A noise reduction method, device, terminal and storage medium

Abstract

The application is suitable for the field of audio processing technology, and provides a noise reduction method and device, a terminal and a storage medium. The noise reduction method comprises the following steps: obtaining an audio signal collected from ambient sound; obtaining amplitudes corresponding to a plurality of preset frequency points of a spectrum of the audio signal in a full frequency band, and averaging the amplitudes corresponding to the plurality of preset frequency points to obtain a full-band amplitude mean value corresponding to the spectrum of the audio signal in the full frequency band; according to different ranges of the full-band amplitude mean value, confirming a category of the ambient sound as one of quiet noise, intermediate noise and noisy noise, and correspondingly adopting a comfortable noise reduction mode, an equalization noise reduction mode or a deep noise reduction mode to reduce the audio signal. The embodiments of the application can reduce the operation power consumption while improving the noise reduction effect.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of audio processing, and particularly relates to a noise reduction method and device, a terminal and a storage medium. BACKGROUND

[0002] The mainstream algorithm for current noise scene recognition is a speech recognition algorithm based on a deep neural network and a hidden Markov model (DNN-HMM). The algorithm performs windowing and framing on collected environmental audio signals, extracts and analyzes spectral features, determines a scene type, and switches to a noise reduction mode matching the current scene to perform noise reduction, thereby meeting noise reduction requirements in different scenes.

[0003] Due to problems such as long running time and large power consumption of the DNN-HMM scene recognition algorithm applied to earphones, earphones in the actual market often use the same noise reduction mode for active noise reduction. Due to the use of fixed noise reduction filter banks, the noise reduction effect is different in different scenes, and the noise reduction effect is significantly poor in some scenes

[0004] Therefore, a noise reduction method is needed that can reduce computational power consumption while improving noise reduction effect. SUMMARY

[0005] Embodiments of the present application provide a noise reduction method, device, terminal and storage medium that can reduce computational power consumption while improving noise reduction effect.

[0006] A first aspect of embodiments of the present application provides a noise reduction method, comprising:

[0007] obtaining an audio signal obtained by collecting environmental sound;

[0008] obtaining amplitudes corresponding to a plurality of preset frequency points of a spectrum of the audio signal in a full frequency band, and averaging the amplitudes corresponding to the plurality of preset frequency points to obtain a full frequency band amplitude mean value corresponding to the spectrum of the audio signal in the full frequency band;

[0009] if the full frequency band amplitude mean value is less than or equal to a first threshold value, the category of the environmental sound is confirmed as quiet noise, and a comfortable noise reduction mode is used to reduce the audio signal;

[0010] if the full frequency band amplitude mean value is greater than the first threshold value and less than or equal to a second threshold value, the category of the environmental sound is confirmed as intermediate noise, and an equalization noise reduction mode is used to reduce the audio signal;

[0011] if the full frequency band amplitude mean value is greater than the second threshold value, the category of the environmental sound is confirmed as noisy noise, and a deep noise reduction mode is used to reduce the audio signal.

[0012] The second aspect of the embodiment of the present application provides a noise reduction device, comprising:

[0013] An acquisition unit is configured to acquire an audio signal collected from an ambient sound;

[0014] A calculation unit is configured to acquire amplitudes corresponding to a plurality of preset frequency points in a full frequency band of a spectrum of the audio signal, and average the amplitudes corresponding to the plurality of preset frequency points to obtain a full frequency band amplitude mean value corresponding to the full frequency band of the spectrum of the audio signal.

[0015] A comfortable noise reduction unit is configured to confirm a category of the ambient sound as quiet noise if the full frequency band amplitude mean value is less than or equal to a first threshold value, and to reduce the audio signal in a comfortable noise reduction mode.

[0016] An equalization noise reduction unit is configured to confirm the category of the ambient sound as intermediate noise if the full frequency band amplitude mean value is greater than the first threshold value and less than or equal to a second threshold value, and to reduce the audio signal in an equalization noise reduction mode.

[0017] A deep noise reduction unit is configured to confirm the category of the ambient sound as noisy noise if the full frequency band amplitude mean value is greater than the second threshold value, and to reduce the audio signal in a deep noise reduction mode.

[0018] The third aspect of the embodiment of the present application provides a terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the above method when executing the computer program.

[0019] The fourth aspect of the embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the steps of the above method.

[0020] The fifth aspect of the embodiment of the present application provides a computer program product, which, when executed on a terminal, enables the terminal to implement the steps of the method.

[0021] In the embodiments of the present application, the audio signal collected by collecting the ambient sound is obtained, and the amplitudes corresponding to a plurality of preset frequency points of the full frequency band of the spectrum of the audio signal are obtained. Then, the amplitudes corresponding to the plurality of preset frequency points are averaged to obtain the full-band amplitude mean value corresponding to the spectrum of the audio signal. Then, according to the range where the full-band amplitude mean value is located, the category of the ambient sound is confirmed as one of quiet noise, intermediate noise and noisy noise, and the audio signal is respectively denoised by using the comfortable noise reduction mode, the balanced noise reduction mode or the deep noise reduction mode. On the one hand, unlike the DNN-HMM scene recognition algorithm, the audio signal does not need to be processed in a complex manner, the operation power consumption of the noise reduction process can be reduced, and the noise reduction efficiency can be improved. On the other hand, different noise reduction modes can be compatible with different scene ambient sounds, and the noise reduction effect is improved. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0023] Figure 1 is an implementation flow diagram of a noise reduction method provided by an embodiment of the present application;

[0024] Figure 2 is a time domain and frequency domain diagram of an office scene ambient sound provided by an embodiment of the present application;

[0025] Figure 3 is a time domain and frequency domain diagram of an in-vehicle scene ambient sound provided by an embodiment of the present application;

[0026] Figure 4 is a time domain and frequency domain diagram of a kindergarten scene ambient sound provided by an embodiment of the present application;

[0027] Figure 5 is a time domain and frequency domain diagram of a road scene ambient sound provided by an embodiment of the present application;

[0028] Figure 6 is a time domain and frequency domain diagram of a field scene ambient sound provided by an embodiment of the present application;

[0029] Figure 7 is a time domain and frequency domain diagram of a drilling scene ambient sound provided by an embodiment of the present application;

[0030] Figure 8 is an amplitude analysis result diagram of scheme one provided by an embodiment of the present application;

[0031] Figure 9is a magnitude analysis result graph of scheme two provided by an embodiment of the present application;

[0032] Figure 10 is a mean square error MES analysis result graph provided by an embodiment of the present application;

[0033] Figure 11 is a magnitude analysis result graph at the same energy level provided by an embodiment of the present application;

[0034] Figure 12 is a structural schematic diagram of a noise reduction device provided by an embodiment of the present application;

[0035] Figure 13 is a structural schematic diagram of a terminal provided by an embodiment of the present application. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor belong to the protection of the present application.

[0037] Since the earphone applies the DNN-HMM scene recognition algorithm, there are problems such as long running time and large power consumption, and the earphone in the actual market often uses the same noise reduction mode for active noise reduction. Since the noise reduction filter set is fixed, the noise reduction effect is different in different scenes, and the noise reduction effect is obviously poor in some scenes

[0038] Therefore, the present application proposes a noise reduction method, which can determine the environmental sound scene category based on the average value of the environmental sound frequency spectrum amplitude, and take the corresponding noise reduction mode for noise reduction. It does not need to perform complex audio signal processing like the DNN-HMM scene recognition algorithm, reduces the operation power consumption of the noise reduction process, improves the noise reduction efficiency, and can be compatible with different noise reduction modes to improve the noise reduction effect.

[0039] In order to illustrate the technical solutions of the present application, the following will be described through specific embodiments.

[0040] Figure 1 An implementation flow schematic diagram of a noise reduction method provided by an embodiment of the present application is shown, which can be applied to a terminal and can be applicable to the case of reducing operation power consumption while improving noise reduction effect.

[0041] In some embodiments of the present application, the above terminal can be an earphone, a sound, a mobile phone, a smart watch and the like, which is a device with audio processing capability, or a smart device connected with a device with audio processing capability.

[0042] Specifically, the above noise reduction method can include the following steps S101 to S105.

[0043] In step S101, an audio signal collected from environmental sound is obtained.

[0044] In the embodiments of the present application, the terminal can pick up the sound of the environment through the microphone installed on the terminal to obtain the above-mentioned audio signal.

[0045] Specifically, the terminal can start to obtain the audio signal and reduce the noise after meeting certain conditions. For example, the terminal can obtain the audio signal collected from the environmental sound and reduce the noise of the environmental sound in response to the terminal entering a call state, so that the user of the terminal can have a clearer call effect. For another example, the terminal can obtain the audio signal collected from the environmental sound and reduce the noise of the environmental sound in response to the terminal entering an audio playing state, so that the user of the terminal can have a clearer audio listening effect.

[0046] In step S102, the amplitudes of the spectrum of the audio signal corresponding to a plurality of preset frequency points in the full frequency band are obtained, and the amplitudes corresponding to the plurality of preset frequency points are averaged to obtain a full-band amplitude mean value corresponding to the spectrum of the audio signal in the full frequency band.

[0047] In the embodiments of the present application, the terminal can analyze the audio signal to obtain the spectrum of the audio signal. The amplitudes of the spectrum of the audio signal corresponding to a plurality of preset frequency points in the full frequency band are extracted, and the amplitudes corresponding to the plurality of preset frequency points are averaged to obtain the above-mentioned full-band amplitude mean value. The obtained full-band amplitude mean value can reflect the overall level of the environmental sound signal energy.

[0048] Wherein, the unit of the spectrum amplitude is the sound pressure, which is a unit for describing the sound intensity, and the unit is dB SPL.

[0049] In the embodiments of the present application, the frequency range of the full frequency band and the preset frequency points selected on the full frequency band can be set according to the actual situation. The selected frequency points should be able to reflect the amplitude change of the spectrum of each frequency band.

[0050] In some embodiments of the present application, the terminal can divide the full frequency band into a plurality of sub-frequency bands according to the set frequency band range, and select a plurality of frequency points as preset frequency points on each sub-frequency band. At this time, the spectrum amplitudes A1, A1,..., A N Full-band amplitude mean value Wherein, N is the number of preset frequency points.

[0051] Specifically, the terminal can set the frequency range of 50 Hz to 5 KHz as the frequency range corresponding to the full frequency band, divide the full frequency band into a low frequency band of 50 Hz to 250 Hz, a medium frequency band of 250 Hz to 1 KHz, and a high frequency band of 1 KHz to 5 KHz, select a plurality of frequency points as preset frequency points on each sub frequency band, obtain the amplitude corresponding to each preset frequency point, and calculate the full frequency band amplitude mean value

[0052] In some embodiments of the present application, the terminal can select at least two frequency points in the high frequency band, the medium frequency band, and the low frequency band, that is, the plurality of preset frequency points include at least two frequency points in the high frequency band, at least two frequency points in the medium frequency band, and at least two frequency points in the low frequency band, so that there are at least two frequency points representing the characteristics of the high frequency band, the medium frequency band, and the low frequency band.

[0053] Step S103, if the full frequency band amplitude mean value is less than or equal to the first threshold value, the category of the environmental sound is confirmed as quiet noise, and the audio signal is denoised using the comfortable noise reduction mode.

[0054] Step S104, if the full frequency band amplitude mean value is greater than the first threshold value and less than or equal to the second threshold value, the category of the environmental sound is confirmed as intermediate noise, and the audio signal is denoised using the balanced noise reduction mode.

[0055] Step S105, if the full frequency band amplitude mean value is greater than the second threshold value, the category of the environmental sound is confirmed as noisy noise, and the audio signal is denoised using the deep noise reduction mode.

[0056] In embodiments of the present application, the terminal can set a plurality of threshold values according to experience or experimental results, determine the range of the full frequency band amplitude mean value according to the threshold values, determine the category of the environmental sound, and denoise the audio signal using the noise reduction mode corresponding to the category of the environmental sound.

[0057] Specifically, in some embodiments of the present application, if the full frequency band amplitude mean value is less than or equal to the first threshold value TH_Low, indicating that the overall energy of the environmental sound signal is at a low level, the terminal can confirm the category of the environmental sound as quiet noise, and denoise the audio signal using the comfortable noise reduction mode. At this time, the category of the environmental sound can be specifically the environmental sound of the office scene and the environmental sound of the field scene.

[0058] If the full frequency band amplitude mean value greater than the first threshold TH_Low and less than or equal to the second threshold TH_High, it indicates that the overall energy of the environmental sound signal is at a medium level, and the terminal can confirm the category of the environmental sound as intermediate noise and adopt an equalization noise reduction mode to reduce the audio signal. At this time, the category of the environmental sound can be specifically the environmental sound of a kindergarten scene.

[0059] If the full-band amplitude mean value is greater than the second threshold TH_High, it indicates that the overall energy of the environmental sound signal is at a higher level, and the terminal can confirm the category of the environmental sound as noisy noise and adopt a deep noise reduction mode to reduce the audio signal. At this time, the category of the environmental sound can be specifically the environmental sound of an in-vehicle scene, the environmental sound of a road scene, and the environmental sound of a drilling scene.

[0060] Different noise reduction modes correspond to different noise reduction filter banks, and different noise reduction filter banks can be used to generate a destructive interference signal that cancels noise in the environmental sound. The energy of the corresponding destructive interference signal in the equalization noise reduction mode is greater than the energy of the corresponding destructive interference signal in the comfort noise reduction mode and less than the energy of the corresponding destructive interference signal in the deep noise reduction mode.

[0061] In order to further subdivide the category of the environmental sound, the terminal can divide the above full-band into a plurality of sub-bands whose frequencies do not overlap, and each sub-band contains one or more preset sub-frequency points. The preset sub-frequency points can be reselected frequency points, or the preset sub-frequency points obtained by dividing the aforementioned preset frequency points according to the sub-bands.

[0062] Correspondingly, the process of the above noise reduction method can further include: obtaining the amplitude corresponding to each preset sub-frequency point, and averaging the amplitudes corresponding to the preset sub-frequency points in each sub-band to obtain the sub-band amplitude mean value corresponding to the frequency spectrum of the audio signal in each sub-band; determining the category of the environmental sound according to the full-band amplitude mean value and the sub-band amplitude mean value corresponding to each sub-band, and adopting a noise reduction mode corresponding to the category of the environmental sound to reduce the audio signal.

[0063] Specifically, the full-band can be divided into a high-frequency band, a medium-frequency band and a low-frequency band. The high-frequency band contains one or more high-frequency points, the medium-frequency band contains one or more medium-frequency points, and the low-frequency band contains one or more low-frequency points.

[0064] The low-frequency point spectrum amplitude A 1低频 , A 2低频 ,..., A N低频 is averaged to obtain a low-frequency mean value , where N 低频 is the number of low-frequency points. If the low-frequency mean value is lower than a preset threshold TH_Low​低频 When , it indicates that the low-frequency energy characteristics of the ambient sound are weak; if the average Above the threshold TH_High 低频 When , it indicates that the low-frequency energy characteristics of the ambient sound are obvious. At this time, the ambient sound with mainly low-frequency energy can be distinguished.

[0065] The spectrum amplitude A of the mid-frequency point 1中频 、A 2中频 ,...,A N中频 The average value of the intermediate frequency can be obtained by averaging Among them, N 中频 is the number of intermediate frequency points. If the intermediate frequency mean Lower than the preset threshold TH_Low 中频 When , it indicates that the energy characteristics of the mid-frequency part of the ambient sound are weak; if the average Above the threshold TH_High 中频 When , it indicates that the energy characteristics of the mid-frequency part of the ambient sound are obvious. At this time, the ambient sound with energy mainly in the mid-frequency range can be distinguished. Combined with the above low-frequency points, the ambient sound with energy mainly in the mid-low frequency range can be distinguished.

[0066] For the high frequency spectrum amplitude A 1高频 、A 2高频 ,...,A N高频 The high frequency mean can be obtained by averaging Among them, N 高频 is the number of high-frequency points. If the high-frequency mean Lower than the preset threshold TH_Low 高频 When , it indicates that the energy characteristics of the high-frequency part of the ambient sound are weak; if the average Above the threshold TH_High 高频 When , it indicates that the high-frequency energy characteristics of the ambient sound are obvious. At this time, the ambient sound with mainly high-frequency energy can be distinguished.

[0067] Specifically, if the full-band amplitude mean Greater than the second threshold TH_High and meets the low frequency mean Less than or equal to the fifth threshold TH_High 低频 and high frequency mean Less than or equal to the sixth threshold TH_High 高频 If at least one of the following conditions is met, the terminal can identify the category of the ambient sound as the first noisy noise and use the first deep noise reduction mode to reduce the noise of the audio signal. At this time, the category of the ambient sound can specifically be the ambient sound of the in-car scene or the ambient sound of the road scene.

[0068] If the full-band amplitude average Greater than the second threshold TH_High, low frequency mean greater than the fifth threshold TH_High 低频 , and the high-frequency mean greater than the sixth threshold TH_High 高频 , the terminal can confirm the category of the environmental sound as the second noisy noise, and adopt the second deep noise reduction mode to reduce the noise of the audio signal. At this time, the category of the environmental sound can be specifically the environmental sound of the drilling scene.

[0069] It should be noted that the noise reduction filter set adopted by the first deep noise reduction mode and the second deep noise reduction mode can be the same or different.

[0070] In actual application, since the overall level of the environmental sound energy in some scenes can be at the critical value of a certain range, that is, the mean of the amplitudes of the full-band frequency spectrum is close to the first threshold TH_Low or the second threshold TH_High. In order to make the classification more accurate, in some embodiments of the present application, the terminal can also calculate the mean square error of the amplitudes corresponding to the multiple preset frequency points of the frequency spectrum of the audio signal in the full-band, and determine the category of the environmental sound according to the mean of the amplitudes of the full-band, the mean of the amplitudes of each sub-band, and the mean square error, and then adopt the noise reduction mode corresponding to the category of the environmental sound to reduce the noise of the audio signal.

[0071] Specifically, the terminal can use the amplitudes A1, A1,..., A N , to obtain the mean square error MSE to evaluate the degree of dispersion, wherein the mean square error If the mean square error MSE is lower than the fourth threshold THL_MSE, it indicates that the corresponding energy between each frequency point of the environmental sound spectrum is relatively balanced, for example, the environmental sound is mixed with noise mainly composed of human voice and low-frequency, and the energy of each frequency band is evenly distributed. If the mean square error MSE is greater than or equal to the fourth threshold THL_MSE, it indicates that the corresponding energy between each frequency point of the environmental sound spectrum is relatively large, for example, the environmental sound is mainly composed of low-frequency noise.

[0072] Specifically, if the mean of the amplitudes of the full-band is greater than the first threshold TH_Low and less than or equal to the second threshold TH_High, the absolute value of the difference between the mean of the amplitudes of each sub-band is less than the third threshold, and the mean square error MSE is less than the fourth threshold THL_MSE, the category of the environmental sound is confirmed as the first intermediate noise, and the first balanced noise reduction mode is adopted to reduce the noise of the audio signal.

[0073] The absolute value of the difference between the mean of the amplitudes of each sub-band can refer to the absolute value of the difference between the high-frequency mean, the medium-frequency mean and the low-frequency mean, and when they are all less than the third threshold, it indicates that the values of the high-frequency mean, the medium-frequency mean and the low-frequency mean are close.

[0074] It should be noted that the specific values of the third threshold, the fourth threshold, the fifth threshold, the sixth threshold and the like can be set according to experience or experimental results.

[0075] In some embodiments of the present application, the energy of the environmental sound in the same scene may be different at different times. The terminal can collect audio signals at a preset collection frequency, and select a corresponding noise reduction mode for noise reduction in real time according to the average value of the amplitude of the spectrum corresponding to the current environmental sound.

[0076] In the embodiments of the present application, the audio signal collected by collecting the environmental sound is obtained, and the amplitudes corresponding to a plurality of preset frequency points of the spectrum of the audio signal in the full frequency band are obtained. Then, the amplitudes corresponding to the plurality of preset frequency points are averaged to obtain the average value of the full frequency band amplitude corresponding to the spectrum of the audio signal. Then, according to the range where the average value of the full frequency band amplitude is located, the category of the environmental sound is confirmed as one of quiet noise, intermediate noise and noisy noise, and the comfortable noise reduction mode, the balanced noise reduction mode or the deep noise reduction mode is correspondingly adopted for noise reduction. On the one hand, unlike the DNN-HMM scene recognition algorithm, it does not need to perform complex audio signal processing, can reduce the operation power consumption of the noise reduction process, and improve the noise reduction efficiency. On the other hand, it can be compatible with different noise reduction modes for different scene environmental sounds, and improve the noise reduction effect.

[0077] In order to illustrate the effect achieved by the embodiments of the present application, two sets of schemes are provided for experimental verification.

[0078] First, audio signals of six types of scenes are collected respectively, and the time domain and frequency domain graphs of the environmental sound in each type of scene are analyzed. Figures 2 to 7 The time domain and frequency domain graphs of the environmental sound in the office scene, the in-vehicle scene, the kindergarten scene, the road scene, the field scene and the drilling scene are respectively analyzed.

[0079] From the above experimental results, it can be seen that the average value of the full frequency band amplitude corresponding to the spectrum of the audio signal can be used to distinguish the category of the environmental sound, and the comfortable noise reduction mode, the balanced noise reduction mode or the deep noise reduction mode can be correspondingly adopted for noise reduction. Figures 2 to 7It can be seen that the spectrum energy of the office scene is small, and the energy is mainly concentrated in the low and medium frequency below 600 Hz, and the spectrum feature can be represented by the spectrum feature of the frequency points selected uniformly in the range of 50 Hz to 600 Hz. The spectrum energy of the in-vehicle scene is mainly concentrated in the low frequency below 250 Hz, and the spectrum feature can be represented by the spectrum feature of the frequency points selected uniformly in the range of 50 Hz to 250 Hz. The spectrum band of the kindergarten scene is wide, and there is a certain uniform distribution below 3 kHz, mainly mixed with different human voices and certain background noise, and the spectrum feature can be represented by the spectrum feature of the frequency points selected uniformly in the range of 50 Hz to 3 kHz. The spectrum energy of the road scene is mainly concentrated below 300 Hz, but the spectrum energy is lifted near 3 kHz, which is affected by the friction sound and whistle sound generated by the motor vehicle, and the spectrum feature can be represented by the spectrum feature of the frequency points selected uniformly in each frequency band in the range of 50 Hz to 3 kHz. The spectrum energy of the field scene is weak, mainly distributed in the low frequency interval below 200 Hz, and the spectrum feature can be represented by the spectrum feature of the frequency points selected uniformly in the range of 50 Hz to 200 Hz. The composition of the environment sound of the drilling scene is complex, and the spectrum energy is mainly concentrated in the low frequency part below 300 Hz, and the spectrum energy of 3 kHz and 5 kHz is large, and the spectrum feature can be represented by the spectrum feature of the frequency points selected uniformly in each frequency band in the range of 50 Hz to 5 kHz.

[0080] Scheme one: the terminal selects the frequency points of 70 Hz, 100 Hz, 150 Hz, 200 Hz, 400 Hz, 500 Hz, 600 Hz, 1000 Hz, 2000 Hz and 3000 Hz as preset frequency points, wherein the low frequency points on the low frequency band are 70 Hz, 100 Hz, 150 Hz and 200 Hz respectively; the medium frequency points on the medium frequency band are 400 Hz, 500 Hz and 600 Hz respectively; and the high frequency points on the high frequency band are 1000 Hz, 2000 Hz and 3000 Hz respectively. The full-band amplitude mean value, high frequency mean value, medium frequency mean value, low frequency mean value and mean square deviation are calculated respectively to obtain Table 1 and the corresponding Figure 8 .

[0081] Figure 8 Among them, type 1 represents the office scene, type 2 represents the in-vehicle scene, type 3 represents the kindergarten scene, type 4 represents the road scene, type 6 represents the field scene, and type 7 represents the drilling scene.

[0082] Table 1: Scheme one amplitude analysis result statistical table

[0083]

[0084] At this time, if the full-band amplitude mean value If the ambient sound level is less than or equal to the first threshold TH_Low, the terminal may determine the category of the ambient sound as quiet noise and use the comfortable noise reduction mode to reduce the noise of the audio signal. In this case, the category of the ambient sound may specifically be the ambient sound of an office scene or the ambient sound of a field scene.

[0085] If the full-band amplitude average If the value of the difference between the sub-band amplitude means corresponding to each sub-band is greater than the first threshold TH_Low and less than or equal to the second threshold TH_High; or if the absolute value of the difference between the sub-band amplitude means corresponding to each sub-band is less than the third threshold, and the mean square error (MSE) is less than the fourth threshold THL_MSE, the terminal can determine the category of the ambient sound as intermediate noise and use the balanced noise reduction mode to reduce noise on the audio signal. In this case, the category of the ambient sound can specifically be the ambient sound of a kindergarten scene.

[0086] If the full-band amplitude average If the ambient sound level is greater than the second threshold TH_High, the terminal can identify the ambient sound as noisy noise and use the deep noise reduction mode to reduce the noise of the audio signal. In this case, the ambient sound level can be specifically classified as the ambient sound of the in-car scene, the ambient sound of the road scene, and the ambient sound of the drilling scene.

[0087] It can be seen from the above data that the solution provided by this application can well distinguish three different scenarios: quiet noise, intermediate noise and noisy noise.

[0088] Solution 2: The terminal selects frequencies of 70Hz, 100Hz, 150Hz, 200Hz, 400Hz, 500Hz, 600Hz, 1000Hz, 2000Hz, 3000Hz, 4000Hz, and 5000Hz as preset frequencies, where the low-frequency points on the low-frequency band are 70Hz, 100Hz, 150Hz, and 200Hz; the intermediate-frequency points on the intermediate-frequency band are 400Hz, 500Hz, and 600Hz; and the high-frequency points on the high-frequency band are 1000Hz, 2000Hz, 3000Hz, 4000Hz, and 5000Hz. Calculating the full-band amplitude mean, high-frequency mean, intermediate-frequency mean, low-frequency mean, and mean square error respectively can yield Table 2 and the corresponding values ​​in Table 2. Figure 9 .

[0089] and Figure 8 same, Figures 9 to 11 Among them, type 1 represents office scenes, type 2 represents in-car scenes, type 3 represents kindergarten scenes, type 4 represents road scenes, type 6 represents field scenes, and type 7 represents drilling scenes.

[0090] Table 2 Statistics of amplitude analysis results of Scheme 2

[0091]

[0092] At this time, if the full-band amplitude mean value is less than or equal to the first threshold TH_Low, the terminal can confirm the category of the environmental sound as quiet noise, and adopt the comfortable noise reduction mode to reduce the audio signal. At this time, the category of the environmental sound can be specifically the environmental sound of the office scene and the environmental sound of the field scene.

[0093] If the full-band amplitude mean value is greater than the first threshold TH_Low and less than or equal to the second threshold TH_High; or on this basis, the absolute value of the difference between the sub-band amplitude mean values corresponding to each sub-band is less than the third threshold, and the mean square error MSE is less than the fourth threshold THL_MSE, the terminal can confirm the category of the environmental sound as intermediate noise, and adopt the balanced noise reduction mode to reduce the audio signal. At this time, the category of the environmental sound can be specifically the environmental sound of the kindergarten scene.

[0094] If the full-band amplitude mean value is greater than the second threshold TH_High, and at least one of the low-frequency mean value is less than or equal to the fifth threshold TH_High 低频 and the high-frequency mean value is less than or equal to the sixth threshold TH_High 高频 , the terminal can confirm the category of the environmental sound as the first noisy noise, and adopt the first deep noise reduction mode to reduce the audio signal. At this time, the category of the environmental sound can be specifically the environmental sound of the in-vehicle scene or the environmental sound of the road scene.

[0095] If the full-band amplitude mean value is greater than the second threshold TH_High, the low-frequency mean value is greater than the fifth threshold TH_High 低频 , and the high-frequency mean value is greater than the sixth threshold TH_High 高频 , the terminal can confirm the category of the environmental sound as the second noisy noise, and adopt the second deep noise reduction mode to reduce the audio signal. At this time, the category of the environmental sound can be specifically the environmental sound of the drilling scene.

[0096] In order to better reflect the difference of the mean square error MSE value, Figure 10 the mean square error MSE analysis result graphs calculated by two schemes are shown.

[0097] From the above data, it can be seen that the scheme provided by the present application can well distinguish the four different scenes of quiet noise, intermediate noise, first noisy noise and second noisy noise.

[0098] It should be noted that, as shown in Table 3 and corresponding to Table 3 Figure 11 As shown in the table, with the increase of the scene type, the accuracy of the environmental sound classification can be reduced. For example, when the energy level is consistent, the environmental sound of the in-vehicle scene and the environmental sound of the road scene are similar in amplitude; the environmental sound of the office scene and the environmental sound of the field scene are similar in amplitude.

[0099] Table 3 Amplitude analysis result statistics table at the same energy level

[0100]

[0101] Therefore, in actual application, based on the accuracy, the required number of divided scenes, and the noise reduction efficiency, one of the two schemes can be selected, or the number of preset frequency points can be further increased or reduced, and the way of selecting the preset frequency points can be further analyzed to adapt to different needs.

[0102] It should be noted that, for the foregoing method embodiments, in order to simply describe, they are all expressed as a series of action combinations, but those skilled in the art should know that the application is not limited by the order of the described actions, because according to the application, certain steps can be performed in other orders.

[0103] As Figure 12 Fig. 1 shows a structure schematic diagram of a noise reduction device 1200 provided by an embodiment of the application, and the noise reduction device 1200 is configured on a terminal.

[0104] Specifically, the noise reduction device 1200 can include:

[0105] An acquisition unit 1201 is configured to acquire an audio signal collected from an environmental sound;

[0106] A calculation unit 1202 is configured to acquire amplitudes of a spectrum of the audio signal at a plurality of preset frequency points in a full frequency band, and average the amplitudes at the plurality of preset frequency points to obtain a full frequency band amplitude mean value corresponding to the spectrum of the audio signal in the full frequency band;

[0107] A comfort noise reduction unit 1203 is configured to, if the full frequency band amplitude mean value is less than or equal to a first threshold value, confirm a category of the environmental sound as quiet noise, and perform noise reduction on the audio signal in a comfort noise reduction mode;

[0108] An equalization noise reduction unit 1204 is configured to, if the full frequency band amplitude mean value is greater than the first threshold value and less than or equal to a second threshold value, confirm the category of the environmental sound as intermediate noise, and perform noise reduction on the audio signal in an equalization noise reduction mode;

[0109] The deep noise reduction unit 1205 is configured to, if the full-band amplitude mean value is greater than the second threshold value, confirm the category of the environmental sound as noisy noise, and adopt a deep noise reduction mode to reduce noise of the audio signal.

[0110] In some embodiments of the present application, the full-band can include a plurality of sub-bands with non-overlapping frequencies, and each of the sub-bands can include one or more preset sub-frequency points. The noise reduction device 1200 can be specifically configured to: obtain an amplitude corresponding to each of the preset sub-frequency points, and average the amplitudes corresponding to the preset sub-frequency points in each of the sub-bands to obtain a sub-band amplitude mean value corresponding to each of the sub-bands of the spectrum of the audio signal; and determine the category of the environmental sound according to the full-band amplitude mean value and the sub-band amplitude mean value corresponding to each of the sub-bands, and reduce noise of the audio signal by adopting a noise reduction mode corresponding to the category of the environmental sound.

[0111] In some embodiments of the present application, the noise reduction device 1200 can be specifically configured to: calculate a mean square error of amplitudes corresponding to a plurality of preset frequency points in the full-band of the spectrum of the audio signal; and determine the category of the environmental sound according to the full-band amplitude mean value, the sub-band amplitude mean value corresponding to each of the sub-bands, and the mean square error, and reduce noise of the audio signal by adopting a noise reduction mode corresponding to the category of the environmental sound.

[0112] In some embodiments of the present application, the noise reduction device 1200 can be specifically configured to: if the full-band amplitude mean value is greater than a first threshold value and less than or equal to a second threshold value, absolute values of differences between the sub-band amplitude mean values corresponding to each of the sub-bands are all less than a third threshold value, and the mean square error is less than a fourth threshold value, confirm the category of the environmental sound as first intermediate noise, and reduce noise of the audio signal by adopting a first equalization noise reduction mode.

[0113] In some embodiments of the present application, the sub-bands can include a high-frequency band, a medium-frequency band, and a low-frequency band, the sub-band amplitude mean value corresponding to the high-frequency band is a high-frequency mean value, the sub-band amplitude mean value corresponding to the medium-frequency band is a medium-frequency mean value, and the sub-band amplitude mean value corresponding to the low-frequency band is a low-frequency mean value. The noise reduction device 1200 can be specifically configured to: if the full-band amplitude mean value is greater than a second threshold value, and at least one of the low-frequency mean value being less than or equal to a fifth threshold value and the high-frequency mean value being less than or equal to a sixth threshold value is satisfied, confirm the category of the environmental sound as first noisy noise, and reduce noise of the audio signal by adopting a first deep noise reduction mode.

[0114] In some embodiments of the present application, the noise reduction device 1200 can further be specifically configured to: if the full-band amplitude mean value is greater than the second threshold value, the low-frequency mean value is greater than the fifth threshold value, and the high-frequency mean value is greater than the sixth threshold value, the category of the environmental sound is confirmed as the second noisy noise, and a second deep noise reduction mode is used to reduce the audio signal.

[0115] In some embodiments of the present application, the plurality of preset frequency points can include at least two frequency points of a high-frequency band, at least two frequency points of a medium-frequency band, and at least two frequency points of a low-frequency band.

[0116] It should be noted that, for the convenience and brevity of description, the specific working process of the noise reduction device 1200 can be referred to Figures 1 to 11 the corresponding process of the method, which will not be repeated here.

[0117] As Figure 13 shown, a schematic diagram of a terminal provided by an embodiment of the present application is shown. The terminal 13 can include a processor 130, a memory 131, and a computer program 132 stored in the memory 131 and executable on the processor 130, such as a noise reduction program. The processor 130 implements the steps in each of the above noise reduction method embodiments when executing the computer program 132, such as Figure 1 shown, steps S101 to S105. Alternatively, the processor 130 implements the functions of each module / unit in each of the above device embodiments when executing the computer program 132, such as Figure 12 shown, the acquisition unit 1201, the calculation unit 1202, and the comfortable noise reduction unit 1203, the equalization noise reduction unit 1204, and the deep noise reduction unit 1205.

[0118] The computer program can be divided into one or more modules / units, which are stored in the memory 131 and executed by the processor 130 to complete the present application. The one or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which are used to describe the execution process of the computer program in the terminal.

[0119] For example, the computer program can be divided into: an acquisition unit, a comfort noise reduction unit, an equalization noise reduction unit, and a deep noise reduction unit. The specific functions of each unit are as follows: the acquisition unit is configured to acquire an audio signal collected from ambient sound; the calculation unit is configured to acquire amplitudes corresponding to a plurality of preset frequency points of a spectrum of the audio signal in a full frequency band, and average the amplitudes corresponding to the plurality of preset frequency points to obtain a full-band amplitude mean value corresponding to the spectrum of the audio signal in the full frequency band; the comfort noise reduction unit is configured to, if the full-band amplitude mean value is less than or equal to a first threshold value, confirm a category of the ambient sound as quiet noise, and reduce noise of the audio signal in a comfort noise reduction mode; the equalization noise reduction unit is configured to, if the full-band amplitude mean value is greater than the first threshold value and less than or equal to a second threshold value, confirm the category of the ambient sound as intermediate noise, and reduce noise of the audio signal in an equalization noise reduction mode; and the deep noise reduction unit is configured to, if the full-band amplitude mean value is greater than the second threshold value, confirm the category of the ambient sound as noisy noise, and reduce noise of the audio signal in a deep noise reduction mode.

[0120] The terminal can include, but is not limited to, a processor 130 and a memory 131. Those skilled in the art can understand that, Figure 13 The terminal is only an example and does not constitute a limitation on the terminal, and can include more or fewer components than the illustration, or combine certain components, or different components, for example, the terminal can also include an input / output device, a network access device, a bus, etc.

[0121] The processor 130 can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0122] The memory 131 can be an internal storage unit of the terminal, for example, a hard disk or a memory of the terminal. The memory 131 can also be an external storage device of the terminal, for example, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. equipped on the terminal. Further, the memory 131 can include both the internal storage unit and the external storage device of the terminal. The memory 131 is used to store the computer program and other programs and data required by the terminal. The memory 131 can also be used to temporarily store data that has been output or will be output.

[0123] It should be clearly understood by those skilled in the art that, for the convenience and brevity of description, only the above-mentioned division of each functional unit and module is exemplified, and in actual application, the above-mentioned functions can be completed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit, and the integrated unit can be realized in the form of hardware or software. In addition, the specific names of each functional unit and module are only for easy distinction, and do not limit the protection scope of the present application. The specific working process of the units and modules in the system can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0124] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described or recorded in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0125] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0126] In the embodiments of the present application, it should be understood that the disclosed apparatus / terminal and method can be implemented in other manners. For example, the embodiments of the apparatus / terminal described above are merely schematic, and the division of the modules or units is merely logical function division, and there can be another division manner in actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections between the units can be indirect couplings or communication connections through some interfaces, devices or units, and can be electrical, mechanical or in other forms.

[0127] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, i.e., can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiments.

[0128] In addition, each functional unit in the various embodiments of the present application can be integrated in one processing unit, or each unit can be a physically independent unit, or two or more units can be integrated in one unit. The integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0129] The integrated module / unit, if implemented in the form of a software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, all or part of the flow of the method in the above embodiments can also be implemented by a computer program instructing related hardware to complete, and the computer program can be stored in a computer readable storage medium. When the processor executes the computer program, the steps of each method embodiment described above can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable file or some intermediate form. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the computer readable medium can include appropriate contents according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer readable medium does not include electric carrier signal and telecommunication signal.

[0130] The above-described embodiments are only used to illustrate the technical solutions of the present application, but not limit them; although the present application is described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method of noise reduction, characterized by, The method comprises: obtaining an audio signal collected from ambient sound; obtaining amplitudes corresponding to a plurality of preset frequency points in a full frequency band of a spectrum of the audio signal, and averaging the amplitudes corresponding to the plurality of preset frequency points to obtain a full-band amplitude mean value corresponding to the full frequency band of the spectrum of the audio signal; the full frequency band comprises a plurality of sub-frequency bands that do not overlap in frequency, and each of the sub-frequency bands comprises one or more preset sub-frequency points; if the full-band amplitude mean value is less than or equal to a first threshold value, the category of the ambient sound is confirmed as quiet noise, and a comfortable noise reduction mode is used to reduce the audio signal; if the full-band amplitude mean value is greater than the first threshold value and less than or equal to a second threshold value, the category of the ambient sound is confirmed as intermediate noise, and an equalization noise reduction mode is used to reduce the audio signal; if the full-band amplitude mean value is greater than the second threshold value, the category of the ambient sound is confirmed as noisy noise, and a deep noise reduction mode is used to reduce the audio signal; the noise reduction method further comprises: obtaining amplitudes corresponding to each of the preset sub-frequency points, and averaging the amplitudes corresponding to the preset sub-frequency points in each of the sub-frequency bands to obtain a sub-band amplitude mean value corresponding to each of the sub-frequency bands of the spectrum of the audio signal; determining the category of the ambient sound according to the full-band amplitude mean value and the sub-band amplitude mean value corresponding to each of the sub-frequency bands, and using a noise reduction mode corresponding to the category of the ambient sound to reduce the audio signal; determining the category of the ambient sound according to the full-band amplitude mean value and the sub-band amplitude mean value corresponding to each of the sub-frequency bands, and using a noise reduction mode corresponding to the category of the ambient sound to reduce the audio signal, comprises: calculating a mean square deviation of the amplitudes corresponding to the plurality of preset frequency points in the full frequency band of the spectrum of the audio signal; determining the category of the ambient sound according to the full-band amplitude mean value, the sub-band amplitude mean value corresponding to each of the sub-frequency bands, and the mean square deviation, and using a noise reduction mode corresponding to the category of the ambient sound to reduce the audio signal.

2. The noise reduction method of claim 1, wherein, determining the category of the ambient sound according to the full-band amplitude mean value, the sub-band amplitude mean value corresponding to each of the sub-frequency bands, and the mean square deviation, and using a noise reduction mode corresponding to the category of the ambient sound to reduce the audio signal, comprises: if the full-band amplitude mean value is greater than a first threshold value and less than or equal to a second threshold value, the absolute value of the difference between each of the sub-band amplitude mean values corresponding to each of the sub-frequency bands is less than a third threshold value, and the mean square deviation is less than a fourth threshold value, the category of the ambient sound is confirmed as first intermediate noise, and a first equalization noise reduction mode is used to reduce the audio signal.

3. The noise reduction method of claim 1, wherein, The sub-frequency bands comprise a high-frequency band, a medium-frequency band, and a low-frequency band, the sub-band amplitude mean value corresponding to the high-frequency band is a high-frequency mean value, the sub-band amplitude mean value corresponding to the medium-frequency band is a medium-frequency mean value, and the sub-band amplitude mean value corresponding to the low-frequency band is a low-frequency mean value; The method comprises the following steps: If the full-band amplitude mean value is greater than the second threshold value, and at least one of the low-frequency mean value being less than or equal to the fifth threshold value and the high-frequency mean value being less than or equal to the sixth threshold value is satisfied, the category of the environmental sound is confirmed as a first noisy noise, and the audio signal is subjected to noise reduction in a first deep noise reduction mode.

4. The noise reduction method of claim 3, wherein, The method further comprises the following steps: If the full-band amplitude mean value is greater than the second threshold value, the low-frequency mean value is greater than the fifth threshold value, and the high-frequency mean value is greater than the sixth threshold value, the category of the environmental sound is confirmed as a second noisy noise, and the audio signal is subjected to noise reduction in a second deep noise reduction mode.

5. The noise reduction method according to any one of claims 1 to 4, wherein The plurality of preset frequency points comprise at least two frequency points in a high-frequency band, at least two frequency points in a medium-frequency band, and at least two frequency points in a low-frequency band.

6. A noise reduction device, characterized by The method comprises the following steps: An acquisition unit is configured to acquire an audio signal obtained by collecting an environmental sound; A calculation unit is configured to acquire amplitudes corresponding to a plurality of preset frequency points in a full band of a spectrum of the audio signal, and average the amplitudes corresponding to the plurality of preset frequency points to obtain a full-band amplitude mean value corresponding to the full band of the spectrum of the audio signal; the full band comprises a plurality of sub-bands with non-overlapping frequencies, and each of the sub-bands comprises one or more preset sub-frequency points; A comfortable noise reduction unit is configured to, if the full-band amplitude mean value is less than or equal to a first threshold value, confirm the category of the environmental sound as quiet noise, and subject the audio signal to noise reduction in a comfortable noise reduction mode; An equalization noise reduction unit is configured to, if the full-band amplitude mean value is greater than the first threshold value and less than or equal to a second threshold value, confirm the category of the environmental sound as intermediate noise, and subject the audio signal to noise reduction in an equalization noise reduction mode; A deep noise reduction unit is configured to, if the full-band amplitude mean value is greater than the second threshold value, confirm the category of the environmental sound as noisy noise, and subject the audio signal to noise reduction in a deep noise reduction mode. The noise reduction device can further be specifically configured to acquire amplitudes corresponding to each of the preset sub-frequency points, and average the amplitudes corresponding to the preset sub-frequency points in each of the sub-bands to obtain a sub-band amplitude mean value corresponding to each of the sub-bands of the spectrum of the audio signal. The method further comprises the following steps: The method further comprises the following steps:

7. A terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, the computer-readable storage medium comprising: The computer program, which is executed by a processor, implements the steps of the method according to any one of claims 1 to 5.

Images