Speech noise reduction method, system and computer storage medium based on loudness level

Abstract

The present invention relates to a method, system and computer storage medium for speech noise reduction based on loudness level, the method comprising: identifying a speech signal and an environmental noise signal from a sound signal acquired in real time; respectively acquiring the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the environmental noise signal; respectively obtaining the speech loudness level of the speech signal and the environmental loudness level of the environmental noise signal according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the environmental noise signal; performing noise reduction processing on the sound signal based on the loudness level difference between the environmental noise signal and the speech signal to obtain a call signal. The technical solution provided by the present invention determines the loudness level difference through the sound pressure level and frequency of the speech signal and the environmental noise signal in the sound signal, performs noise reduction processing on the sound signal according to the loudness level difference, and performs speech noise reduction based on a loudness level that is more in line with the human ear perception, so that the noise reduction effect is better.

Description

Technical Field

[0001] The present invention relates to the field of speech noise reduction, and in particular to a speech noise reduction method, system and computer storage medium based on loudness level. Background Art

[0002] Currently, in order to improve the clarity of voice calls, mobile phones are usually equipped with a noise reduction function.

[0003] Existing noise reduction is usually based on the amplitude difference between the voice signal and the ambient noise signal. Although this noise reduction based on the signal amplitude can achieve a certain noise reduction effect, due to the difference in the human ear's perception of sounds of different frequencies, the effect of noise reduction based on the amplitude of the sound signal is not good.

[0004] Therefore, it is necessary to provide a method, system and computer storage medium for speech noise reduction based on loudness level to solve the deficiencies of the prior art. Summary of the invention

[0005] In order to solve the problems in the prior art, the present invention provides a method, system and computer storage medium for speech noise reduction based on loudness level.

[0006] The present application provides a method for reducing speech noise based on loudness level, comprising:

[0007] Identify speech signals and environmental noise signals from sound signals acquired in real time;

[0008] Respectively acquiring the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal;

[0009] According to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal, respectively obtain the speech loudness level of the speech signal and the ambient loudness level of the ambient noise signal;

[0010] The sound signal is subjected to noise reduction processing based on the loudness level difference between the ambient noise signal and the voice signal to obtain a call signal.

[0011] Furthermore, the identifying of the speech signal and the environmental noise signal from the sound signal acquired in real time includes:

[0012] Extracting speech feature information from the sound signal;

[0013] A speech signal and an environmental noise signal are identified according to the speech feature information.

[0014] Further, the obtaining, according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal, a speech loudness level of the speech signal and an ambient loudness level of the ambient noise signal respectively comprises:

[0015] Finding the speech loudness level of the speech signal in an equal loudness curve model according to the sound pressure level and frequency of the speech signal;

[0016] The ambient loudness level of the ambient noise signal is searched in the equal loudness curve model according to the sound pressure level and frequency of the ambient noise signal.

[0017] Furthermore, the method for constructing the equal loudness curve model comprises:

[0018] Obtaining the user's judgment results on the softness and loudness of pure tones of different sound pressure levels and frequencies;

[0019] The soft-loudness judgment result is input into a neural network model trained by equal-loudness curves to obtain the equal-loudness curve model.

[0020] Furthermore, the noise reduction intensity of the noise reduction processing is proportional to the loudness level difference.

[0021] Further, the sound signal is subjected to noise reduction processing based on the loudness level difference between the ambient noise signal and the voice signal to obtain a call signal, including:

[0022] Calculating a loudness level difference between an ambient loudness level of the ambient noise signal and a voice loudness level of the voice signal;

[0023] Filtering out the maximum loudness level difference value among the loudness level difference values;

[0024] The sound signal is subjected to noise reduction processing based on the maximum loudness level difference value to obtain a call signal.

[0025] Further, after respectively acquiring the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal, the method further includes:

[0026] Determining whether the sound signal meets a loudness level noise reduction condition according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal;

[0027] When the sound signal meets the loudness level noise reduction condition, the speech loudness level of the speech signal and the environment loudness level of the environment noise signal are obtained according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the environment noise signal, and the sound signal is denoised based on the difference between the environment loudness level and the speech loudness level to obtain a call signal;

[0028] When the sound signal does not meet the loudness level noise reduction condition, the sound signal is denoised based on the sound pressure level difference between the ambient noise signal and the voice signal to obtain a call signal.

[0029] Further, judging whether the sound signal satisfies a loudness level noise reduction condition according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal includes:

[0030] When the difference between the sound pressure levels of the ambient noise signal and the speech signal is greater than a sound pressure level threshold, or when there is an ambient noise signal with a frequency within a preset frequency range in the ambient noise signal, determining that the sound signal meets the loudness level noise reduction condition;

[0031] When the sound pressure level difference between the ambient noise signal and the speech signal is less than or equal to the sound pressure level threshold, and there is no ambient noise signal with a frequency within a preset frequency range in the ambient noise signal, it is determined that the sound signal does not meet the loudness level noise reduction condition.

[0032] The present invention also provides a loudness level-based speech noise reduction system, comprising:

[0033] A first acquisition module is used to identify a speech signal and an environmental noise signal from a sound signal acquired in real time;

[0034] A second acquisition module, used to respectively acquire the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal;

[0035] A determination module, configured to obtain a speech loudness level of the speech signal and an environmental loudness level of the environmental noise signal according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the environmental noise signal;

[0036] The noise reduction module is used to perform noise reduction processing on the sound signal based on the loudness level difference between the ambient noise signal and the voice signal to obtain a call signal.

[0037] The present invention also provides a computer storage medium, on which a program of a method for speech noise reduction based on loudness level is stored. When the program of the method for speech noise reduction based on loudness level is executed by a processor, the steps of any of the above-mentioned methods for speech noise reduction based on loudness level are implemented.

[0038] The technical solution of the present invention has the following advantages compared with the closest prior art:

[0039] The technical solution provided by the present invention first identifies the voice signal and the environmental noise signal from the sound signal acquired in real time, then respectively acquires the sound pressure level and frequency of the voice signal and the sound pressure level and frequency of the environmental noise signal, obtains the voice loudness level of the voice signal and the environmental loudness level of the environmental noise signal according to the sound pressure level and frequency of the voice signal and the sound pressure level and frequency of the environmental noise signal, and finally performs noise reduction processing on the voice signal based on the loudness level difference between the voice signal and the environmental noise signal to obtain a call signal. The technical solution provided by the present invention determines the loudness level difference between the environmental noise signal and the voice signal through the sound pressure level and frequency of the voice signal and the environmental noise signal in the sound signal, and performs noise reduction processing on the sound signal according to the loudness level difference. The present invention performs voice noise reduction based on a loudness level that is more in line with the human ear perception, so that the noise reduction effect is better. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 is a flow chart of a method for reducing speech noise based on loudness level in an embodiment of the present invention;

[0041] Figure 2 is a schematic diagram of equal loudness curves provided in an embodiment of the present invention;

[0042] Figure 3 Schematic diagram of a loudness level-based speech noise reduction system in an embodiment of the present invention. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] like Figure 1 As shown, the present application provides a method for reducing the noise of a voice based on loudness level, which can realize real-time noise reduction processing of a sound signal during a call. The method for reducing the noise of a voice based on loudness level can include the following steps:

[0045] S100, identifying a speech signal and an environmental noise signal from a sound signal acquired in real time;

[0046] S200, respectively acquiring the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal;

[0047] S300, obtaining a speech loudness level of the speech signal and an environmental loudness level of the environmental noise signal according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the environmental noise signal;

[0048] S400, performing noise reduction processing on the sound signal based on the loudness level difference between the ambient noise signal and the voice signal to obtain a call signal.

[0049] In an embodiment of the present application, a speech signal and an ambient noise signal are first identified from a sound signal acquired in real time, and then the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal are acquired respectively. According to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal, the speech loudness level of the speech signal and the ambient loudness level of the ambient noise signal are acquired respectively. Finally, based on the loudness level difference between the speech signal and the ambient noise signal, the sound signal is subjected to noise reduction processing to obtain a call signal. The technical solution provided by the present invention determines the loudness level difference between the ambient noise signal and the speech signal through the sound pressure level and frequency of the speech signal and the ambient noise signal in the sound signal, and performs noise reduction processing on the sound signal according to the loudness level difference. The present application performs speech noise reduction based on a loudness level that is more in line with the human ear perception, so that the noise reduction effect is better.

[0050] In a possible implementation of the present application, step S100 specifically includes:

[0051] Extracting speech feature information from the sound signal;

[0052] A speech signal and an environmental noise signal are identified according to the speech feature information.

[0053] Specifically, it is first necessary to identify the sound signal according to the signal characteristics of the speech signal and the ambient noise signal. The above-mentioned speech signal and ambient noise signal can be identified by pre-storing a human voice model and a noise model. The above-mentioned model contains speech features of the corresponding sound, such as sound intensity, loudness, frequency, signal-to-noise ratio, short-time average zero-crossing rate, etc. For example, after sampling the sound signal, the sound signal is matched with the model. If the sound signal contains all the features in the human voice model, it means that someone is currently speaking. If the sound signal cannot match the human voice model, it may be that the user's voice is too small or the background noise is too large, indicating that the human voice cannot be recognized from the currently acquired sound. Similarly, the acquired sound signal can be identified according to the pre-stored tire noise model, air conditioning noise model or wind noise model, so that the type of noise contained in the current sound signal can be determined.

[0054] Furthermore, the speech signal and the environmental noise signal in the sound signal are identified through the speech feature information.

[0055] In a possible implementation manner of the present application, step S300 specifically includes:

[0056] Finding the speech loudness level of the speech signal in an equal loudness curve model according to the sound pressure level and frequency of the speech signal;

[0057] The ambient loudness level of the ambient noise signal is searched in the equal loudness curve model according to the sound pressure level and frequency of the ambient noise signal.

[0058] In the embodiment of the present application, the speech loudness level of the speech signal and the ambient loudness level of the ambient noise signal can both be found and determined from the equal loudness curve model.

[0059] In a possible implementation of the present application, a method for constructing an equal loudness curve model specifically includes:

[0060] Obtaining the user's judgment results on the softness and loudness of pure tones of different sound pressure levels and frequencies;

[0061] The soft-loudness judgment result is input into a neural network model trained by equal-loudness curves to obtain the equal-loudness curve model.

[0062] In the embodiment of the present application, a preset program can be started, in which pure tones of the same sound pressure level but different frequencies and pure tones of different sound pressure levels but the same frequency are compared in pairs, to obtain the soft-loudness judgment result input by the user, and then the soft-loudness judgment result is input into the neural network model to obtain the equal-loudness curve model. The neural network model is a neural network model trained by the equal-loudness curve.

[0063] In a possible implementation manner of the present application, the noise reduction intensity of the noise reduction process is proportional to the loudness level difference.

[0064] In a possible implementation manner of the present application, step S400 specifically includes:

[0065] Calculating a loudness level difference between an ambient loudness level of the ambient noise signal and a voice loudness level of the voice signal;

[0066] Filtering out the maximum loudness level difference value among the loudness level difference values;

[0067] The sound signal is subjected to noise reduction processing based on the maximum loudness level difference value to obtain a call signal.

[0068] In the embodiment of the present application, since the ambient noise signal may contain noises of various frequencies and sound pressure levels, when calculating the loudness level difference between the ambient loudness level of the ambient noise signal and the voice loudness level of the voice signal, multiple loudness level difference values ​​may appear, and the maximum loudness level difference value with the largest loudness level difference value is screened out, and the sound signal is subjected to noise reduction processing according to the maximum loudness level difference value to obtain a call signal. Specifically, the noise reduction intensity is proportional to the maximum loudness level difference value.

[0069] In a possible implementation manner of the present application, after step S200, the following steps are further included:

[0070] Determining whether the sound signal meets a loudness level noise reduction condition according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal;

[0071] When the sound signal meets the loudness level noise reduction condition, the speech loudness level of the speech signal and the environment loudness level of the environment noise signal are obtained according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the environment noise signal, and the sound signal is denoised based on the difference between the environment loudness level and the speech loudness level to obtain a call signal;

[0072] When the sound signal does not meet the loudness level noise reduction condition, the sound signal is denoised based on the sound pressure level difference between the ambient noise signal and the voice signal to obtain a call signal.

[0073] That is, the noise reduction processing of the sound signal is divided into two cases, one is whether the sound signal meets the loudness level noise reduction condition, and the noise reduction processing is performed through the loudness level difference, which is the case described in the above embodiment; the other is that the sound signal does not meet the loudness level noise reduction condition, and the noise reduction processing is directly performed according to the sound pressure level difference between the ambient noise signal and the voice signal. By first determining whether the sound signal meets the loudness level noise reduction condition and then determining the noise reduction processing method, the amount of calculation can be reduced and the noise reduction efficiency can be improved.

[0074] In a possible implementation of the present application, judging whether the sound signal satisfies the loudness level noise reduction condition according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal specifically includes:

[0075] When the difference between the sound pressure levels of the ambient noise signal and the speech signal is greater than a sound pressure level threshold, or when there is an ambient noise signal with a frequency within a preset frequency range in the ambient noise signal, determining that the sound signal meets the loudness level noise reduction condition;

[0076] When the sound pressure level difference between the ambient noise signal and the speech signal is less than or equal to the sound pressure level threshold, and there is no ambient noise signal with a frequency within a preset frequency range in the ambient noise signal, it is determined that the sound signal does not meet the loudness level noise reduction condition.

[0077] That is, the sound signal satisfies the loudness level noise reduction condition when: the difference in sound pressure level between the ambient noise signal and the speech signal is greater than the sound pressure level threshold, or the ambient noise signal contains an ambient noise signal whose frequency is within a preset frequency range.

[0078] Specifically, the preset frequency range can be set to 2k-5kHz. According to the equal loudness curve, Figure 2 As shown, the human ear is most sensitive to sounds in the 2k-5kHz range, that is, at the same sound pressure level, the sounds in this frequency range are louder than sounds of other frequencies. Therefore, in order to prevent the noise in this frequency range from having a significant impact on call quality, the focus is on processing the noise in this frequency range.

[0079] Furthermore, the case where the sound signal does not meet the loudness level noise reduction condition is described in detail in the following embodiments.

[0080] In a possible implementation of the present application, determining whether the sound signal meets the loudness level noise reduction condition specifically includes:

[0081] Calculating a sound pressure level difference between the ambient noise signal and the speech signal;

[0082] Determining whether the sound pressure level difference is greater than a sound pressure level threshold;

[0083] If the sound pressure level difference is less than or equal to the sound pressure level threshold, it means that the sound signal does not meet the loudness level noise reduction condition, and the sound is denoised according to the sound pressure level difference.

[0084] In the embodiment of the present application, by calculating whether there is an environmental noise signal in the environmental noise signal whose sound pressure level difference with the voice signal is greater than the sound pressure level threshold, if not, the above-mentioned sound signal is denoised according to the sound pressure level difference. Among them, the noise reduction intensity is proportional to the sound pressure level difference. If it exists, the noise reduction is performed according to the sound pressure level and frequency of the environmental noise signal. The noise reduction method has been described in detail in the above embodiment and will not be repeated here.

[0085] In a possible implementation of the present application, the loudness level-based speech noise reduction method may further include:

[0086] Calculate the loudness level difference between the ambient loudness level after noise reduction and the speech loudness level;

[0087] Determining whether the loudness level difference is greater than a loudness threshold;

[0088] If the loudness level difference is greater than the loudness threshold, it is determined that the noise-reduced sound signal is a call signal.

[0089] In the embodiment of the present application, after the above noise reduction process is completed, the loudness level difference between the ambient loudness level and the voice loudness level after noise reduction is further calculated, and the magnitude of the loudness level difference and the loudness threshold is determined. If the loudness level difference is greater than the loudness threshold, the sound signal after noise reduction is determined to be the final call signal. That is, the sound signal at this time is already very close to the human ear, and the noise reduction effect is very good.

[0090] In a possible implementation of the present application, the loudness level-based speech noise reduction method may further include:

[0091] The call signal is transmitted to another terminal via the 5G network.

[0092] In the embodiment of the present application, the sound signal is transmitted through the 5G network, making the sound clearer, the transmission speed faster, and enhancing the user experience.

[0093] In another embodiment of the present application, a speech noise reduction system based on loudness level is also provided. Figure 3 As shown, the loudness level-based speech noise reduction system may include the following modules: a first acquisition module 1 , a second acquisition module 2 , a determination module 3 and a noise reduction module 4 .

[0094] The first acquisition module 1 is configured to identify a speech signal and an environmental noise signal from a sound signal acquired in real time.

[0095] The second acquisition module 2 is configured to respectively acquire the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal.

[0096] The determination module 3 is configured to obtain the speech loudness level of the speech signal and the environmental loudness level of the environmental noise signal according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the environmental noise signal.

[0097] The noise reduction module 4 is configured to perform noise reduction processing on the sound signal based on the loudness level difference between the ambient noise signal and the voice signal to obtain a call signal.

[0098] In the embodiment of the present application, a voice signal and an environmental noise signal are identified from the sound signal of the first terminal acquired in real time by a first acquisition module, and then the sound pressure level and frequency of the voice signal and the sound pressure level and frequency of the environmental noise signal are acquired respectively by a second acquisition module, and then the voice loudness level of the voice signal and the environmental loudness level of the environmental noise signal are obtained respectively by a determination module according to the sound pressure level and frequency of the voice signal and the sound pressure level and frequency of the environmental noise signal, and finally the noise reduction module performs noise reduction processing on the sound signal based on the loudness level difference between the environmental noise signal and the voice signal to obtain a call signal. The technical solution provided by the present invention determines the loudness level difference between the environmental noise signal and the voice signal by the sound pressure level and frequency of the voice signal and the environmental noise signal in the sound signal, performs noise reduction processing on the sound signal according to the loudness level difference, and performs voice noise reduction based on a loudness level that is more in line with the human ear perception, so that the noise reduction effect is better.

[0099] Furthermore, the first acquisition module 1 is specifically used for:

[0100] Extracting speech feature information from the sound signal;

[0101] A speech signal and an environmental noise signal are identified according to the speech feature information.

[0102] Furthermore, it is determined that module 3 is specifically used for:

[0103] Finding the speech loudness level of the speech signal in an equal loudness curve model according to the sound pressure level and frequency of the speech signal;

[0104] The ambient loudness level of the ambient noise signal is searched in the equal loudness curve model according to the sound pressure level and frequency of the ambient noise signal.

[0105] Furthermore, the determination module 3 may also be used to:

[0106] Obtaining the user's judgment results on the softness and loudness of pure tones of different sound pressure levels and frequencies;

[0107] The soft-loudness judgment result is input into a neural network model trained by equal-loudness curves to obtain the equal-loudness curve model.

[0108] Furthermore, the noise reduction module 4 is specifically used for:

[0109] Calculating a loudness level difference between an ambient loudness level of the ambient noise signal and a voice loudness level of the voice signal;

[0110] Filtering out the maximum loudness level difference value among the loudness level difference values;

[0111] The sound signal is subjected to noise reduction processing based on the maximum loudness level difference value to obtain a call signal.

[0112] Furthermore, the loudness level-based speech noise reduction system may further include: a judgment module.

[0113] The judgment module is configured to judge whether the sound signal meets the loudness level noise reduction condition according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal;

[0114] When the sound signal meets the loudness level noise reduction condition, the speech loudness level of the speech signal and the environment loudness level of the environment noise signal are obtained according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the environment noise signal, and the sound signal is denoised based on the difference between the environment loudness level and the speech loudness level to obtain a call signal;

[0115] When the sound signal does not meet the loudness level noise reduction condition, the sound signal is denoised based on the sound pressure level difference between the ambient noise signal and the voice signal to obtain a call signal.

[0116] Furthermore, the judgment module is specifically used for:

[0117] When the difference between the sound pressure levels of the ambient noise signal and the speech signal is greater than a sound pressure level threshold, or when there is an ambient noise signal with a frequency within a preset frequency range in the ambient noise signal, determining that the sound signal meets the loudness level noise reduction condition;

[0118] When the sound pressure level difference between the ambient noise signal and the speech signal is less than or equal to the sound pressure level threshold, and there is no ambient noise signal with a frequency within a preset frequency range in the ambient noise signal, it is determined that the sound signal does not meet the loudness level noise reduction condition.

[0119] In another embodiment of the present disclosure, a computer storage medium is further provided, on which a method program for speech noise reduction based on loudness level is stored. When the method program for speech noise reduction based on loudness level is executed by a processor, the following steps are implemented:

[0120] Identify speech signals and environmental noise signals from sound signals acquired in real time;

[0121] Respectively acquiring the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal;

[0122] According to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal, respectively obtain the speech loudness level of the speech signal and the ambient loudness level of the ambient noise signal;

[0123] The sound signal is denoised based on the loudness level difference between the ambient noise signal and the voice signal to obtain a call signal.

[0124] In the embodiment of the present application, the voice signal and the environmental noise signal are first identified from the sound signal acquired in real time, and then the sound pressure level and frequency of the voice signal and the sound pressure level and frequency of the environmental noise signal are respectively acquired. According to the sound pressure level and frequency of the voice signal and the sound pressure level and frequency of the environmental noise signal, the voice loudness level of the voice signal and the environmental loudness level of the environmental noise signal are respectively obtained. Finally, based on the loudness level difference between the voice signal and the environmental noise signal, the voice signal is subjected to noise reduction processing to obtain a call signal. The technical solution provided by the present invention determines the loudness level difference between the environmental noise signal and the voice signal through the sound pressure level and frequency of the voice signal and the environmental noise signal in the sound signal, performs noise reduction processing on the sound signal according to the loudness level difference, and performs voice noise reduction based on a loudness level that is more in line with the human ear perception, so that the noise reduction effect is better.

[0125] In a possible implementation of the present application, identifying a speech signal and an environmental noise signal from a sound signal acquired in real time specifically includes:

[0126] Extracting speech feature information from the sound signal;

[0127] A speech signal and an environmental noise signal are identified according to the speech feature information.

[0128] Specifically, it is first necessary to identify the sound signal according to the signal characteristics of the speech signal and the ambient noise signal. The above-mentioned speech signal and ambient noise signal can be identified by pre-storing a human voice model and a noise model. The above-mentioned model contains speech features of the corresponding sound, such as sound intensity, loudness, frequency, signal-to-noise ratio, short-time average zero-crossing rate, etc. For example, after sampling the sound signal, the sound signal is matched with the model. If the sound signal contains all the features in the human voice model, it means that someone is currently speaking. If the sound signal cannot match the human voice model, it may be that the user's voice is too small or the background noise is too large, indicating that the human voice cannot be recognized from the currently acquired sound. Similarly, the acquired sound signal can be identified according to the pre-stored tire noise model, air conditioning noise model or wind noise model, so that the type of noise contained in the current sound signal can be determined.

[0129] Furthermore, the speech signal and the environmental noise signal in the sound signal are identified through the speech feature information.

[0130] In a possible implementation manner of the present application, obtaining the speech loudness level of the speech signal and the ambient loudness level of the ambient noise signal respectively according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal specifically includes:

[0131] Finding the speech loudness level of the speech signal in an equal loudness curve model according to the sound pressure level and frequency of the speech signal;

[0132] The ambient loudness level of the ambient noise signal is searched in the equal loudness curve model according to the sound pressure level and frequency of the ambient noise signal.

[0133] In the embodiment of the present application, the speech loudness level of the speech signal and the ambient loudness level of the ambient noise signal can both be found and determined from the equal loudness curve model.

[0134] In a possible implementation of the present application, a method for constructing an equal loudness curve model specifically includes:

[0135] Obtaining the user's judgment results on the softness and loudness of pure tones of different sound pressure levels and frequencies;

[0136] The soft-loudness judgment result is input into a neural network model trained by equal-loudness curves to obtain the equal-loudness curve model.

[0137] In the embodiment of the present application, a preset program can be started, in which pure tones of the same sound pressure level but different frequencies and pure tones of different sound pressure levels but the same frequency are compared in pairs, to obtain the soft-loudness judgment result input by the user, and then the soft-loudness judgment result is input into the neural network model to obtain the equal-loudness curve model. The neural network model is a neural network model trained by the equal-loudness curve.

[0138] In a possible implementation manner of the present application, the noise reduction intensity of the noise reduction process is proportional to the loudness level difference.

[0139] In a possible implementation of the present application, performing noise reduction processing on the sound signal based on the loudness level difference between the ambient noise signal and the voice signal to obtain a call signal specifically includes:

[0140] Calculating a loudness level difference between an ambient loudness level of the ambient noise signal and a voice loudness level of the voice signal;

[0141] Filtering out the maximum loudness level difference value among the loudness level difference values;

[0142] The sound signal is subjected to noise reduction processing based on the maximum loudness level difference value to obtain a call signal.

[0143] In the embodiment of the present application, since the ambient noise signal may contain noises of various frequencies and sound pressure levels, when calculating the loudness level difference between the ambient loudness level of the ambient noise signal and the voice loudness level of the voice signal, multiple loudness level difference values ​​may appear, and the maximum loudness level difference value with the largest loudness level difference value is screened out, and the sound signal is subjected to noise reduction processing according to the maximum loudness level difference value to obtain a call signal. Specifically, the noise reduction intensity is proportional to the maximum loudness level difference value.

[0144] In a possible implementation manner of the present application, after respectively acquiring the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal, the method further includes:

[0145] Determining whether the sound signal meets a loudness level noise reduction condition according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal;

[0146] When the sound signal meets the loudness level noise reduction condition, the speech loudness level of the speech signal and the environment loudness level of the environment noise signal are obtained according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the environment noise signal, and the sound signal is denoised based on the difference between the environment loudness level and the speech loudness level to obtain a call signal;

[0147] When the sound signal does not meet the loudness level noise reduction condition, the sound signal is denoised based on the sound pressure level difference between the ambient noise signal and the voice signal to obtain a call signal.

[0148] That is, the noise reduction processing of the sound signal is divided into two cases, one is whether the sound signal meets the loudness level noise reduction condition, and the noise reduction processing is performed through the loudness level difference, which is the case described in the above embodiment; the other is that the sound signal does not meet the loudness level noise reduction condition, and the noise reduction processing is directly performed according to the sound pressure level difference between the ambient noise signal and the voice signal. By first determining whether the sound signal meets the loudness level noise reduction condition and then determining the noise reduction processing method, the amount of calculation can be reduced and the noise reduction efficiency can be improved.

[0149] In a possible implementation of the present application, judging whether the sound signal satisfies the loudness level noise reduction condition according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal specifically includes:

[0150] When the difference between the sound pressure levels of the ambient noise signal and the speech signal is greater than a sound pressure level threshold, or when there is an ambient noise signal with a frequency within a preset frequency range in the ambient noise signal, determining that the sound signal meets the loudness level noise reduction condition;

[0151] When the sound pressure level difference between the ambient noise signal and the speech signal is less than or equal to the sound pressure level threshold, and there is no ambient noise signal with a frequency within a preset frequency range in the ambient noise signal, it is determined that the sound signal does not meet the loudness level noise reduction condition.

[0152] That is, the sound signal satisfies the loudness level noise reduction condition when: the difference in sound pressure level between the ambient noise signal and the speech signal is greater than the sound pressure level threshold, or the ambient noise signal contains an ambient noise signal whose frequency is within a preset frequency range.

[0153] Specifically, the preset frequency range can be set to 2k-5kHz. According to the equal loudness curve, Figure 2 As shown, the human ear is most sensitive to sounds in the 2k-5kHz range, that is, at the same sound pressure level, the sounds in this frequency range are louder than sounds of other frequencies. Therefore, in order to prevent the noise in this frequency range from having a significant impact on call quality, the focus is on processing the noise in this frequency range.

[0154] In a possible implementation of the present application, the loudness level-based speech noise reduction method further includes:

[0155] The call signal is transmitted to another terminal via the 5G network.

[0156] In the embodiment of the present application, the sound signal is transmitted through the 5G network, making the sound clearer, the transmission speed faster, and enhancing the user experience.

[0157] It should be noted that, in this document, relational terms such as “first” and “second” are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations.

[0158] It is understood that the embodiments described herein may be implemented in hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit may be implemented in one or more application specific integrated circuits (ASIC), digital signal processors (DSP), digital signal processing devices (DSPD), programmable logic devices (PLD), field programmable gate arrays (FPGA), general purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described in the present application, or a combination thereof.

[0159] For software implementation, the technology described herein can be implemented by a unit that performs the functions described herein. The software code can be stored in a memory and executed by a processor. The memory can be implemented in the processor or outside the processor.

[0160] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0161] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0162] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0163] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, ROM, RAM, magnetic disks or optical disks.

[0164] It should be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0165] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for reducing speech noise based on loudness level, characterized in that: include: Identify speech signals and environmental noise signals from sound signals acquired in real time; Respectively acquiring the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal; According to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal, respectively obtain the speech loudness level of the speech signal and the ambient loudness level of the ambient noise signal; Performing noise reduction processing on the sound signal based on the loudness level difference between the ambient noise signal and the voice signal to obtain a call signal; After respectively acquiring the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal, the method further includes: Determining whether the sound signal meets a loudness level noise reduction condition according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal; When the sound signal meets the loudness level noise reduction condition, the speech loudness level of the speech signal and the environment loudness level of the environment noise signal are obtained according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the environment noise signal, and the sound signal is denoised based on the difference between the environment loudness level and the speech loudness level to obtain a call signal; When the sound signal does not meet the loudness level noise reduction condition, the sound signal is denoised based on the sound pressure level difference between the ambient noise signal and the voice signal to obtain a call signal.

2. A method for speech noise reduction based on loudness level according to claim 1, characterized in that: The step of identifying a speech signal and an environmental noise signal from a sound signal acquired in real time includes: Extracting speech feature information from the sound signal; A speech signal and an environmental noise signal are identified according to the speech feature information.

3. The method for speech noise reduction based on loudness level according to claim 1, characterized in that: The obtaining, according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal, the speech loudness level of the speech signal and the ambient loudness level of the ambient noise signal respectively comprises: Finding the speech loudness level of the speech signal in an equal loudness curve model according to the sound pressure level and frequency of the speech signal; The ambient loudness level of the ambient noise signal is searched in the equal loudness curve model according to the sound pressure level and frequency of the ambient noise signal.

4. The method for speech noise reduction based on loudness level according to claim 3, characterized in that: The method for constructing the equal loudness curve model comprises: Obtaining the user's judgment results on the softness and loudness of pure tones of different sound pressure levels and frequencies; The soft-loudness judgment result is input into a neural network model trained by equal-loudness curves to obtain an equal-loudness curve model.

5. The method for speech noise reduction based on loudness level according to claim 1, characterized in that: The noise reduction intensity of the noise reduction process is proportional to the loudness level difference.

6. A method for reducing speech noise based on loudness level according to claim 1 or 5, characterized in that: The method further comprises: performing noise reduction processing on the sound signal based on the loudness level difference between the ambient noise signal and the voice signal to obtain a call signal, comprising: Calculating a loudness level difference between an ambient loudness level of the ambient noise signal and a voice loudness level of the voice signal; Filtering out the maximum loudness level difference value among the loudness level difference values; The sound signal is subjected to noise reduction processing based on the maximum loudness level difference value to obtain a call signal.

7. The method for speech noise reduction based on loudness level according to claim 1, characterized in that: The determining, based on the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal, whether the sound signal satisfies the loudness level noise reduction condition comprises: When the difference between the sound pressure levels of the ambient noise signal and the speech signal is greater than a sound pressure level threshold, or when there is an ambient noise signal with a frequency within a preset frequency range in the ambient noise signal, determining that the sound signal meets the loudness level noise reduction condition; When the sound pressure level difference between the ambient noise signal and the speech signal is less than or equal to the sound pressure level threshold, and there is no ambient noise signal with a frequency within a preset frequency range in the ambient noise signal, it is determined that the sound signal does not meet the loudness level noise reduction condition.

8. A speech noise reduction system based on loudness level, characterized in that: include: A first acquisition module is used to identify a speech signal and an environmental noise signal from a sound signal acquired in real time; A second acquisition module, used to respectively acquire the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal; A determination module, configured to obtain a speech loudness level of the speech signal and an environmental loudness level of the environmental noise signal according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the environmental noise signal; A noise reduction module, configured to perform noise reduction processing on the sound signal based on a loudness level difference between the ambient noise signal and the voice signal to obtain a call signal; Wherein, the loudness level-based speech noise reduction system further includes: a judgment module; The judgment module is configured to judge whether the sound signal meets the loudness level noise reduction condition according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the ambient noise signal; When the sound signal meets the loudness level noise reduction condition, the speech loudness level of the speech signal and the environment loudness level of the environment noise signal are obtained according to the sound pressure level and frequency of the speech signal and the sound pressure level and frequency of the environment noise signal, and the sound signal is denoised based on the difference between the environment loudness level and the speech loudness level to obtain a call signal; When the sound signal does not meet the loudness level noise reduction condition, the sound signal is denoised based on the sound pressure level difference between the ambient noise signal and the voice signal to obtain a call signal.

9. A computer storage medium, characterized in that: The computer storage medium stores a program for a method for speech noise reduction based on loudness level, and when the program for speech noise reduction based on loudness level is executed by a processor, the steps of the method for speech noise reduction based on loudness level according to any one of claims 1 to 7 are implemented.

Images