Resolution enhancement method and system based on audio spectrum analysis

Abstract

The application discloses a resolution improvement method and system based on audio spectrum analysis, and the method comprises the following steps: a pressure sensor converts the vibration of a mute drumhead into an electric voltage signal output, and converts the electric voltage signal into an electric current signal through a coupling circuit; a consensus frequency component template is constructed based on the electric current signal, and the spectrum of an audio signal is filtered based on the consensus frequency component template; after the spectrum of the denoised audio signal is smoothed according to time frames, inverse discrete Fourier transform is performed to obtain a reconstructed audio signal with clear tone color. The application significantly improves the resolution and retention capability of each frequency component by collecting multiple coherent current signals, can maximize the retention of effective frequency components while suppressing noise, further processes the signal change in the time dimension to reduce signal fluctuation and distortion, and finally generates an audio feedback with high fidelity and high definition, thereby providing more accurate and reliable practice reference for users.

Description

Technical Field

[0001] This invention relates to the field of silent drum technology, and in particular to a method and system for improving resolution based on audio spectrum analysis. Background Technology

[0002] The silent drum, as an important percussion practice device, works by replacing the traditional drumhead with a high-damping material (such as rubber, silicone, or a mesh polymer drumhead) and combining it with an acoustically optimized structural design. This converts the intense sound wave energy generated by striking the drum into mechanical vibrations and absorbs them, significantly suppressing the airborne sound pressure level radiated into the environment, achieving a volume output close to a whisper. While this feature effectively solves the problem of noise pollution during practice, it also presents new challenges for performers: the extremely low initial volume makes it difficult for them to accurately judge the force of the strikes, rhythmic precision, timbre, and dynamic control effects based on direct auditory feedback, thus hindering effective evaluation and adjustment of playing techniques based on real-time auditory information.

[0003] To overcome the aforementioned limitations, a more advanced technical solution currently involves using a microphone to collect the weak audio signal from the mute drum, amplifying it, and then playing it back to the learner to provide timely and effective feedback. However, this audio signal collection and reconstruction process faces several technical challenges: First, due to the inherent characteristics of the mute drum's sound production mechanism, the inherent amplitude of certain frequency components (especially high-frequency harmonic components) in the effective audio signal it generates is low, making them easily masked by environmental noise or system background noise. Second, the signal acquisition process inevitably introduces electromagnetic interference, circuit noise, and environmental background noise, which may be close to the useful signal in the frequency domain. Third, during subsequent noise reduction using signal processing techniques (such as filtering, threshold discrimination, spectral subtraction, etc.), it is often difficult to accurately distinguish between small-amplitude useful frequency components and noise, resulting in the accidental removal of some low-amplitude useful frequency components while suppressing noise, causing signal distortion or loss of detail.

[0004] Furthermore, due to the transient nature of the striking action and the real-time constraints of the signal processing system, it cannot be guaranteed that the microphone can completely capture and retain all frequency component information at every moment. Moreover, under the short-time analysis framework, the frequency resolution capability is limited by the time-frequency uncertainty principle, making it difficult to achieve both high time resolution and high frequency resolution simultaneously. This further affects the integrity and accuracy of the reconstructed audio signal. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for improving the resolution of audio spectrum analysis of mute drums. This method can effectively preserve the weak characteristic frequency components of mute drums in a strong noise background and improve the accuracy of time-frequency analysis, truly meeting the needs of professional players for perceiving performance details and improving their techniques.

[0006] The technical solution proposed in this invention is a resolution enhancement method based on audio spectrum analysis, comprising the following steps:

[0007] S1. Install pressure sensors inside the silent drum. The pressure sensors are placed in a distributed array. The pressure sensors convert the drumhead vibration into a voltage signal output.

[0008] S2. Assign a unique number to all pressure sensors. The numbering rule should ensure that pressure sensors that are physically adjacent have consecutive numbers. Add a coupling circuit between each pair of consecutively numbered pressure sensors and connect the voltage signal output by each pressure sensor to the connected coupling circuit to output a current signal with the same frequency as the voltage signal.

[0009] S3. Perform spectrum analysis on each current signal obtained in step S2 to obtain the spectrum of the current signal. Count the number of times the frequency components appear in the spectrum of all current signals and retain the frequency components whose frequency exceeds the threshold to form a consensus frequency component template.

[0010] S4: Superimpose all current signals obtained in step S2 to generate an audio signal; divide the audio signal into time-domain segments and divide it into fixed-duration frame audio signals; uniformly sample each frame audio signal to obtain discrete frame audio signals; apply a window function to each discrete frame audio signal to reduce spectral leakage, and convert it into a frequency domain representation through discrete Fourier transform to obtain the spectrum of the frame audio signal.

[0011] The spectrum of each frame audio signal is filtered using the consensus frequency component template obtained in step S3, retaining only the frequency components present in the consensus frequency component template, to obtain the denoised frame audio signal spectrum.

[0012] S5. Arrange the spectrum of the denoised frame audio signal in the order of time frames and extract each frequency component. The amplitude of the sequence changes over time and is then processed by moving average filtering.

[0013] S6. Perform an inverse discrete Fourier transform on the spectrum of the filtered frame audio signal to obtain the reconstructed time-domain segment of the frame audio signal; splice all the obtained reconstructed time-domain segments to obtain the reconstructed audio signal; transmit the reconstructed audio signal to the headphones for playback.

[0014] Optionally, S1 includes:

[0015] The diaphragm of the pressure sensor is attached to the inside of the drumhead, and the resistance strain gauge is attached to the inside of the diaphragm.

[0016] When the drumhead vibrates, the diaphragm moves, and the resistance strain gauge stretches or compresses synchronously with the diaphragm, causing the resistance value to change.

[0017] The resistance change is converted into a voltage signal change by measuring the bridge circuit, and then the voltage signal is output.

[0018] Optionally, S3 includes:

[0019] Each current signal obtained in step S2 is uniformly sampled to obtain discrete current signals; each discrete current signal is converted into the spectrum of the current signal by discrete Fourier transform.

[0020] Optionally, S4 includes:

[0021] For the spectrum of each frame of audio signal, iterate through each frequency component. ;if In the consensus frequency component template obtained in step S3, the spectrum of the audio signal of that frame is retained. The amplitude information; if The consensus frequency component template obtained in step S3 is not included in the spectrum of the audio signal of this frame. The amplitude is set to 0.

[0022] Optionally, S5 includes:

[0023] S51: Arrange the spectrum of the denoised frame audio signal in the order of time frames to construct a three-dimensional data structure; the three dimensions of this three-dimensional data structure are the X-axis (time frame dimension), the Y-axis (frequency dimension) and the Z-axis (amplitude dimension).

[0024] S52: In a three-dimensional data structure, for each frequency component Its amplitude value varies with time frame The changing sequence is treated as an independent time-domain signal For each frequency component time domain signal A moving average filter is applied to suppress the time-domain signal. The high-frequency fluctuations were used to obtain a smoothed time-domain signal. .

[0025] Optionally, S52 includes:

[0026] For the smoothed time-domain signal The calculation is divided into boundary frames and intermediate frames;

[0027] For boundary frames, when time frames satisfy: or At that time, among them, This represents the window length of the moving average filter. It is an odd number. Represents time-domain signal In the length of the time frame dimension;

[0028] ;

[0029] For intermediate frames, when time frames satisfy: hour,

[0030] ;

[0031] in, Represents the time-domain signal When performing sliding filtering, the leftmost value in the window, Represents the time-domain signal The second value from the left in the window when performing sliding filtering. Represents the time-domain signal The rightmost value in the window when performing sliding filtering. Represents the time-domain signal The starting point of the time frame when performing sliding filtering. Represents the time-domain signal The end point of the time frame when performing sliding filtering.

[0032] The present invention also provides a resolution enhancement system based on audio spectrum analysis, comprising:

[0033] Pressure sensor module: The pressure sensors are installed in a distributed array inside the mute drum; the pressure sensors convert the drumhead vibration into a voltage signal output.

[0034] Voltage-to-current conversion module: Pressure sensor numbering; coupling circuits are added between consecutively numbered pressure sensors; voltage signals are converted into current signals through the coupling circuits;

[0035] Consensus Spectrum Component Template Module: Calculates the spectrum of the current signal, counts the number of occurrences of the spectrum components, retains the frequency components whose frequency exceeds the threshold, and forms a consensus frequency component template;

[0036] Audio signal denoising module: superimposes current signals to obtain audio signals; performs truncation, sampling, and discrete Fourier transform on the audio signals to obtain the spectrum of each frame of audio signals; retains the frequency components present in the consensus spectrum component template for the spectrum of the frame audio signals.

[0037] Audio signal time filtering module: Arranges the spectrum of the frame audio signal in the order of time frames, and performs moving average filtering on the sequence of amplitude changes of each frequency component over time;

[0038] Audio signal reconstruction module: Performs discrete Fourier inverse transform on the spectrum of the frame audio signal to obtain the reconstructed time domain segment of the frame audio signal; splices the reconstructed time domain segments of each frame audio signal in time frame order to obtain the reconstructed audio signal.

[0039] Beneficial effects:

[0040] This invention provides an audio signal improvement method suitable for mute drums. Compared with the original signal, the audio signal output by this invention removes noise and interference, reduces fluctuations and distortion, and has clear timbre and appropriate loudness.

[0041] In the current signal acquisition stage, this invention, based on the initial propagation signal acquired at close range, actively adjusts the phase and amplitude relationship through a coupling circuit to realistically simulate the multiple reflection effects of indoor acoustics and reproduce the spatial sound field characteristics of a real drum kit performance. At the same time, it constructs a consensus spectrum component template based on multiple coherent current signals, which significantly improves the resolution and preservation of each frequency component in the frame audio signal spectrum, and can retain effective frequency components to the maximum extent while suppressing noise. In addition, it smooths the fluctuations in the amplitude of frequency components in the time domain, which helps to reduce glitches and frequency component loss caused by acquisition and transmission. Attached Figure Description

[0042] Figure 1 This is a flowchart illustrating a resolution enhancement method based on audio spectrum analysis according to an embodiment of the present invention.

[0043] Figure 2 This is a schematic diagram of the distributed array of pressure sensors in this invention, wherein... Figure 2 (a) is a schematic diagram of a pressure sensor using a rectangular array. Figure 2 (b) is a schematic diagram of a pressure sensor array arranged in a ring with equal spacing. Figure 2 (c) is a schematic diagram of a pressure sensor using a ring array and arranged non-uniformly according to specific directional requirements;

[0044] Figure 3 This is a schematic diagram of the installation of the vibrating diaphragm of the pressure sensor in this invention;

[0045] Figure 4This is a schematic diagram of the pressure sensor in this invention connected in series via a coupling circuit, wherein... Figure 4 (a) is a schematic diagram of two pressure sensors connected in series through a coupling circuit; Figure 4 (b) is a schematic diagram of three pressure sensors connected in series through a coupling circuit;

[0046] Figure 5 This is a schematic diagram of the coupling circuit in the present invention, wherein... Figure 5 (a) Pressure sensor 1, pressure sensor 2, and pressure sensor 3 are connected in series through a coupling circuit, and each outputs a voltage signal. , and Result diagram; Figure 5 (b) Pressure sensor 1, pressure sensor 2, pressure sensor 3, and pressure sensor 4 are connected in series through coupling circuits, and each outputs a voltage signal. , , and The result is illustrated in the diagram.

[0047] Figure Labels

[0048] 1. Silent drumhead; 2. Vibrating diaphragm; 3. Damping; 4. Spring; 5. Fixed plate. Detailed Implementation

[0049] The present invention will be further described below with reference to the accompanying drawings, but this is not intended to limit the present invention in any way. Any modifications or substitutions made based on the teachings of the present invention shall fall within the protection scope of the present invention.

[0050] Example 1:

[0051] Resolution enhancement methods based on audio spectrum analysis, such as Figure 1 As shown, it includes the following steps:

[0052] S1: A pressure sensor is installed inside the mute drum, and the pressure sensor is placed in a distributed array; the pressure sensor converts the vibration of the mute drumhead 1 into a voltage signal output.

[0053] The diaphragm 2 of the pressure sensor is attached to the inside of the silent drum skin 1, and the resistance strain gauge is attached to the inside of the diaphragm 2.

[0054] When the drumhead 1 vibrates, the diaphragm 2 moves, and the resistance strain gauge is stretched or compressed synchronously with the diaphragm 2, causing the resistance value to change.

[0055] The resistance change is converted into a voltage signal change by measuring the bridge circuit, and the voltage signal is output.

[0056] like Figure 2 As shown in the embodiment of the present invention, the distributed array of pressure sensors can be selected as a rectangular array or a ring array according to the shape and size of the silent drum; Figure 2 (a) shows a rectangular array of pressure sensors arranged in a row-column grid structure, with uniform row and column spacing. Figure 2 (b) and Figure 2 (c) shows that the pressure sensors are arranged in a ring array, and the physical positions of all pressure sensors are constrained to one or more closed circular trajectories with the center point of the silent drum as the center. Figure 2 (b) The distribution of pressure sensors in the ring array shown can be equally spaced. Figure 2 (c) shows pressure sensors arranged non-uniformly according to specific directional requirements;

[0057] In this embodiment of the invention, the number of pressure sensors is determined based on the striking force, the material of the drumhead, and the shape and size of the silent drum.

[0058] like Figure 3 As shown, in this embodiment of the invention, the vibrating diaphragm 2 is attached to the inside of the silent drum skin 1, and the vibrating diaphragm 2 is connected to the fixed plate 5 through the spring 4 and the damper 4.

[0059] It should be noted that, in this embodiment, on the one hand, the diaphragm 2 is installed inside the mute drum skin 1, where space is limited, so it cannot vibrate violently; on the other hand, the effective displacement amplitude of the diaphragm 2 needs to be able to cause a change in the resistance value. Therefore, in this invention, the spring 4 and the damper 3 are used together to control the vibration amplitude of the diaphragm 2. By adjusting the values ​​of the spring 4 and the damper 3, the vibration amplitude of the diaphragm 2 is kept within a suitable range. Before formal use, the user can calibrate the values ​​of the spring 4 and the damper 3 to ensure that the loudness of the headphone output audio is within a suitable range.

[0060] S2: Assign a unique number to all pressure sensors. The numbering rule should ensure that pressure sensors that are physically adjacent have consecutive numbers. Add a coupling circuit between each pair of consecutively numbered pressure sensors and connect the voltage signal output by each pressure sensor to the connected coupling circuit to output a current signal with the same frequency as the voltage signal.

[0061] It should be noted that in this embodiment of the invention, a coupling circuit is added between adjacent pressure sensors, and all pressure sensors are connected in series through the coupling circuit.

[0062] like Figure 4 The diagram shows a pressure sensor connected in series via a coupling circuit, where the hollow circle represents the pressure sensor and the solid circle represents the coupling circuit. Figure 4 (a) shows a schematic diagram of two pressure sensors connected in series via a coupling circuit; Figure 4 (b) shows a schematic diagram of the three pressure sensors connected in series via a coupling circuit;

[0063] Figure 5 This is a schematic diagram of the coupling circuit in this invention; Figure 5 In (a), pressure sensor 1, pressure sensor 2, and pressure sensor 3 are connected in series through coupling circuits. The output voltage signal of pressure sensor 1 is... The pressure sensor 2 outputs a voltage signal of The pressure sensor 3 outputs a voltage signal of Measure the resistance respectively The current at the point is used to obtain the output current signal of pressure sensor 1, pressure sensor 2, and pressure sensor 3 after passing through the coupling circuit. , and ;

[0064] Figure 5 The coupled circuit in (a) satisfies the following equation:

[0065] ;

[0066] , and The variable to be solved in the coupled equation is the pressure sensor output current signal through the coupling circuit, which satisfies the following:

[0067] ;

[0068] in, This represents the inductance value of the inductor in a coupled circuit. , and These represent the input excitations respectively. , and The resulting system response , and They represent respectively to , and Calculate time The derivative, and These represent the capacitance values ​​of the capacitors in the coupling circuit. and These represent the resistance values ​​of the resistors in the coupling circuit;

[0069] This coupling circuit can be considered as a response system. , and As input stimulus, , and To obtain the system response, take the time derivative of the system response; this is the required output current signal. , and ;

[0070] We can connect the pressure sensors in series by building a coupling circuit, and then measure the resistance of each sensor. The current at the point is used to obtain the output current signal of pressure sensor 1, pressure sensor 2, and pressure sensor 3 after passing through the coupling circuit. , and Alternatively, the output voltage of the pressure sensor can be... , and Substitute the solutions into the constructed coupled circuit equations and solve for the system response. , and Then through Obtain the output current signal , and ;

[0071] exist Figure 5 In (b), pressure sensor 1, pressure sensor 2, pressure sensor 3, and pressure sensor 4 are connected in series via coupling circuits. The output voltage signal of pressure sensor 1 is... The pressure sensor 2 outputs a voltage signal of The pressure sensor 3 outputs a voltage signal of The pressure sensor 4 outputs a voltage signal of Measure the resistance respectively The current at the point is used to obtain the output current signal of pressure sensors 1, 2, 3, and 4 after passing through the coupling circuit. , , and ;

[0072] Figure 5 The coupled circuit in (b) satisfies the following equation:

[0073] ;

[0074] , , and The variable to be solved in the coupled equation is the pressure sensor output current signal through the coupling circuit, which satisfies the following:

[0075] ;

[0076] in, , , and These represent the input excitations respectively. , , and The resulting system response Indicates to Calculate time The derivative;

[0077] In this coupling circuit, , , and As input stimulus, , , and To obtain the system response, take the time derivative of the system response; this is the required output current signal. , , and ;

[0078] We can connect the pressure sensors in series by building a coupling circuit, and then measure the resistance of each sensor. The current at the point is used to obtain the output current signal of pressure sensors 1, 2, 3, and 4 after passing through the coupling circuit. , , and Alternatively, the output voltage of the pressure sensor can be... , , and Substitute the solutions into the constructed coupled circuit equations and solve for the system response. , , and Then through Obtain the output current signal , , and ;

[0079] When shared When a pressure sensor is connected in series through a coupling circuit;

[0080] For the first pressure sensor, the output current signal after the coupling circuit satisfies:

[0081] ;

[0082] ;

[0083] For the last pressure sensor, the output current signal after the coupling circuit... satisfy:

[0084] ;

[0085] ;

[0086] in, and These represent the input excitations respectively. and The resulting system response and They represent respectively to and Calculate time The derivative, Indicates pressure sensor The output voltage signal, Pressure sensor The output voltage signal;

[0087] Other pressure sensors The output current signal after the coupling circuit satisfy:

[0088] ;

[0089] in, Indicates pressure sensor The output voltage signal, , and These represent the input excitations respectively. , and The resulting system response and These represent pressure sensors. and The output voltage signal, , and They represent respectively to , and Calculate time The derivative;

[0090] Solving the simultaneous equations yields the output current signal of each pressure sensor after passing through the coupling circuit.

[0091] We can do it by... A coupling circuit is built between the pressure sensors to connect them in series, and then the resistance of each sensor is measured. The current at that point is obtained The output current signal of a pressure sensor after passing through a coupling circuit; or the output voltage of the pressure sensor. , ... Substitute the solutions into the constructed coupled circuit equations and solve for the system response. , ... Then, the time is calculated by analyzing the system response. The derivative of the output current signal is obtained. , ... .

[0092] It should be noted that in this embodiment of the invention, multiple coherent voltage signals are acquired through a pressure sensor array, and a coupling circuit is introduced to process the signals. Compared to the input voltage signal, the phase difference of the current signal output by the coupling circuit is significantly increased in the same frequency components. This design, based on the initial propagation signal acquired at close range, effectively simulates the propagation effect of sound under multiple reflections indoors by actively increasing the phase difference, thereby reproducing spatial acoustic characteristics similar to those of a real indoor drum kit performance in the audio signal. Compared to the voltage signal directly output by the pressure sensor, the current signal output by the coupling circuit further enhances the amplitude difference of the same frequency components in the current signal, better distinguishing the same frequency components from different current signals. In addition, the coupling circuit processing does not eliminate existing frequency components, thus avoiding the omission of effective frequency components and providing a more complete and reliable raw data foundation for subsequent signal processing and audio reconstruction.

[0093] S3. Perform spectral analysis on each current signal obtained in step S2 to obtain the spectrum of the current signal. Count the frequency components that appear in the spectrum of all current signals and retain the frequency components that appear more than a threshold to form a consensus frequency component template.

[0094] Each current signal obtained in step S2 is uniformly sampled to obtain discrete current signals; each discrete current signal is converted into the spectrum of the current signal by discrete Fourier transform.

[0095] Specifically, in this embodiment of the invention, for each current signal obtained in step S2 Perform uniform sampling. This indicates that the current signal originates from the first... A pressure sensor obtains discrete current signals. , Indicates the sampling sequence. For each discrete current signal, the total number of samples is [number]. Perform a Discrete Fourier Transform to obtain the spectrum of the current signal. , Represents frequency components;

[0096] Traversing the spectrum of all current signals ( Each frequency component (representing the total number of pressure sensors) Statistical frequency components Number of times If the number of occurrences If the value is greater than the threshold, the frequency component will be... Mark it as a reserved frequency component; otherwise, leave the frequency component as is. The frequency components are marked as to be discarded; all retained frequency components constitute the consensus frequency component template.

[0097] It should be noted that in the implementation of this invention, the threshold value is positively correlated with the total number of pressure sensors, and this threshold value can be manually adjusted by the user according to the actual auditory effect requirements. When a purer and more prominent audio signal is needed (e.g., simulating the individual sound effect of a certain part in a drum kit), the threshold value can be increased accordingly to filter out signal components with higher consistency; while when a richer and more integrated composite sound effect is desired (e.g., simulating the reverberation effect of a multi-part ensemble or group singing), the threshold value can be appropriately decreased to retain more signal characteristics from the source. This design allows users to flexibly adjust the system's signal synthesis strategy to adapt to different playing styles and auditory performance requirements.

[0098] S4: Superimpose all current signals obtained in step S2 to generate an audio signal; divide the audio signal into time-domain segments and divide it into fixed-duration frame audio signals; uniformly sample each frame audio signal to obtain discrete frame audio signals; apply a window function to each discrete frame audio signal to reduce spectral leakage, and convert it into a frequency domain representation through discrete Fourier transform to obtain the spectrum of the frame audio signal.

[0099] The spectrum of each frame audio signal is filtered using the consensus frequency component template obtained in step S3, retaining only the frequency components present in the consensus frequency component template, to obtain the denoised frame audio signal spectrum:

[0100] For the spectrum of each frame of audio signal, iterate through each frequency component. ;if In the consensus frequency component template obtained in step S3, the spectrum of the audio signal of that frame is retained. The amplitude information; if The consensus frequency component template obtained in step S3 is not included in the spectrum of the audio signal of this frame. The amplitude is set to 0;

[0101] Specifically, in this embodiment of the invention, the superimposed audio signal... The audio signals are segmented according to time sequence, into frames of fixed duration. ,in Indicates the frame index; for the first... Frame audio signal Uniform sampling is performed to obtain discrete frame audio signals. ,in This represents the sampling sequence of a discrete frame audio signal. The total number of samples in the discrete frame audio signal;

[0102] For the Discrete frame audio signals Applying a window function to reduce spectral leakage and converting it to a frequency domain representation using a discrete Fourier transform, we obtain the first... Frame audio signal Spectrum , Represents the spectrum Frequency components in;

[0103] For the spectrum Iterate through each frequency component ;if In the consensus frequency component template obtained in step S3, the spectrum of the audio signal of that frame is retained. The amplitude information; if The frequency component is not determined in the consensus frequency component template obtained in step S3. To determine if the audio signal spectrum of this frame is noise or interference. middle The amplitude is set to 0;

[0104] In addition, regarding the spectrum Filtering is performed using a bandpass filter to ensure that the frequency components are preserved. Within the effective frequency range, the bandpass filter Represented as:

[0105] ;

[0106] in, Represents the minimum frequency within the effective frequency range. This indicates the maximum frequency within the effective frequency range.

[0107] It should be noted that in the process of constructing the consensus frequency component template, this step involves statistically analyzing the frequency of occurrence of each frequency component in the spectrum of all current signals. This ensures that the effective frequency components generated by the mute drum throughout the entire striking process are comprehensively covered, guaranteeing their systematic identification and recording. In step S4, the audio signal is segmented into frame audio signals in the time domain, and then each frame is compared and analyzed against the aforementioned consensus frequency component template. This method effectively avoids the problem of missing effective frequency components that may occur in the frequency domain processing of traditional methods, significantly improving the completeness of signal analysis and the accuracy of audio reconstruction.

[0108] S5. Arrange the spectrum of the denoised frame audio signal in the order of time frames and extract each frequency component. The amplitude of the sequence changes over time, and a moving average filter is applied:

[0109] S51: Arrange the spectrum of the denoised frame audio signal in the order of time frames to construct a three-dimensional data structure; the three dimensions of this three-dimensional data structure are the X-axis (time frame dimension), the Y-axis (frequency dimension) and the Z-axis (amplitude dimension).

[0110] S52: In this three-dimensional data structure, for each frequency component Its amplitude value varies with time frame The changing sequence is treated as an independent time-domain signal For each frequency component Time domain information A moving average filter is applied to suppress the time-domain signal. The high-frequency fluctuations were used to obtain a smoothed time-domain signal. :

[0111] For the smoothed time-domain signal The calculation is divided into boundary frames and intermediate frames;

[0112] For boundary frames, when time frames satisfy: or At that time, among them, This represents the window length of the moving average filter. It is an odd number. Represents time-domain signal In the length of the time frame dimension;

[0113] ;

[0114] For intermediate frames, when time frames satisfy: hour,

[0115] ;

[0116] in, Represents the time-domain signal When performing sliding filtering, the leftmost value in the window, Represents the time-domain signal The second value from the left in the window when performing sliding filtering. Represents the time-domain signal The rightmost value in the window when performing sliding filtering. Represents the time-domain signal The starting point of the time frame when performing sliding filtering. Represents the time-domain signal The end point of the time frame when performing sliding filtering.

[0117] S6: Perform an inverse discrete Fourier transform on the spectrum of the filtered frame audio signal to obtain the reconstructed time-domain segment of the frame audio signal; splice all the obtained reconstructed time-domain segments to obtain the reconstructed audio signal; transmit the reconstructed audio signal to the headphones for playback.

[0118] Example 2: The present invention also provides a resolution enhancement system based on audio spectrum analysis, comprising the following six modules:

[0119] Pressure sensor module: The pressure sensors are installed in a distributed array inside the mute drum; the pressure sensors convert the drumhead vibration into a voltage signal output.

[0120] Voltage-to-current conversion module: Pressure sensor numbering; coupling circuits are added between consecutively numbered pressure sensors; voltage signals are converted into current signals through the coupling circuits;

[0121] Consensus Spectrum Component Template Module: Calculates the spectrum of the current signal, counts the number of occurrences of the spectrum components, retains the frequency components whose frequency exceeds the threshold, and forms a consensus frequency component template;

[0122] Audio signal denoising module: superimposes current signals to obtain audio signals; performs truncation, sampling, and discrete Fourier transform on the audio signals to obtain the spectrum of each frame of audio signals; retains the frequency components present in the consensus spectrum component template for the spectrum of the frame audio signals.

[0123] Audio signal time filtering module: Arranges the spectrum of the frame audio signal in the order of time frames, and performs moving average filtering on the sequence of amplitude changes of each frequency component over time;

[0124] Audio signal reconstruction module: Performs discrete Fourier inverse transform on the spectrum of the frame audio signal to obtain the reconstructed time domain segment of the frame audio signal; splices the reconstructed time domain segments of each frame audio signal in time frame order to obtain the reconstructed audio signal.

[0125] It should be noted that the sequence numbers of the above embodiments of the present invention are merely for descriptive purposes and do not represent the superiority or inferiority of the embodiments. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, apparatus, article, or method. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0126] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0127] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A resolution enhancement method based on audio spectrum analysis, characterized in that, Includes the following steps: S1. Install pressure sensors inside the silent drum. The pressure sensors are placed in a distributed array. The pressure sensors convert the drumhead vibration into a voltage signal output. S2. Assign a unique number to all pressure sensors. The numbering rule should ensure that pressure sensors that are physically adjacent have consecutive numbers. Add a coupling circuit between each pair of consecutively numbered pressure sensors and connect the voltage signal output by each pressure sensor to the connected coupling circuit to output a current signal with the same frequency as the voltage signal. S3. Perform spectrum analysis on each current signal obtained in step S2 to obtain the spectrum of the current signal. Count the number of times the frequency components appear in the spectrum of all current signals and retain the frequency components whose frequency exceeds the threshold to form a consensus frequency component template. S4: Superimpose all current signals obtained in step S2 to generate an audio signal; divide the audio signal into time-domain segments and divide it into fixed-duration frame audio signals; uniformly sample each frame audio signal to obtain discrete frame audio signals; apply a window function to each discrete frame audio signal to reduce spectral leakage, and convert it into a frequency domain representation through discrete Fourier transform to obtain the spectrum of the frame audio signal. The spectrum of each frame audio signal is filtered using the consensus frequency component template obtained in step S3, retaining only the frequency components present in the consensus frequency component template, to obtain the denoised frame audio signal spectrum. S5. Arrange the spectrum of the denoised frame audio signal in the order of time frames and extract each frequency component. The amplitude of the sequence changes over time and is then processed by moving average filtering. Step S5 includes: S51: Arrange the spectrum of the denoised frame audio signal in the order of time frames to construct a three-dimensional data structure; the three dimensions of this three-dimensional data structure are X-axis - time frame dimension, Y-axis - frequency dimension and Z-axis - amplitude dimension; S52: In the three-dimensional data structure, for each frequency component Its amplitude value varies with time frame The changing sequence is treated as an independent time-domain signal For each frequency component time domain signal A moving average filter is applied to suppress the time-domain signal. The high-frequency fluctuations were used to obtain a smoothed time-domain signal. ; S6. Perform an inverse discrete Fourier transform on the spectrum of the filtered frame audio signal to obtain the reconstructed time-domain segment of the frame audio signal; splice all the obtained reconstructed time-domain segments to obtain the reconstructed audio signal; transmit the reconstructed audio signal to the headphones for playback.

2. The resolution enhancement method based on audio spectrum analysis according to claim 1, characterized in that, Step S1 includes: The diaphragm of the pressure sensor is attached to the inside of the drumhead, and the resistance strain gauge is attached to the inside of the diaphragm. When the drumhead vibrates, the diaphragm moves, and the resistance strain gauge stretches or compresses synchronously with the diaphragm, causing the resistance value to change. The resistance change is converted into a voltage signal change by measuring the bridge circuit, and then the voltage signal is output.

3. The resolution enhancement method based on audio spectrum analysis according to claim 1, characterized in that, Step S3 includes: Each current signal obtained in step S2 is uniformly sampled to obtain discrete current signals; each discrete current signal is converted into the spectrum of the current signal by discrete Fourier transform.

4. The resolution enhancement method based on audio spectrum analysis according to claim 1, characterized in that, Step S4 includes: For the spectrum of each frame of audio signal, iterate through each frequency component. ;if In the consensus frequency component template obtained in step S3, the spectrum of the audio signal of that frame is retained. The amplitude information; if The consensus frequency component template obtained in step S3 is not included in the spectrum of the audio signal of this frame. The amplitude is set to 0.

5. The resolution enhancement method based on audio spectrum analysis according to claim 1, characterized in that, Step S52 includes: For the smoothed time-domain signal The calculation is divided into boundary frames and intermediate frames; For boundary frames, when time frames satisfy: or At that time, among them, This represents the window length of the moving average filter. It is an odd number. Represents time-domain signal In the length of the time frame dimension; ； For intermediate frames, when time frames satisfy: hour, ； in, Represents the time-domain signal When performing sliding filtering, the leftmost value in the window, Represents the time-domain signal The second value from the left in the window when performing sliding filtering. Represents the time-domain signal The rightmost value in the window when performing sliding filtering. Represents the time-domain signal The starting point of the time frame when performing sliding filtering. Represents the time-domain signal The end point of the time frame when performing sliding filtering.

6. A resolution enhancement system based on audio spectrum analysis, characterized in that, include: Pressure sensor module: The pressure sensors are installed in a distributed array inside the mute drum; the pressure sensors convert the drumhead vibration into a voltage signal output. Voltage-to-current conversion module: pressure sensor numbering; coupling circuits are added between pressure sensors with consecutive numbers; The voltage signal is converted into a current signal through a coupling circuit; Consensus Spectrum Component Template Module: Calculates the spectrum of the current signal, counts the number of occurrences of the spectrum components, retains the frequency components whose frequency exceeds the threshold, and forms a consensus frequency component template; Audio signal denoising module: superimposes current signals to obtain audio signals; performs truncation, sampling, and discrete Fourier transform on the audio signals to obtain the spectrum of each frame of audio signals; retains the frequency components present in the consensus spectrum component template for the spectrum of the frame audio signals. Audio signal time filtering module: Arranges the spectrum of the frame audio signal in the order of time frames, and performs moving average filtering on the sequence of amplitude changes of each frequency component over time; Audio signal reconstruction module: Performs discrete Fourier inverse transform on the spectrum of the frame audio signal to obtain the reconstructed time domain segment of the frame audio signal; splices the reconstructed time domain segments of each frame audio signal in time frame order to obtain the reconstructed audio signal; To achieve the resolution enhancement method based on audio spectrum analysis as described in any one of claims 1-5.

Images