deep noise reduction algorithm may be used, namely the DNS (Deep noisesuppression) noise reduction algorithm, which is a type of neural network noise reduction algorithm. The greater the DNS noise reduction intensity (Over_Drive), the stronger the noise reduction, and the smaller the DNS gain coefficient (Gain_Floor), the stronger the noise reduction. The noise reduction intensity is controlled by adjusting the OverDrive parameter and the GainFloor parameter. GainFloor is the lower limit of the gain, which can prevent excessive noise suppression of small sounds; OverDrive is used to adjust the noise reduction intensity to meet the needs of different noise reduction intensities. Since the DNS noise reduction algorithm has a stronger noise reduction, there will also be more voice loss than other algorithms. Therefore, for the DNS noise reduction algorithm, the speech compensation of the present invention can accurately select the speech compensation spectrum for the predetermined noise environment, and the speech compensation effect is very good.

[0129] In an optional embodiment, the noise reduction algorithm in the speech compensation method of the present invention can adaptively adjust the noise reduction intensity based on the ambient noise. The noise reduction algorithm adjusts the noise reduction intensity based on the current ambient noise. If the current ambient noise is too high, the noise reduction intensity is increased; if the current ambient noise is low, the noise reduction intensity is decreased, ensuring that the current speech is not excessively affected.

[0130] In an optional embodiment, for example, if the user is not in a call and the Bluetooth headset does not detect an uplink audio signal during use of the Bluetooth headset, the current noise environment can be re-detected to determine whether the noise environment falls within one of the predetermined noise environments. If so, the speech compensation spectrum corresponding to the predetermined noise environment is updated as the current speech compensation spectrum. This allows for timely selection of a more appropriate speech compensation curve as the noise environment changes, thereby achieving more accurate compensation for the current speech.

[0131] This embodiment also discloses a voice compensation device for a Bluetooth headset, which is used to compensate the uplink voice of the Bluetooth headset in at least one predetermined noise environment, such as Figure 10 FIG. 1 is a schematic diagram of the structure of a speech compensation device for a Bluetooth headset disclosed in this embodiment. The speech compensation device includes: a speech compensation spectrum storage module 701, a speech compensation spectrum selection module 702, a noise reduction module 703, and a compensation module 704, wherein:

[0132] The speech compensation spectrum storage module 701 is used to pre-store at least one speech compensation spectrum, each speech compensation spectrum corresponds to a predetermined noise environment, and each speech compensation spectrum includes speech amplitude compensation values ​​at multiple preset frequency points in the corresponding predetermined noise environment, and the multiple preset frequency points are frequency points distributed within the voice call bandwidth of the Bluetooth headset; the speech compensation spectrum selection module 702 is used to determine whether the current noise environment belongs to one of the predetermined noise environments. If so, the speech compensation spectrum in the predetermined noise environment corresponding to the current noise environment is selected in the speech compensation spectrum storage module 701 as the current speech compensation spectrum; the noise reduction module 703 is used to perform noise reduction processing on the current uplink audio signal to obtain the current noise-reduced speech signal; the compensation module 704 is used to superimpose the speech amplitude compensation value corresponding to the frequency point in the current speech compensation spectrum for each target frequency point on the basis of the current noise-reduced speech signal. The target frequency point is the frequency point included in the frequency band of the current noise-reduced speech signal and belongs to a subset of multiple preset frequency points.

[0133] The speech compensation device of the present invention, for the case of working in a predetermined noise environment, pre-stores the speech compensation spectrum corresponding to the predetermined noise environment through the speech compensation spectrum storage module 701. If the current noise environment is one of the predetermined noise environments, the speech compensation spectrum selection module 702 selects the speech compensation spectrum of this predetermined noise environment. The compensation module 704 uses the speech compensation spectrum of the predetermined noise environment selected by the speech compensation spectrum selection module 702 to perform accurate speech compensation on the current noise reduction speech signal at multiple preset frequency points. The algorithm is simple, the speech compensation effect is good, and the call quality is greatly improved.

[0134] In an alternative embodiment, Figure 10The speech compensation device in the embodiment may further include a speech amplitude compensation value calculation module 705, such as Figure 11 As shown, the speech amplitude compensation value calculation module 705 is used to calculate the speech amplitude difference between a predetermined first speech signal and a second speech signal at multiple preset frequency points, where the first speech signal is a signal of a pure speech signal after the noise reduction processing; and the second speech signal is a signal of a pure speech signal after the noise reduction processing in the predetermined noise environment.

[0135] In an optional embodiment, the speech compensation spectrum selection module 702 may calculate the noise intensity of the current ambient noise to obtain the current noise intensity; and when the current noise intensity falls within a preset noise intensity level range, the module may determine that the current noise environment falls within one of the predetermined noise environments. In a specific embodiment, the preset noise intensity levels may include weak noise, medium noise, and strong noise. Weak noise may be defined as a noise intensity less than 60 dB, medium noise may be defined as a noise intensity greater than or equal to 60 dB and less than 80 dB, and strong noise may be defined as a noise intensity greater than or equal to 80 dB and less than 100 dB.

[0136] In an optional embodiment, if Figure 12 As shown, the speech compensation spectrum selection module 702 may include a noise amplitude spectrum storage module 801, a noise amplitude spectrum calculation unit 802, a noise amplitude difference calculation unit 803, and a noise environment judgment unit 804. The noise amplitude spectrum storage module 801 is configured to pre-store at least two noise amplitude spectra, each corresponding to a predetermined noise environment, and the noise amplitude spectrum includes the noise amplitudes at multiple preset frequency points in the corresponding predetermined noise environment. The noise amplitude spectrum calculation unit 802 is configured to calculate the current environment noise amplitude spectrum, which is the amplitude of the current environment noise at the multiple preset frequency points. The noise amplitude difference calculation unit 803 is configured to respectively calculate the amplitude differences between the current environment noise amplitude spectrum and the noise amplitude spectrum of each predetermined noise environment at the corresponding frequency points, and respectively calculate the cumulative value of the amplitude differences corresponding to each predetermined noise environment. The noise environment judgment unit 804 is configured to determine the predetermined noise environment corresponding to the minimum cumulative value as the predetermined noise environment to which the current noise environment belongs.

[0137] In an alternative embodiment, Figure 12The noise amplitude spectrum storage module 801 in the speech compensation spectrum selection module pre-stores at least one noise amplitude spectrum, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes the noise amplitudes at multiple preset frequency points in the corresponding predetermined noise environment; the functions of the noise amplitude spectrum calculation unit 802 and the noise amplitude difference calculation unit 903 are the same as those in the above embodiment; the noise environment judgment unit 904 is used to determine the predetermined noise environment corresponding to any accumulated value as the predetermined noise environment to which the current noise environment belongs when any accumulated value is less than or equal to the preset value.

[0138] In an optional embodiment, if Figure 13 As shown, the speech compensation spectrum selection module 60 may also include a noise amplitude spectrum storage module 901, a noise intensity calculation unit 902, a noise intensity level judgment unit 903, a noise amplitude spectrum calculation unit 904, a noise amplitude difference calculation unit 905 and a noise environment judgment unit 906, wherein the noise amplitude spectrum storage module 901 is used to pre-store at least two noise amplitude spectra, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes the noise amplitudes at multiple preset frequency points under the corresponding predetermined noise environment; the noise intensity calculation unit 902 is used to divide the pre-stored noise amplitude spectrum into levels according to the noise intensity of the corresponding noise environment, and calculate the noise intensity of the current environmental noise to obtain the current noise intensity; the noise intensity level judgment unit 903 Used to determine whether the current noise intensity is within the preset noise intensity level range; the noise amplitude spectrum calculation unit 904 is used to calculate the current environment noise amplitude spectrum when the noise intensity level judgment unit 903 judges as yes, the current environment noise amplitude spectrum is the amplitude of the current environment noise at the multiple preset frequency points; the noise amplitude difference calculation unit 905 is used to respectively calculate the amplitude difference between the current environment noise amplitude spectrum and each noise amplitude spectrum pre-stored within the current noise intensity level range at the corresponding frequency points, and respectively calculate the cumulative value of each amplitude difference corresponding to each pre-stored noise amplitude spectrum within the level range of the current noise intensity; the noise environment judgment unit 906 is used to determine the predetermined noise environment corresponding to the minimum cumulative value as the predetermined noise environment to which the current noise environment belongs.

[0139] In an alternative embodiment, Figure 13The noise amplitude spectrum storage module 901 in the system pre-stores at least one noise amplitude spectrum, each noise amplitude spectrum corresponding to a predetermined noise environment, and the noise amplitude spectrum includes the noise amplitudes at multiple preset frequency points in the corresponding predetermined noise environment; the functions of the noise intensity calculation unit 902, the noise intensity level judgment unit 903, the noise amplitude spectrum calculation unit 904 and the noise amplitude difference calculation unit 905 are the same as those in the above embodiment; the noise environment judgment unit 906 is used to determine the predetermined noise environment corresponding to any accumulated value as the predetermined noise environment to which the current noise environment belongs when any accumulated value is less than or equal to the preset value.

[0140] In an optional embodiment, the preset noise intensity levels may include weak noise, medium noise and strong noise, where weak noise means noise intensity less than 60dB, medium noise means noise intensity greater than or equal to 60dB and less than 80dB, and strong noise means noise intensity greater than or equal to 80dB and less than 100dB.

[0141] In an optional implementation, the noise reduction module 70 may include a deep noise reduction unit, which is configured to perform noise reduction processing on the current uplink audio signal using a deep noise reduction algorithm (DNS).

[0142] In an optional embodiment, the noise reduction module 70 may further include a noise reduction intensity adjustment unit for adjusting the noise reduction intensity according to the ambient noise.

[0143] In an optional embodiment, the speech compensation spectrum selection module 60 can also be used to determine whether the noise environment belongs to one of the predetermined noise environments when the user is not making a call and the Bluetooth headset does not detect an uplink audio signal during the use of the Bluetooth headset. If so, the speech compensation spectrum in the predetermined noise environment corresponding to the noise environment is updated as the current speech compensation spectrum.

[0144] This embodiment also discloses a Bluetooth headset, which adopts a voice compensation method for a Bluetooth headset disclosed in any of the above embodiments, or includes a voice compensation device for a Bluetooth headset disclosed in any of the above embodiments.

[0145] This embodiment further discloses a computer-readable storage medium having a computer program stored thereon. When the computer program stored in the storage medium is executed, a voice compensation method for a Bluetooth headset disclosed in any of the above embodiments can be implemented.

[0146] This embodiment also discloses an audio processing chip, including a processor and a memory, wherein a computer program is stored in the memory, and the processor can execute the voice compensation method for Bluetooth headset disclosed in any of the above embodiments to complete the voice compensation function.

[0147] According to a speech compensation method for Bluetooth headsets disclosed in an embodiment of the present invention, for the situation of working in a predetermined noise environment, by pre-storing the speech compensation spectrum corresponding to the predetermined noise environment, if the current noise environment is one of the predetermined noise environments, the speech compensation spectrum of this predetermined noise environment is selected to perform accurate speech compensation on the current noise reduction speech signal at multiple preset frequency points. The algorithm is very simple, and the speech compensation effect is very good, which greatly improves the call quality.

[0148] It should be noted that the use of step numbers (letters or numbers) to refer to certain specific method steps in the present invention is solely for the purpose of convenience and brevity, and is in no way intended to limit the order of these method steps. Those skilled in the art will appreciate that the order of the relevant method steps is determined by the technology itself and should not be unduly limited by the use of step numbers.

[0149] Those skilled in the art will appreciate that, provided there is no conflict, the above preferred solutions can be freely combined and superimposed.

[0150] It should be understood that the above-mentioned embodiments are merely illustrative and non-restrictive. Without departing from the basic principles of the present invention, various obvious or equivalent modifications or substitutions that can be made by those skilled in the art to the above-mentioned details will be included in the scope of the claims of the present invention.

Claims

1. A speech compensation method for a Bluetooth headset, used to compensate for the uplink speech of the Bluetooth headset in at least one predetermined noise environment, characterized in that: The method comprises: At least one speech compensation spectrum is pre-stored, each speech compensation spectrum corresponds to a predetermined noise environment, and each speech compensation spectrum includes speech amplitude compensation values ​​at a plurality of preset frequency points in the corresponding predetermined noise environment, wherein the plurality of preset frequency points are frequency points distributed within the voice call bandwidth of the Bluetooth headset; the speech amplitude compensation value is a predetermined speech amplitude difference between a first speech signal and a second speech signal at the plurality of preset frequency points, wherein the first speech signal is a signal of a pure speech signal after noise reduction processing, and the second speech signal is a signal of the pure speech signal after noise reduction processing in the predetermined noise environment; Receiving current environmental noise, determining whether the current noise environment belongs to one of the predetermined noise environments, and if so, selecting a speech compensation spectrum in a predetermined noise environment corresponding to the current noise environment as the current speech compensation spectrum; when no uplink audio signal is detected, receiving environmental noise, determining whether the noise environment belongs to one of the predetermined noise environments, and if so, updating the speech compensation spectrum in the predetermined noise environment corresponding to the noise environment as the current speech compensation spectrum; Receive a current uplink audio signal, perform noise reduction processing on the current uplink audio signal, and obtain a current noise-reduced speech signal, wherein the noise reduction processing adopts a deep noise reduction algorithm (DNS); On the basis of the current noise reduction speech signal, the speech amplitude compensation value corresponding to the frequency point in the current speech compensation spectrum is superimposed for each target frequency point, and the target frequency point is a frequency point included in the frequency band of the current noise reduction speech signal and belongs to a subset of the multiple preset frequency points.

2. The voice compensation method for a Bluetooth headset according to claim 1, characterized in that: The determining whether the current noise environment belongs to one of the predetermined noise environments includes: Pre-store at least two noise amplitude spectrums, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes noise amplitudes at a plurality of preset frequency points in the corresponding predetermined noise environment; Calculating a current environmental noise amplitude spectrum, where the current environmental noise amplitude spectrum includes the amplitudes of the current environmental noise at the plurality of preset frequency points; Calculating the amplitude differences between the current environmental noise amplitude spectrum and the noise amplitude spectrum of each predetermined noise environment at corresponding frequency points respectively; Calculate the accumulated value of each amplitude difference corresponding to each predetermined noise environment respectively; The predetermined noise environment corresponding to the minimum accumulated value is determined as the predetermined noise environment to which the current noise environment belongs.

3. The voice compensation method for a Bluetooth headset according to claim 1, characterized in that: The determining whether the current noise environment belongs to one of the predetermined noise environments includes: Pre-store at least one noise amplitude spectrum, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes noise amplitudes at a plurality of preset frequency points in the corresponding predetermined noise environment; Calculating a current environmental noise amplitude spectrum, where the current environmental noise amplitude spectrum includes the amplitudes of the current environmental noise at the plurality of preset frequency points; Calculating the amplitude differences between the current environmental noise amplitude spectrum and the noise amplitude spectrum of each predetermined noise environment at corresponding frequency points respectively; Calculate the accumulated value of each amplitude difference corresponding to each predetermined noise environment respectively; If any accumulated value is less than or equal to the preset value, the predetermined noise environment corresponding to the accumulated value is determined as the predetermined noise environment to which the current noise environment belongs.

4. The voice compensation method for a Bluetooth headset according to claim 1, characterized in that: The determining whether the current noise environment belongs to one of the predetermined noise environments includes: Pre-store at least two noise amplitude spectrums, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes noise amplitudes at a plurality of preset frequency points in the corresponding predetermined noise environment; Classify the pre-stored noise amplitude spectrum into levels according to the noise intensity of the corresponding noise environment; Calculating the noise intensity of the current environmental noise to obtain the current noise intensity; If the current noise intensity is within the preset noise intensity level range, then calculating the current environmental noise amplitude spectrum, the current environmental noise amplitude spectrum including the amplitudes of the current environmental noise at the plurality of preset frequency points; Respectively calculating the amplitude differences between the current environmental noise amplitude spectrum and the pre-stored noise amplitude spectrum within the level range of the current noise intensity at corresponding frequency points; Calculate the accumulated value of each amplitude difference corresponding to each pre-stored noise amplitude spectrum respectively; The predetermined noise environment corresponding to the minimum accumulated value is determined as the predetermined noise environment to which the current noise environment belongs.

5. The voice compensation method for a Bluetooth headset according to claim 1, characterized in that: The determining whether the current noise environment belongs to one of the predetermined noise environments includes: Pre-store at least one noise amplitude spectrum, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes noise amplitudes at a plurality of preset frequency points in the corresponding predetermined noise environment; Classify the pre-stored noise amplitude spectrum into levels according to the noise intensity of the corresponding noise environment; Calculating the noise intensity of the current environmental noise to obtain the current noise intensity; If the current noise intensity is within the preset noise intensity level range, then calculating the current environmental noise amplitude spectrum, the current environmental noise amplitude spectrum including the amplitudes of the current environmental noise at the plurality of preset frequency points; Respectively calculating the amplitude differences between the current environmental noise amplitude spectrum and the pre-stored noise amplitude spectrum within the level range of the current noise intensity at corresponding frequency points; Calculate the accumulated value of each amplitude difference corresponding to each pre-stored noise amplitude spectrum respectively; If any accumulated value is less than or equal to the preset value, the predetermined noise environment corresponding to the accumulated value is determined as the predetermined noise environment to which the current noise environment belongs.

6. The voice compensation method for a Bluetooth headset according to claim 4 or 5, characterized in that: The preset noise intensity levels include weak noise, medium noise and strong noise. The weak noise refers to a noise intensity less than 60dB; The medium noise refers to a noise intensity greater than or equal to 60dB and less than 80dB; The strong noise refers to noise intensity greater than or equal to 80dB and less than 100dB.

7. A speech compensation device for a Bluetooth headset, used to compensate for the uplink speech of the Bluetooth headset in at least one predetermined noise environment, characterized in that: It includes a speech amplitude compensation value calculation module, a speech compensation spectrum storage module, a speech compensation spectrum selection module, a noise reduction module and a compensation module, wherein: The speech amplitude compensation value calculation module is used to calculate the speech amplitude difference between a predetermined first speech signal and a second speech signal at a plurality of preset frequency points, wherein the first speech signal is a signal of a pure speech signal after noise reduction processing, and the second speech signal is a signal of a pure speech signal after noise reduction processing in the predetermined noise environment; The speech compensation spectrum storage module is used to pre-store at least one speech compensation spectrum, each speech compensation spectrum corresponds to a predetermined noise environment, and each speech compensation spectrum includes speech amplitude compensation values ​​at multiple preset frequency points in the corresponding predetermined noise environment, and the multiple preset frequency points are frequency points distributed within the voice call bandwidth of the Bluetooth headset; The speech compensation spectrum selection module is used to determine whether the current noise environment belongs to one of the predetermined noise environments. If so, the speech compensation spectrum in the predetermined noise environment corresponding to the current noise environment is selected in the speech compensation spectrum storage module as the current speech compensation spectrum. When there is no uplink audio signal, it is determined whether the noise environment belongs to one of the predetermined noise environments. If so, the speech compensation spectrum in the predetermined noise environment corresponding to the noise environment is updated as the current speech compensation spectrum. The noise reduction module includes a deep noise reduction unit, which is used to perform noise reduction processing on the current uplink audio signal using a deep noise reduction algorithm to obtain a current noise-reduced speech signal; The compensation module is used to superimpose the speech amplitude compensation value corresponding to each target frequency point in the current speech compensation spectrum on the basis of the current noise reduction speech signal, and the target frequency point is a frequency point included in the frequency band of the current noise reduction speech signal and belongs to a subset of the multiple preset frequency points.

8. The voice compensation device for a Bluetooth headset according to claim 7, characterized in that: It also includes a voice amplitude compensation value calculation module, which is used to calculate the voice amplitude difference between the first voice signal and the second voice signal at the plurality of preset frequency points. The first speech signal is a signal of a pure speech signal after the noise reduction processing; The second speech signal is a signal obtained by subjecting the pure speech signal to the noise reduction processing in the predetermined noise environment.

9. The voice compensation device for a Bluetooth headset according to claim 7, characterized in that: The speech compensation spectrum selection module includes a noise amplitude spectrum storage module, a noise amplitude spectrum calculation unit, a noise amplitude difference calculation unit and a noise environment judgment unit, wherein: The noise amplitude spectrum storage module is used to pre-store at least two noise amplitude spectrums, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes noise amplitudes at multiple preset frequency points in the corresponding predetermined noise environment; The noise amplitude spectrum calculation unit is used to calculate the current environmental noise amplitude spectrum, and the current environmental noise amplitude spectrum includes the amplitude of the current environmental noise at the plurality of preset frequency points; The noise amplitude difference calculation unit is used to respectively calculate the amplitude differences between the current environmental noise amplitude spectrum and the noise amplitude spectrum of each predetermined noise environment at corresponding frequency points, and respectively calculate the accumulated value of each amplitude difference corresponding to each predetermined noise environment; The noise environment judgment unit is used to determine the predetermined noise environment corresponding to the minimum accumulated value as the predetermined noise environment to which the current noise environment belongs.

10. The voice compensation device for a Bluetooth headset according to claim 7, characterized in that: The speech compensation spectrum selection module includes a noise amplitude spectrum storage module, a noise amplitude spectrum calculation unit, a noise amplitude difference calculation unit and a noise environment judgment unit, wherein: The noise amplitude spectrum storage module is used to pre-store at least one noise amplitude spectrum, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes noise amplitudes at multiple preset frequency points in the corresponding predetermined noise environment; The noise amplitude spectrum calculation unit is used to calculate the current environmental noise amplitude spectrum, and the current environmental noise amplitude spectrum includes the amplitude of the current environmental noise at the plurality of preset frequency points; The noise amplitude difference calculation unit is used to respectively calculate the amplitude differences between the current environmental noise amplitude spectrum and the noise amplitude spectrum of each predetermined noise environment at corresponding frequency points, and respectively calculate the accumulated value of each amplitude difference corresponding to each predetermined noise environment; The noise environment judgment unit is used to determine the predetermined noise environment corresponding to any accumulated value as the predetermined noise environment to which the current noise environment belongs when any accumulated value is less than or equal to a preset value.

11. The voice compensation device for a Bluetooth headset according to claim 7, characterized in that: The speech compensation spectrum selection module includes a noise amplitude spectrum storage module, a noise intensity calculation unit, a noise intensity level judgment unit, a noise amplitude spectrum calculation unit, a noise amplitude difference calculation unit and a noise environment judgment unit, wherein: The noise amplitude spectrum storage module is used to pre-store at least two noise amplitude spectrums, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes noise amplitudes at multiple preset frequency points in the corresponding predetermined noise environment; The noise intensity calculation unit is used to classify the pre-stored noise amplitude spectrum according to the noise intensity of the corresponding noise environment, and calculate the noise intensity of the current environmental noise to obtain the current noise intensity; The noise intensity level determination unit is used to determine whether the current noise intensity is within a preset noise intensity level range; The noise amplitude spectrum calculation unit is used to calculate the current environmental noise amplitude spectrum when the noise intensity level judgment unit judges that it is yes, and the current environmental noise amplitude spectrum includes the amplitude of the current environmental noise at the plurality of preset frequency points; The noise amplitude difference calculation unit is used to respectively calculate the amplitude difference between the current environmental noise amplitude spectrum and each pre-stored noise amplitude spectrum within the level range where the current noise intensity is located at the corresponding frequency point, and respectively calculate the cumulative value of each amplitude difference corresponding to each pre-stored noise amplitude spectrum within the level range where the current noise intensity is located; The noise environment judgment unit is used to determine the predetermined noise environment corresponding to the minimum accumulated value as the predetermined noise environment to which the current noise environment belongs.

12. The voice compensation device for a Bluetooth headset according to claim 7, characterized in that: The speech compensation spectrum selection module includes a noise amplitude spectrum storage module, a noise intensity calculation unit, a noise intensity level judgment unit, a noise amplitude spectrum calculation unit, a noise amplitude difference calculation unit and a noise environment judgment unit, wherein: The noise amplitude spectrum storage module is also used to pre-store at least one noise amplitude spectrum, each noise amplitude spectrum corresponds to a predetermined noise environment, and the noise amplitude spectrum includes noise amplitudes at multiple preset frequency points in the corresponding predetermined noise environment; The noise intensity calculation unit is used to classify the pre-stored noise amplitude spectrum according to the noise intensity of the corresponding noise environment, and calculate the noise intensity of the current environmental noise to obtain the current noise intensity; The noise intensity level determination unit is used to determine whether the current noise intensity is within a preset noise intensity level range; The noise amplitude spectrum calculation unit is used to calculate the current environment noise amplitude spectrum when the noise environment judgment unit judges that it is yes, and the current environment noise amplitude spectrum includes the amplitude of the current environment noise at the plurality of preset frequency points; The noise amplitude difference calculation unit is used to respectively calculate the amplitude differences between the current environmental noise amplitude spectrum and the pre-stored noise amplitude spectrum within the level range where the current noise intensity is located at the corresponding frequency points, and respectively calculate the accumulated values ​​of the amplitude differences corresponding to each pre-stored noise amplitude spectrum within the level range where the current noise intensity is located; The noise environment judgment unit is used to determine the predetermined noise environment corresponding to any accumulated value as the predetermined noise environment to which the current noise environment belongs when any accumulated value is less than or equal to a preset value.

13. The speech compensation device for a Bluetooth headset according to claim 11 or 12, characterized in that: The preset noise intensity levels include weak noise, medium noise and strong noise. The weak noise refers to a noise intensity less than 60dB; The medium noise refers to a noise intensity greater than or equal to 60dB and less than 80dB; The strong noise refers to a noise intensity greater than or equal to 80dB and less than 100dB.

14. A Bluetooth headset, characterized in that: It comprises a speech compensation device as described in any one of claims 7 to 13.

15. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed, it can implement the method according to any one of claims 1 to 6.

16. An audio processing chip, comprising a processor and a memory, characterized in that: The memory stores a computer program, and the processor can execute the computer program to implement the method according to any one of claims 1 to 6.

Images