A single microphone long distance sound pickup method and device

Abstract

The application discloses a single microphone long-distance sound pickup method and device, which comprises the following steps: determining a microphone bias voltage according to the microphone type, circuit and characteristics of the single microphone to be used, directly picking up sound from the sound source by the single microphone to obtain a to-be-processed audio signal; automatically adjusting the overall level of the to-be-processed audio signal by using an automatic level control algorithm; converting the to-be-processed audio signal after the level adjustment into a digital signal; sequentially performing noise reduction, speech enhancement and gain adjustment processing on the digital signal by using a speech enhancement algorithm and an automatic gain control algorithm, that is, using a method combining an analog domain and a digital domain to improve the sound pickup distance; and using an automatic level control (ALC) technology to optimize the quality of the audio signal collected by the single microphone.

Description

Technical Field

[0001] This invention relates to the field of audio and video technology, specifically to a method and apparatus for long-distance sound pickup with a single microphone, an electronic device, and a storage medium. Background Technology

[0002] In video surveillance applications, the sound source and the camera are usually far apart. The purpose of this invention is to solve the problem of sound pickup by microphones at long distances.

[0003] Existing technologies typically use microphone arrays, which employ multiple microphones and array algorithms to achieve long-distance sound pickup. However, due to limitations in hardware size and product cost, microphone array solutions are not always feasible.

[0004] In a single-microphone scenario, the smaller signal, i.e. the signal collected from a distant sound source, can be amplified by increasing the analog gain. However, if there is a larger signal (a closer sound source), the collected sound will be clipped.

[0005] Some solutions employ automatic gain control in the digital domain, but the results are unsatisfactory because the analog gain cannot be maximized. Summary of the Invention

[0006] One of the objectives of this invention is to provide a single-microphone long-distance sound pickup method, apparatus, electronic device, and storage medium, which addresses the shortcomings of the prior art by providing a single-microphone long-distance sound pickup solution that combines analog and digital domains.

[0007] To address the aforementioned technical problems, in a first aspect, embodiments of the present invention provide a single-microphone long-distance sound pickup method, the method comprising:

[0008] The microphone bias voltage is determined based on the microphone type, circuit, and characteristics of the single microphone used. The single microphone is then positioned directly in front of the sound source to pick up the sound and obtain the audio signal to be processed.

[0009] An automatic level control algorithm is used to automatically adjust the overall level of the audio signal to be processed;

[0010] Convert the adjusted audio signal into a digital signal;

[0011] The digital signal is sequentially processed by noise reduction, speech enhancement, and gain adjustment using a speech enhancement algorithm and an automatic gain control algorithm.

[0012] Preferably, the automatic level control algorithm for automatically adjusting the overall level of the audio signal to be processed specifically includes:

[0013] The input signal level is obtained by sampling and measuring the amplitude of the input audio signal to be processed;

[0014] Calculate the required target output signal level based on the preset target level;

[0015] Calculate the required gain value based on the difference between the input signal level and the target output signal level, and adjust the output signal level according to the calculated gain value;

[0016] The input audio signal to be processed is detected and adjusted based on the output signal level.

[0017] Preferably, the step of converting the adjusted audio signal to a digital signal specifically includes:

[0018] During the sampling period T s The audio signal x(t) to be processed is sampled to obtain a series of discrete sample values ​​x(nT). s ), where n = 0, 1, ...;

[0019] The amplitude A of each sampled value is compared with a predefined quantization level to obtain the quantized value, wherein the quantization level is denoted by q0, q1, q2, ..., q L-1 This indicates that, where L is the number of quantization levels, if the amplitude A is between the nth and (n+1)th quantization levels, and A is closer to the nth quantization level, then the quantization value is q. n-1 If it is closer to the (n+1)th quantization level, then the quantization value is q. n ;

[0020] Each quantized value q after quantization i It is converted into the corresponding binary digital signal x(n) according to the preset encoding rules.

[0021] Preferably, the process of sequentially performing noise reduction, speech enhancement, and gain adjustment on the digital signal using a speech enhancement algorithm and an automatic gain control algorithm specifically includes:

[0022] The digital signal is segmented and windowed, and a short-time Fourier transform is performed on each frame to obtain the spectral characteristics of the audio signal. The formula for the Fourier transform is as follows:

[0023]

[0024] Calculate the noise power spectrum based on the aforementioned spectral characteristics;

[0025] Based on the noise power spectrum, the probability of speech presence is calculated, and the noise power spectrum estimate is updated based on the probability of speech presence.

[0026] Preferably, the frame segmentation process specifically includes:

[0027] When the audio signal is x(n), n=0,1,…,N-1, where N is the signal length, then the frame length is M, the frame shift is S, and the signal x of the i-th frame is... i (m)=x(iS+m), m=0,1,…,M-1;

[0028] The windowing process specifically includes: expressing the Hamming window function w(m) as... Then the windowed signal y i (m)=x i (m)w(m)=x(iS+m)w(m), m=0,1,…,M-1.

[0029] Preferably, the noise power spectrum is calculated based on the spectral characteristics;

[0030] The square of the complex spectral amplitude of the audio signal in frame 0 is used as the initial noise power spectrum, i.e., P(k,0)=|Y(k,0)| 2 Where P(k,0) is the noise power spectrum at the k-th frequency point of the 0th frame, and Y(k,0) is the complex spectrum at the k-th frequency point of the 0th frame;

[0031] If, in the frequency domain, the power spectral density of each subsequent frame of noisy speech after spectrum smoothing is... b(i) represents the standardized window function with a length of 2ω+1, and Y(ki,l) represents the amplitude of the noisy speech under the short-time Fourier transform in the time-frequency domain. Recursive smoothing is then used to calculate the power spectrum of the current frame, i.e., P(k,l)=α. P P(k,l-1)+(1-α P )|Y(k,l)| 2 , where α P The smoothing factor is set to 0.9-0.95. P(k,l) is the noise power spectrum of the k-th frequency point in the l-th frame, P(k,l-1) is the noise power spectrum of the k-th frequency point in the (l-1)-th frame, and Y(k,l) is the complex spectrum of the k-th frequency point in the l-th frame.

[0032] Preferably, calculating the probability of speech presence based on the noise power spectrum specifically includes:

[0033] In the initial stage, if the speech signal is absent or the speech signal strength is less than the preset signal strength threshold, the power spectrum of the acquired audio signal is directly used as the initial estimate of the noise power spectrum.

[0034] For each subsequent frame of audio signal, the minimum value of the spectral amplitude of the current frame is obtained. When both speech and noise follow a Gaussian distribution, the probability of speech presence is calculated by comparing the spectral amplitude of the audio signal in the current frame with the noise power spectrum estimate.

[0035] Secondly, embodiments of the present invention also provide a single-microphone long-range sound pickup device, the device comprising:

[0036] The audio acquisition unit is used to determine the microphone bias voltage based on the microphone type, circuit and characteristics of the single microphone used, and to pick up the sound from the sound source by pointing the single microphone directly at the sound source to obtain the audio signal to be processed.

[0037] A level control unit is used to automatically adjust the overall level of the audio signal to be processed using an automatic level control algorithm;

[0038] The analog-to-digital converter unit is used to convert the adjusted audio signal into a digital signal.

[0039] The digital domain signal processing unit employs a speech enhancement algorithm and an automatic gain control algorithm to sequentially perform noise reduction, speech enhancement, and gain adjustment processing on the digital signal.

[0040] Thirdly, the present invention also provides an electronic device, which includes a processor and a memory, the memory being used to store a computer program, the computer program including program instructions, and the processor being configured to invoke the program instructions to execute the method described above.

[0041] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing a computer program, the computer program including program instructions that, when executed by a processor or calculator, cause the processor to perform the method described above.

[0042] Compared with the prior art, the single-microphone long-distance sound pickup method and device provided in this embodiment of the invention have at least the following beneficial effects:

[0043] This invention, in its embodiments, determines the microphone bias voltage based on the microphone type, circuitry, and characteristics of the single microphone used. The single microphone is then positioned directly facing the sound source to pick up the audio signal to be processed. An automatic level control algorithm is employed to automatically adjust the overall level of the audio signal. The adjusted audio signal is then converted into a digital signal. A speech enhancement algorithm and an automatic gain control algorithm are used to sequentially process the digital signal through noise reduction, speech enhancement, and gain adjustment, effectively combining analog and digital domain techniques to improve the pickup distance. Finally, automatic level control (ALC) technology is used to optimize the quality of the audio signal acquired by the single microphone. Attached Figure Description

[0044] The preferred embodiments will now be described in a clear and easy-to-understand manner, in conjunction with the accompanying drawings, to further explain the above-mentioned characteristics, technical features, advantages, and implementation methods of the present invention.

[0045] Figure 1 This is a schematic flowchart of a single-microphone long-distance sound pickup method according to an embodiment of the present invention;

[0046] Figure 2 This is a schematic diagram of a single-microphone long-distance sound pickup device according to an embodiment of the present invention;

[0047] Figure 3 This is a schematic diagram of an electronic device structure according to an embodiment of the present invention. Detailed Implementation

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the specific implementation methods of the present invention will be described below with reference to the accompanying drawings. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings and other implementation methods can be obtained based on these drawings without creative effort.

[0049] To keep the drawings concise, each figure only schematically shows the parts relevant to the invention, and these do not represent the actual structure of the product. Furthermore, to facilitate understanding, in some figures, only one of components with the same structure or function is schematically depicted, or only one is labeled. In this document, "one" not only means "only one," but can also mean "more than one."

[0050] The following detailed description of the implementation of the technical solution of the present invention will be based on some specific embodiments.

[0051] like Figure 1 As shown, in order to achieve the objective of this invention, an embodiment of this invention provides a single-microphone long-distance sound pickup method, the method comprising:

[0052] The microphone bias voltage is determined based on the microphone type, circuit, and characteristics of the single microphone used. The single microphone is then positioned directly in front of the sound source to pick up the sound and obtain the audio signal to be processed.

[0053] An automatic level control algorithm is used to automatically adjust the overall level of the audio signal to be processed;

[0054] Convert the adjusted audio signal into a digital signal;

[0055] The digital signal is sequentially processed by noise reduction, speech enhancement, and gain adjustment using a speech enhancement algorithm and an automatic gain control algorithm.

[0056] The embodiments of the present invention employ a method combining analog and digital domains to improve the sound acquisition distance.

[0057] The analog domain introduces automatic level control to automatically and dynamically adjust the signal magnitude. When the signal is large, a smaller gain is applied to prevent clipping; when the signal is small, a larger gain is applied to amplify the small signal, ultimately achieving a more balanced output level in the analog domain.

[0058] The digital domain introduces noise reduction and automatic gain control.

[0059] Since the analog domain output amplifies small signals, it also amplifies noise. Therefore, a noise reduction module is first introduced in the digital domain to reduce the noise amplification problem caused by automatic level control.

[0060] Finally, automatic gain control is added because the amplification factor of the automatic level control mentioned above is usually limited. Adding automatic gain control with voice detection further increases the signal strength and improves the signal-to-noise ratio.

[0061] The specific implementation steps of this invention are as follows:

[0062] Sound pickup steps:

[0063] A single microphone is used, and the appropriate microphone bias voltage is determined according to the microphone type, circuit and characteristics. Generally speaking, the bias voltage of electret microphones is usually between 2V and 5V. For specific values, please refer to the microphone's specification. The single microphone is placed directly in front of the sound source to pick up the audio signal to be processed.

[0064] Simulation domain processing steps:

[0065] The Automatic Level Control (ALC) algorithm is used to automatically adjust the overall level of the audio signal to keep it constant within a certain range, preventing signal clipping and distortion.

[0066] Simulation domain processing includes the following steps:

[0067] Detecting the level of the input audio signal: The level of the current audio signal is obtained by sampling and measuring the amplitude of the input signal;

[0068] Output signal level calculation: Calculate the required output signal level based on the set target level.

[0069] Gain control: Calculate the required gain value based on the difference between the input signal level and the target output signal level, and adjust the output signal level according to the calculated gain value.

[0070] Feedback control: The input signal is detected and adjusted based on the level of the output signal.

[0071] The analog-to-digital conversion process mainly includes three steps: sampling, quantization, and encoding.

[0072] Sampling: at regular time intervals Ts (Sampling period) is used to take values ​​for the analog signal x(t), resulting in a series of discrete sampled values ​​x(nT). s ), where n = 0, 1, ...

[0073] Quantization: The amplitude A of each sampled value is compared with a predefined quantization level, typically q0, q1, q2, ..., q... L-1 This indicates that L is the number of quantization levels. If the amplitude A is between the nth and (n+1)th quantization levels, and if A is closer to the nth quantization level, then it is quantized as q. n-1 If it is closer to the (n+1)th quantization level, then it is quantized to q. n .

[0074] Encoding: Encoding each quantized value q i It is converted into the corresponding binary code according to certain encoding rules. For example, if the quantization value is 12345, it will be converted into 16-bit binary code as 0011000000111001.

[0075] Digital domain signal processing steps:

[0076] The acquired audio signal is subjected to noise reduction, speech enhancement, and gain adjustment using speech enhancement and automatic gain control algorithms. The specific steps are as follows:

[0077] Signal preprocessing: The acquired audio signal is segmented and windowed, and then a short-time Fourier transform is performed on each frame to obtain the spectral characteristics of the audio signal.

[0078] The specific steps of frame segmentation are as follows: Assume the audio signal is x(n), n = 0, 1, ..., N-1, where N is the signal length. The frame length is M, the frame shift is S, and the signal x of the i-th frame... i (m)=x(iS+m), m=0,1,…,M-1.

[0079] Taking the Hamming window as an example, the specific steps of windowing are as follows: The expression for the Hamming window function w(m) is: Windowed signal y i (m)=x i (m)w(m)=x(iS+m)w(m), m=0,1,…,M-1.

[0080] The formula for the Fourier transform is:

[0081]

[0082] Calculate the noise power spectrum based on the aforementioned spectral characteristics;

[0083] Based on the noise power spectrum, the probability of speech presence is calculated, and the noise power spectrum estimate is updated based on the probability of speech presence.

[0084] In the noise power spectrum estimation, at the initial stage, assuming that the speech signal is absent or very weak, the power spectrum of the acquired audio signal is directly used as the initial estimate of the noise power spectrum. For each subsequent frame of audio signal, the minimum value of the spectral amplitude is calculated. Assuming that both speech and noise follow a Gaussian distribution, the probability of speech presence is calculated by comparing the spectral amplitude of the audio signal in the current frame with the noise power spectrum estimate. Finally, the noise power spectrum estimate is updated based on the probability of speech presence. If the probability of speech presence is low, the minimum value is used to update the noise power spectrum estimate; if the probability of speech presence is high, the noise power spectrum estimate is not updated.

[0085] Specifically, calculating the noise power spectrum based on the spectral characteristics includes:

[0086] The square of the complex spectral amplitude of the audio signal in frame 0 is used as the initial noise power spectrum, i.e., P(k,0)=|Y(k,0)| 2 Where P(k,0) is the noise power spectrum at the k-th frequency point of the 0th frame, and Y(k,0) is the complex spectrum at the k-th frequency point of the 0th frame;

[0087] If, in the frequency domain, the power spectral density of each subsequent frame of noisy speech after spectrum smoothing is... b(i) represents the standardized window function with a length of 2ω+1, and Y(ki,l) represents the amplitude of the noisy speech under the short-time Fourier transform in the time-frequency domain. Recursive smoothing is then used to calculate the power spectrum of the current frame, i.e., P(k,l)=α. P P(k,l-1)+(1-α P )|Y(k,l)| 2 , where α P The smoothing factor is set to 0.9-0.95. P(k,l) is the noise power spectrum of the k-th frequency point in the l-th frame, P(k,l-1) is the noise power spectrum of the k-th frequency point in the (l-1)-th frame, and Y(k,l) is the complex spectrum of the k-th frequency point in the l-th frame.

[0088] For the initial noise power spectrum:

[0089] The square of the complex spectral amplitude of the audio signal in frame 0 is used as the initial noise power spectrum, i.e., P(k,0)=|Y(k,0)| 2 , where P(k,0) is the noise power spectrum of the k-th frequency point in the 0th frame, and Y(k,0) is the complex spectrum of the k-th frequency point in the 0th frame.

[0090] Other frame noise power spectra:

[0091] The power spectral density value after smoothing the noisy speech spectrum for each frame in the frequency domain is... b(i) represents the standardized window function with a length of 2ω+1, and Y(ki,l) represents the amplitude of the noisy speech in the time-frequency domain using the short-time Fourier transform.

[0092] Next, a recursive smoothing method is used to calculate the power spectrum of the current frame, i.e., P(k,l)=α. P P(k,l-1)+(1-α P )|Y(k,l)| 2 , where α P The smoothing factor is usually set to 0.9-0.95. P(k,l) is the noise power spectrum of the k-th frequency point in the l-th frame, P(k,l-1) is the noise power spectrum of the k-th frequency point in the (l-1)-th frame, and Y(k,l) is the complex spectrum of the k-th frequency point in the l-th frame.

[0093] Minimum tracking: Search for the minimum value within a time window as a noise candidate. N is the length of the search window.

[0094] Define variables B min =1.66 is the bias compensation factor for the minimum noise power spectrum estimation.

[0095] definition I(k,l) = 1 indicates that the speech does not exist, and I(k,l) = 0 indicates that the speech exists;

[0096] Among them, the empirical values ​​are γ0 = 4.6 and ζ0 = 1.67.

[0097] The smoothed power spectrum values ​​are obtained by performing a second smoothing on the power spectrum at different frequencies:

[0098]

[0099] The power spectrum value obtained by using the first-order recursive smoothing method in the time domain is:

[0100]

[0101] Tracking the second-order smoothed power spectrum Minimum value:

[0102]

[0103] Define two more variables:

[0104]

[0105] Calculate the estimated probability of speech presence. ζ0=1.67, γ1=3

[0106]

[0107] The noise power spectrum estimate is updated based on the probability of speech presence, specifically including:

[0108] Recursively update the noise power spectrum estimation by combining the estimated probability of speech presence:

[0109]

[0110] Where, α d This is the noise smoothing factor, typically between 0.85 and 0.95. This is the updated result of the noise power spectrum estimation for the k-th frequency point in the l-th frame.

[0111] To achieve noise reduction and elimination, the following steps are required:

[0112] Speech enhancement: Based on the updated noise power spectrum estimation results, the posterior signal-to-noise ratio is calculated. Then, the prior signal-to-noise ratio is calculated by combining the updated noise power spectrum estimation and the posterior signal-to-noise ratio. Next, the gain function is calculated according to the minimum mean square error criterion. Finally, the logarithmic spectral amplitude of the audio signal and the logarithm of the gain function are added together to obtain the adjusted logarithmic spectral amplitude.

[0113] The formula for calculating the posterior signal-to-noise ratio is as follows:

[0114] Prior signal-to-noise ratio calculation formula:

[0115]

[0116] The formula for the gain function when speech is present is: in

[0117] Calculate the speech presence value based on the estimated probability of speech presence:

[0118] The enhanced audio signal spectrum is obtained as follows:

[0119] Signal recovery: The adjusted logarithmic spectrum amplitude is converted back to the time domain signal using the inverse short-time Fourier transform (ISFT) to obtain the enhanced audio signal. The IFT is as follows: Assuming X(k,l) is the frequency domain signal after the IFT, where k represents the frequency index, l represents the time frame index, and the window function is ω(l), then the recovered time domain signal is... N is the number of points in the Fourier transform.

[0120] Next, automatic gain control is performed to stabilize the voice volume, including the following steps:

[0121] First, signal detection and analysis are performed. The enhanced audio signal is monitored and analyzed in real time to obtain relevant signal features. The volume is determined by calculating the amplitude of the audio signal samples. The speech activity detection (VAD) algorithm is used to determine whether there is speech.

[0122] The speech activity detection algorithm is as follows: First, calculate the energy of the audio signal within a certain time window. Where x(i) is the value of the audio signal at the i-th sampling point, n represents the sequence number of the current frame, and N is the number of sampling points corresponding to the frame length. n usually starts from N-1, representing the last sampling point of the complete data of the first frame, and then increases at intervals of m corresponding to the frame shift, i.e., n = N-1, N-1+m, N-1+2m, ...

[0123] Then calculate the zero-crossing rate. ZCR(n) represents the number of times the audio signal crosses zero within a certain time period, i.e., the zero-crossing rate. sgn(·) is the sign function; if the energy is higher than a certain threshold T... E Furthermore, if the zero-crossing rate exceeds the threshold, it can be determined that there is voice activity. The energy threshold and the zero-crossing rate threshold need to be determined based on the specific application scenario and the characteristics of the audio data.

[0124] Gain calculation: Based on the target volume and the current signal volume, calculate the required gain value to amplify or attenuate the input audio signal.

[0125] Let the input audio signal be x(n), the output audio signal be y(n), and the gain be G(n).

[0126] In the speech segment, the gain is calculated based on the target volume T and the current signal energy E(n), such as... Then y(n) = G(n) × x(n).

[0127] In non-speech segments, a fixed, relatively small gain G can be set. min It is generally between 0dB and 6dB, and the specific value varies depending on the application scenario.

[0128] like Figure 2 As shown, this embodiment of the invention also provides a single-microphone long-distance sound pickup device, the device comprising:

[0129] The audio acquisition unit is used to determine the microphone bias voltage based on the microphone type, circuit, and characteristics of the single microphone used, and to pick up sound from the sound source using the single microphone to obtain the audio signal to be processed; the level control unit is used to automatically adjust the overall level of the audio signal to be processed using an automatic level control algorithm; the analog-to-digital conversion unit is used to convert the adjusted audio signal to be processed into a digital signal; the digital domain signal processing unit uses a speech enhancement algorithm and an automatic gain control algorithm to sequentially perform noise reduction, speech enhancement, and gain adjustment processing on the digital signal.

[0130] Here, the implementation of the device is the same as that of the method described above, and will not be repeated here.

[0131] The embodiments of the present invention use a combination of analog and digital domain methods to improve the pickup distance and use automatic level control (ALC) technology to optimize the quality of audio signals acquired by a single microphone. As a result, it has significant advantages over existing solutions under different intensities of sound sources. For example, the test results of two other solutions are shown in Table 1, which shows that the embodiments of the present invention have significant advantages.

[0132] It's important to note that competitor product two, when using a strong audio source, reduces noise signals to 0 levels through noise reduction, resulting in a maximum SNR (expressed as 100dB). However, this approach doesn't produce a pleasant listening experience and is generally not recommended. A signal of zero is represented by -100dBFS, indicating no signal at all.

[0133] (Table 1)

[0134]

[0135] Thirdly, the present invention also provides an electronic device, which includes a processor and a memory. The memory is used to store a computer program, the computer program including program instructions, and the processor or calculator is configured to call the program instructions to execute the method described above.

[0136] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing a computer program, the computer program including program instructions that, when executed by a processor or calculator, cause the processor or calculator to perform the method described above.

[0137] like Figure 3As shown in the figure, an electronic device 1000 provided in this application embodiment includes a processor or calculator (not shown) 1001 and a memory 1002. The processor or calculator 1001 and the memory 1002 can be interconnected via a communication bus 1003. The communication bus 1003 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus 1003 can be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, the memory 1002 is used to store a computer program, which includes program instructions. The processor 1001 is configured to call the program instructions. The program includes steps for executing some or all of the steps in the aforementioned methods.

[0138] The processor 1001 may be a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits used to control the execution of the above-mentioned program.

[0139] The memory 1002 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital versatile optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory may exist independently and be connected to the processor via a bus. The memory may also be integrated with the processor.

[0140] The electronic device 1000 may further include a communication module 1004 and a display 1005. The communication module 1004 can communicate with the optical tracking device. The communication module 1004 can be a wireless communication module (e.g., a WiFi module, a Bluetooth module, etc.) or a wired communication module.

[0141] In addition, the electronic device 1000 may also include general components such as communication interfaces (e.g., USB interfaces, microphone interfaces, etc.) and antennas, which will not be described in detail here.

[0142] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0143] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0144] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another apparatus, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical or other forms.

[0145] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0146] Furthermore, the functional units in the various embodiments of the application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software program module.

[0147] If the integrated unit is implemented as a software program module and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0148] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage device, which may include: a flash drive, a read-only memory, a random access memory, a magnetic disk, or an optical disk, etc.

[0149] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

[0150] It should be noted that the above embodiments can be freely combined as needed. The above are merely preferred embodiments of the present invention. It should be pointed out that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for long-distance sound pickup using a single microphone, characterized in that, The method includes: The microphone bias voltage is determined based on the microphone type, circuit, and characteristics of the single microphone used. The single microphone is then positioned directly in front of the sound source to pick up the sound and obtain the audio signal to be processed. An automatic level control algorithm is used to automatically adjust the overall level of the audio signal to be processed; Convert the adjusted audio signal into a digital signal; The digital signal is sequentially processed using a speech enhancement algorithm and an automatic gain control algorithm, performing noise reduction, speech enhancement, and gain adjustment. Specifically, this process includes: The digital signal is segmented and windowed, and a short-time Fourier transform is performed on each frame to obtain the spectral characteristics of the audio signal. The formula for the Fourier transform is as follows: Among them, Y i (k) represents the signal of the i-th frame digital signal in the k-frequency band, y i (m) represents the value of the m-th point of the time-domain signal in the i-th frame, where M is the length of a frame of digital signal; Calculate the noise power spectrum based on the aforementioned spectral characteristics; Based on the noise power spectrum, the probability of speech presence is calculated, and the noise power spectrum estimate is updated based on the probability of speech presence. The noise reduction and speech enhancement are specifically as follows: Based on the updated noise power spectrum estimation, the posterior signal-to-noise ratio (SNR) is calculated. Then, the prior SNR is calculated by combining the updated noise power spectrum estimation and the posterior SNR. Next, the gain function is calculated according to the minimum mean square error criterion. Finally, the logarithmic spectral amplitude of the audio signal and the logarithm of the gain function are added to obtain the adjusted logarithmic spectral amplitude. The adjusted logarithmic spectral amplitude is then converted back to the time domain signal using an inverse short-time Fourier transform (ISFT) to obtain the enhanced audio signal. The IFT is as follows: Assuming X(k,l) is the frequency domain signal after the IFT, where k represents the frequency index, l represents the time frame index, and the window function is ω(l), then the recovered time domain signal is... N is the number of points in the Fourier transform; The gain adjustment process involves automatic gain control to stabilize the voice volume, specifically including: First, signal detection and analysis are performed. The enhanced audio signal is monitored and analyzed in real time. The volume of the current signal is determined by calculating the amplitude of the audio signal samples. The presence of speech is determined by using a speech activity detection algorithm. The speech activity detection algorithm is as follows: First, the energy of the audio signal sample within a certain time window is calculated. Where x(i) is the value of the audio signal at the i-th sampling point, n represents the sequence number of the current frame, and N is the number of sampling points corresponding to the frame length. n usually starts from N-1, representing the last sampling point of the complete data of the first frame, and then increases at intervals of m corresponding to the frame shift, i.e., n = N-1, N-1+m, N-1+2m, ...; then the zero-crossing rate is calculated. ZCR(n) represents the number of times the audio signal crosses zero within a certain time period, i.e., the zero-crossing rate. sgn(·) is the sign function; if the energy is higher than a certain threshold T... E Furthermore, if the zero-crossing rate exceeds the threshold, it is determined that speech is present; Then, gain calculation is performed. Based on the target volume and the current signal volume, the required gain value is calculated to amplify or attenuate the input audio signal.

2. The single-microphone long-distance sound pickup method as described in claim 1, characterized in that, The automatic level control algorithm used to automatically adjust the overall level of the audio signal to be processed specifically includes: The input signal level is obtained by sampling and measuring the amplitude of the input audio signal to be processed; Calculate the required target output signal level based on the preset target level; Calculate the required gain value based on the difference between the input signal level and the target output signal level, and adjust the output signal level according to the calculated gain value; The input audio signal to be processed is detected and adjusted based on the output signal level.

3. The single-microphone long-distance sound pickup method as described in claim 1, characterized in that, The process of converting the adjusted audio signal into a digital signal specifically includes: During the sampling period T s The audio signal x(t) to be processed is sampled to obtain a series of discrete sample values ​​x(nT). s ), where n = 0, 1, ...; The amplitude A of each sampled value is compared with a predefined quantization level to obtain the quantized value, wherein the quantization level is denoted by q0, q1, q2, ..., q L-1 This indicates that, where L is the number of quantization levels, if the amplitude A is between the nth and (n+1)th quantization levels, and A is closer to the nth quantization level, then the quantization value is q. n-1 If it is closer to the (n+1)th quantization level, then the quantization value is q. n ; Each quantized value q after quantization i It is converted into the corresponding binary digital signal x(n) according to the preset encoding rules.

4. The single-microphone long-distance sound pickup method as described in claim 1, characterized in that, Frame division processing specifically includes: When the audio signal is x(n), n=0,1,…,N-1, where N is the signal length, then the frame length is M, the frame shift is S, and the signal x of the i-th frame is... i (m)=x(iS+m), m=0,1,…,M-1; The windowing process specifically includes: expressing the Hamming window function w(m) as... Then the windowed signal y i (m)=x i (m)w(m)=x(iS)+m)w(m), m=0,1,…,M-1.

5. The single-microphone long-distance sound pickup method as described in claim 1, characterized in that, The calculation of the noise power spectrum based on the spectral characteristics specifically includes: The square of the complex spectral amplitude of the audio signal in frame 0 is used as the initial noise power spectrum, i.e., P(k,0)=|Y(k,0)| 2 Where P(k,0) is the noise power spectrum at the k-th frequency point of the 0th frame, and Y(k,0) is the complex spectrum at the k-th frequency point of the 0th frame; If, in the frequency domain, the power spectral density of each subsequent frame of noisy speech after spectrum smoothing is... b(i) represents the standardized window function with a length of 2ω+1, and Y(ki,l) represents the amplitude of the noisy speech under the short-time Fourier transform in the time-frequency domain. Recursive smoothing is then used to calculate the power spectrum of the current frame, i.e., P(k,l)=α. P P(k,l-1)+(1-α P )|Y(k,l)| 2 , where α P The smoothing factor is set to 0.9-0.

95. P(k,l) is the noise power spectrum of the k-th frequency point in the l-th frame, P(k,l-1) is the noise power spectrum of the k-th frequency point in the (l-1)-th frame, and Y(k,l) is the complex spectrum of the k-th frequency point in the l-th frame.

6. The single-microphone long-distance sound pickup method as described in claim 1, characterized in that, The calculation of the probability of speech presence based on the noise power spectrum specifically includes: In the initial stage, if the speech signal is absent or the speech signal strength is less than the preset signal strength threshold, the power spectrum of the acquired audio signal is directly used as the initial estimate of the noise power spectrum. For each subsequent frame of audio signal, the minimum value of the spectral amplitude of the current frame is obtained. When both speech and noise follow a Gaussian distribution, the probability of speech presence is calculated by comparing the spectral amplitude of the audio signal in the current frame with the noise power spectrum estimate.

7. A single-microphone long-distance sound pickup device, characterized in that, The device includes: The audio acquisition unit is used to determine the microphone bias voltage based on the microphone type, circuit and characteristics of the single microphone used, and to pick up the sound from the sound source by pointing the single microphone directly at the sound source to obtain the audio signal to be processed. A level control unit is used to automatically adjust the overall level of the audio signal to be processed using an automatic level control algorithm; The analog-to-digital converter unit is used to convert the adjusted audio signal into a digital signal. The digital domain signal processing unit employs a speech enhancement algorithm and an automatic gain control algorithm to sequentially perform noise reduction, speech enhancement, and gain adjustment processing on the digital signal. Specifically, this sequential noise reduction, speech enhancement, and gain adjustment processing using the speech enhancement algorithm and automatic gain control algorithm includes: The digital signal is segmented and windowed, and a short-time Fourier transform is performed on each frame to obtain the spectral characteristics of the audio signal. The formula for the Fourier transform is as follows: Among them, Y i (k) represents the signal of the i-th frame digital signal in the k-frequency band, y i (m) represents the value of the m-th point of the time-domain signal in the i-th frame, where M is the length of a frame of digital signal; Calculate the noise power spectrum based on the aforementioned spectral characteristics; Based on the noise power spectrum, the probability of speech presence is calculated, and the noise power spectrum estimate is updated based on the probability of speech presence. The noise reduction and speech enhancement are specifically as follows: Based on the updated noise power spectrum estimation, the posterior signal-to-noise ratio (SNR) is calculated. Then, the prior SNR is calculated by combining the updated noise power spectrum estimation and the posterior SNR. Next, the gain function is calculated according to the minimum mean square error criterion. Finally, the logarithmic spectral amplitude of the audio signal and the logarithm of the gain function are added to obtain the adjusted logarithmic spectral amplitude. The adjusted logarithmic spectral amplitude is then converted back to the time domain signal using an inverse short-time Fourier transform (ISFT) to obtain the enhanced audio signal. The IFT is as follows: Assuming X(k,l) is the frequency domain signal after the IFT, where k represents the frequency index, l represents the time frame index, and the window function is ω(l), then the recovered time domain signal is... N is the number of points in the Fourier transform; The gain adjustment process involves automatic gain control to stabilize the voice volume, specifically including: First, signal detection and analysis are performed. The enhanced audio signal is monitored and analyzed in real time. The volume of the current signal is determined by calculating the amplitude of the audio signal samples. The presence of speech is determined by using a speech activity detection algorithm. The speech activity detection algorithm is as follows: First, the energy of the audio signal sample within a certain time window is calculated. Where x(i) is the value of the audio signal at the i-th sampling point, n represents the sequence number of the current frame, and N is the number of sampling points corresponding to the frame length. n usually starts from N-1, representing the last sampling point of the complete data of the first frame, and then increases at intervals of m corresponding to the frame shift, i.e., n = N-1, N-1+m, N-1+2m, ...; then the zero-crossing rate is calculated. ZCR(n) represents the number of times the audio signal crosses zero within a certain time period, i.e., the zero-crossing rate. sgn(·) is the sign function; if the energy is higher than a certain threshold T... E Furthermore, if the zero-crossing rate exceeds the threshold, it is determined that speech is present; Then, gain calculation is performed. Based on the target volume and the current signal volume, the required gain value is calculated to amplify or attenuate the input audio signal.

8. An electronic device, characterized in that, The electronic device includes a processor and a memory, the memory being used to store a computer program, the computer program including program instructions, and the processor being configured to invoke the program instructions to execute the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, the computer program including program instructions that, when executed by a processor, cause the processor or calculator to perform the method as described in any one of claims 1 to 6.

Images