An acoustic calibration method and device based on fast Fourier transform, electronic equipment and storage medium

Abstract

The application relates to an acoustic calibration method and device based on a fast Fourier transform, electronic equipment and a storage medium, the method comprising the following steps: generating a sweep signal and an inverse sweep signal corresponding to the sweep signal; performing filling processing on the sweep signal and the inverse sweep signal, and performing fast Fourier transform on the filled sweep signal and inverse sweep signal to obtain frequency domain signals corresponding to the sweep signal and the inverse sweep signal respectively; establishing a frequency response curve of the frequency domain signals; based on a frequency response feature of the frequency response curve, performing windowing processing on a time domain filter by using a Blackman window function; determining a filter coefficient according to the Blackman window function, and compensating for the influence of an environment on a sound signal according to the filter coefficient, so that the influence of the environment on the sound signal is compensated for, accurate acoustic calibration is realized, the operation speed is improved, the adaptability and accuracy of acoustic calibration are enhanced, and multi-scene requirements are met.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to an acoustic calibration method, apparatus, electronic device and storage medium based on Fast Fourier Transform. Background Technology

[0002] Room calibration technology is a technique that combines acoustics, electronic engineering, and digital signal processing to optimize the performance of an audio system in a specific space. It uses calibration microphones to collect the acoustic response of speakers (such as frequency response, reverberation time, phase difference, etc.) in the listening area. Algorithms are then used to identify room acoustic problems (such as standing waves, reflections, low-frequency buildup, etc.), and digital signal processing (DSP) is used to adjust parameters such as equalizer, delay, and phase to compensate for room acoustic defects and improve sound quality.

[0003] However, existing technologies can cause sound field deformation in multi-channel joint optimization strategies. When the left and right channels share the same phase parameters, the uniformity of sound energy distribution on the horizontal plane is disrupted, resulting in a shrinkage of the stereo image in the listening area. This error stems from the fact that the installation position differences of the speakers in different channels are not independently modeled, and forced phase alignment destroys the unique room transfer function characteristics of each channel. In terms of computing speed, existing technologies have high computational complexity in time-domain convolution, making it difficult to process high sampling rate audio in real time. Furthermore, parameters such as frequency range, maximum gain, and target curve are fixed, making it difficult to adapt to the needs of multiple scenarios.

[0004] Therefore, there is an urgent need to develop an acoustic calibration method, device, electronic device, and storage medium based on Fast Fourier Transform to solve one or more of the aforementioned problems. Summary of the Invention

[0005] In view of this, in order to solve the above-mentioned technical problems or some of the technical problems, the present invention provides an acoustic calibration method, apparatus, electronic device and storage medium based on fast Fourier transform.

[0006] Firstly, this application provides an acoustic calibration method based on Fast Fourier Transform, the method comprising:

[0007] Generate a sweep frequency signal and a corresponding inverse sweep frequency signal;

[0008] The frequency sweep signal and the inverse frequency sweep signal are padded, and a fast Fourier transform is performed on the padded frequency sweep signal and the inverse frequency sweep signal to obtain the frequency domain signals corresponding to the frequency sweep signal and the inverse frequency sweep signal respectively.

[0009] Establish the frequency response curve of the frequency domain signal;

[0010] Based on the frequency response characteristics of the frequency response curve, the Blackman window function is used to perform windowing processing on the time-domain filter in order to suppress the Gibbs phenomenon of the time-domain and frequency-domain signals.

[0011] The filter coefficients are determined based on the Blackman window function, and the environmental influence on the sound signal is compensated based on the filter coefficients to achieve acoustic calibration.

[0012] In one possible implementation, after obtaining the frequency domain signals corresponding to the sweep frequency signal and the inverse sweep frequency signal, the method further includes:

[0013] The frequency domain signal is matched with the microphone standard data to obtain the adjustment parameters of the frequency domain signal;

[0014] The audio system is adjusted using the aforementioned adjustment parameters to achieve sound compensation.

[0015] In one possible implementation, the adjustment parameters include compensation parameters and constraint parameters;

[0016] The adjustment of the audio system using the adjustment parameters includes:

[0017] The audio system is initially adjusted based on the compensation parameters, and the adjustment range is controlled to conform to the constraint parameters during the initial adjustment process;

[0018] Determine whether the audio characteristics of the adjusted audio system meet the preset requirements;

[0019] If the preset requirements are not met, the adjustment steps are repeated on the adjusted audio system to make the audio characteristics of the audio system meet the preset requirements.

[0020] In one possible implementation, the formula for generating the frequency sweep signal includes:

[0021]

[0022] Where s(t) is the instantaneous amplitude of the swept frequency signal, A is the normalized amplitude, f1 is the lowest frequency, T is the signal duration, f2 is the highest frequency, and t is the time variable.

[0023] In one possible implementation, the filling process for the frequency sweep signal and the inverse frequency sweep signal includes:

[0024] The signal lengths of the sweep frequency signal and the inverse sweep frequency signal are filled to a preset length, where the preset length is a power of 2.

[0025] In one possible implementation, establishing the frequency response curve of the frequency domain signal includes:

[0026] An initial frequency response curve for the frequency domain signal is established using Bézier curve interpolation technology. The initial frequency response curve is a third-order continuously differentiable curve.

[0027] Determine whether there are abnormal regions in the initial frequency response curve, including abnormal fluctuation regions and abnormal abrupt change regions;

[0028] If there are abnormal regions in the initial frequency response curve, eliminate the abnormal regions;

[0029] The initial frequency response curve after elimination is homogenized by using a fractional octave smoothing algorithm to obtain the frequency response curve of the frequency domain signal.

[0030] In one possible implementation, the fast Fourier transform formula includes:

[0031]

[0032] Where X[k] is the instantaneous amplitude of the frequency domain signal, and Xeven[k] is the sample with all even indices in X[k]. Xodd[k] is the rotation factor, and Xodd[k] is a sample of all odd indices in X[k].

[0033] Secondly, this application provides an acoustic calibration device based on Fast Fourier Transform, the device comprising:

[0034] The generation module is used to generate a sweep frequency signal and a corresponding inverse sweep frequency signal.

[0035] The filling module is used to fill the sweep frequency signal and the inverse sweep frequency signal, and perform a fast Fourier transform on the filled sweep frequency signal and the inverse sweep frequency signal to obtain the frequency domain signals corresponding to the sweep frequency signal and the inverse sweep frequency signal respectively.

[0036] A module is created to generate the frequency response curve of a frequency domain signal;

[0037] A windowing module is used to perform windowing processing on the time-domain filter using the Blackman window function based on the frequency response characteristics of the frequency response curve, so as to suppress the Gibbs phenomenon of the time-domain and frequency-domain signals.

[0038] The compensation module is used to determine the filter coefficients based on the Blackman window function, and to compensate for the influence of the environment on the sound signal based on the filter coefficients, so as to achieve acoustic calibration.

[0039] Thirdly, this application provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the acoustic calibration method based on fast Fourier transform as described in any embodiment of the first aspect.

[0040] Fourthly, this application also provides a computer storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the acoustic calibration method based on fast Fourier transform as described in any embodiment of the first aspect.

[0041] Compared with the prior art, the technical solution provided in this application has the following advantages: The method provided in this application generates a swept frequency signal and its corresponding inverse swept frequency signal, performs filling processing and fast Fourier transform on these signals, and can efficiently obtain frequency domain signals. Furthermore, it establishes the frequency response curve of the frequency domain signal, and based on the characteristics of the frequency response curve, it uses the Blackman window function to window the time domain filter to effectively suppress the Gibbs phenomenon. Finally, based on the filter coefficients determined by the Blackman window function, it compensates for the influence of the environment on the sound signal, thereby achieving accurate acoustic calibration. This not only improves the calculation speed, but also enhances the adaptability and accuracy of acoustic calibration, meeting the needs of multiple scenarios. Attached Figure Description

[0042] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] One or more embodiments are illustrated by way of example with reference numerals in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0045] Figure 1 A flowchart illustrating an acoustic calibration method based on Fast Fourier Transform provided for an embodiment of this application;

[0046] Figure 2 A flowchart illustrating another acoustic calibration method based on Fast Fourier Transform provided for embodiments of this application;

[0047] Figure 3 A flowchart illustrating the steps of an acoustic alignment device based on Fast Fourier Transform provided in this application embodiment;

[0048] Figure 4 This is a schematic diagram illustrating a step for obtaining a frequency response curve, as provided in an embodiment of this application.

[0049] Figure 5 A schematic diagram of a 3dB gain compensation curve provided for an embodiment of this application;

[0050] Figure 6 A schematic diagram of a 5dB gain compensation curve provided for an embodiment of this application;

[0051] Figure 7 This is a schematic diagram of 1 / 12 octave band smoothing provided in an embodiment of this application;

[0052] Figure 8 This is a schematic diagram of 1 / 13 octave band smoothing provided in an embodiment of this application;

[0053] Figure 9 A schematic diagram illustrating the effect of windowed and windowless filter coefficients provided in the embodiments of this application;

[0054] Figure 10 A schematic diagram of an acoustic calibration device based on fast Fourier transform provided in an embodiment of this application;

[0055] Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0057] The following disclosure provides numerous different embodiments or examples for implementing various structures of the invention. To simplify the disclosure, specific examples of components and arrangements are described below. These are merely examples and are not intended to limit the scope of the invention. Furthermore, reference numerals and / or letters may be repeated in different examples. Such repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed.

[0058] To address the shortcomings of existing technologies, such as the lack of independent modeling of speaker installation locations across different channels and the disruption of unique room transfer function characteristics by forced phase alignment, and the high computational complexity of time-domain convolution operations making real-time processing of high-sampling-rate audio, coupled with the fixed parameters like frequency range, maximum gain, and target curves that hinder adaptability to diverse scenarios, this application provides an acoustic calibration method, apparatus, electronic device, and storage medium based on Fast Fourier Transform (FFT). By generating a swept-frequency signal and its corresponding inverse swept-frequency signal, and performing fill processing and FFT on these signals, the frequency domain signal can be efficiently acquired. Furthermore, the frequency response curve of the frequency domain signal is established, and based on the characteristics of this curve, a Blackman window function is used to window the time-domain filter to effectively suppress the Gibbs phenomenon. Finally, based on the filter coefficients determined by the Blackman window function, the influence of the environment on the sound signal is compensated, thereby achieving accurate acoustic calibration. This not only improves computational speed but also enhances the adaptability and accuracy of acoustic calibration, meeting the needs of multiple scenarios.

[0059] Figure 1 A flowchart illustrating an acoustic calibration method based on Fast Fourier Transform is provided for embodiments of this application, as shown below. Figure 1 As shown, the method specifically includes:

[0060] S101, Generate a sweep frequency signal and a corresponding inverse sweep frequency signal;

[0061] In this embodiment, an exponentially swept frequency signal is generated according to a formula, covering a frequency range of 20Hz to 20kHz. The duration and amplitude are adjusted to suit different sampling rates. Simultaneously, an inverse swept frequency signal is generated for subsequent convolution processing to extract the impulse response.

[0062] S102. The sweep frequency signal and the inverse sweep frequency signal are filled, and the filled sweep frequency signal and the inverse sweep frequency signal are subjected to fast Fourier transform to obtain the frequency domain signals corresponding to the sweep frequency signal and the inverse sweep frequency signal respectively.

[0063] In this embodiment, the swept frequency signal and the inverse swept frequency signal are first padded to ensure that the signal can maintain its integrity and improve the accuracy of subsequent transformation processing. After the padded processing is completed, a Fast Fourier Transform (FFT) is performed on the padded swept frequency signal and the inverse swept frequency signal. The FFT is an efficient algorithm used to convert time-domain signals into frequency-domain signals. Through the FFT, the frequency domain representations of the swept frequency signal and the inverse swept frequency signal can be obtained, which can analyze and process the characteristics of the signal in the frequency domain, such as the intensity and phase information of the frequency components.

[0064] Specifically, padding typically involves adding a certain number of zeros to the end of a signal sequence. This reduces spectral leakage and helps improve the resolution of the Fast Fourier Transform (FFT).

[0065] S103. Establish the frequency response curve of the frequency domain signal;

[0066] In this embodiment, in order to accurately analyze and understand the response characteristics of the signal processing system to signals of different frequencies, a frequency response curve of the frequency domain signal is established. This frequency response curve can intuitively show the trend of system gain or attenuation changing with frequency.

[0067] S104. Based on the frequency response characteristics of the frequency response curve, the Blackman window function is used to perform windowing processing on the time-domain filter to suppress the Gibbs phenomenon of the time-domain and frequency-domain signals.

[0068] In this embodiment, after analyzing the frequency response characteristics shown by the specific frequency response curve, the Blackman window function is used to window the time-domain filter, which effectively suppresses the Gibbs phenomenon that is common in time-domain and frequency-domain signal processing, thereby improving the accuracy and stability of signal processing.

[0069] The Blackman window function is a window function with low sidelobe amplitude and high sidelobe decay rate. Compared with other window functions, such as the rectangular window and Hanning window, the Blackman window function can better suppress spectral leakage and reduce sidelobe interference to the main lobe, thus it is widely used in signal processing. In time-domain filter design, using the Blackman window function to window the filter coefficients can make the filter's frequency response smoother, reduce transition band fluctuations, and improve filter performance.

[0070] This embodiment employs advanced technologies such as Fast Fourier Transform and Blackman window function, enabling it to efficiently process high sampling rate audio and adapt to different frequency ranges, maximum gain, and target curve requirements. It can deliver excellent performance in various scenarios, achieving accurate acoustic calibration in venues such as home theaters, concert halls, and professional recording studios.

[0071] S105. Determine the filter coefficients according to the Blackman window function, and compensate for the influence of the environment on the sound signal according to the filter coefficients to achieve acoustic calibration.

[0072] In this embodiment, the filter coefficients are determined based on the Blackman window function. The filter coefficients can accurately reflect the influence of the environment on the sound signal, and the audio system is compensated and adjusted accordingly. This not only takes into account the physical characteristics of the room, but also combines the specific parameters of the audio system, thereby ensuring the comprehensiveness and accuracy of the acoustic calibration.

[0073] In practice, the compensation module first calculates the filter coefficients based on the Blackman window function. These coefficients represent the amount of compensation needed for the audio signal at different frequencies. Then, the compensation module applies these coefficients to the audio system, adjusting parameters such as equalizer settings, delay, and phase to compensate for the environmental influences on the audio signal.

[0074] The filter coefficients are obtained by defining the ideal frequency response through frequency domain sampling and interpolation, and then through inverse transformation. Since the number of frequency domain sampling points is usually much larger than the filter order, the final filter coefficients are extracted from the central part of the transformation result to obtain the finite-length impulse response that is closest to the ideal response. This impulse response is then multiplied by the Blackman window function to generate the FIR filter coefficients.

[0075] It should be noted that the compensation adjustment process is real-time, and the audio system output will dynamically adjust as the environment changes to always maintain the best sound quality.

[0076] The acoustic calibration method based on Fast Fourier Transform (FFT) provided in this application combines multiple steps, including generating a swept frequency signal and its inverse swept frequency signal, performing FFT, establishing a frequency response curve, applying a Blackman window function, and compensating for filter coefficients. This achieves efficient and accurate calibration of the audio signal, not only improving the processing speed of audio but also significantly enhancing the adaptability and accuracy of acoustic calibration, enabling it to flexibly meet the needs of various complex environments and different audio systems. Furthermore, the real-time compensation adjustment function ensures that the audio system outputs optimal sound quality in any environment, providing users with an ultimate listening experience.

[0077] In an optional embodiment of the present invention, after obtaining the frequency domain signals corresponding to the sweep frequency signal and the inverse sweep frequency signal, the method further includes:

[0078] The frequency domain signal is matched with the microphone standard data to obtain the adjustment parameters of the frequency domain signal;

[0079] The audio system is adjusted using the aforementioned adjustment parameters to achieve sound compensation.

[0080] In this embodiment, the acquired frequency domain signal is first matched with the pre-set microphone standard data to obtain precise parameters for adjusting the frequency domain signal. Then, the audio system is adjusted according to the adjustment parameters to achieve sound compensation of the audio system and ensure that the output sound quality reaches the best state.

[0081] Specifically, in this embodiment, the microphone calibration data is first matched with the impulse response spectrum at frequency points, and the compensation gain is calculated point by point at a resolution of 1 / 6 octave band. The process of matching the frequency domain signal with the pre-set microphone standard data includes the following steps:

[0082] Step 1: Data Preparation

[0083] Acquiring frequency domain signals: Audio signals are acquired using a microphone, and then converted from time-domain signals to frequency-domain signals using algorithms such as Fast Fourier Transform (FFT). For example, in an audio testing scenario, after the microphone acquires the speech signal, it undergoes an FFT transformation to obtain a frequency domain representation with frequency as the horizontal axis and signal amplitude as the vertical axis.

[0084] Define standard specifications: Determine the microphone's standard specifications, such as frequency response range and gain at various frequencies. These specifications are typically provided by the microphone manufacturer or determined according to relevant industry standards. For example, a certain microphone model may have a standard frequency response of 20Hz-20kHz and a standard gain of 0dB at 1kHz.

[0085] Step 2: Frequency Matching

[0086] Frequency range alignment: Check if the frequency range of the acquired frequency domain signal matches the frequency range of the microphone's standard data. If they do not match, adjustments are required. For example, if the acquired signal frequency range is 30Hz-18kHz, while the standard data is 20Hz-20kHz, interpolation or truncation can be used to match the two frequency ranges.

[0087] Frequency-by-frequency comparison: Within the same frequency range, compare the amplitude, phase, and other parameters of the acquired signal with those of the standard data at each frequency point. For example, compare the amplitude value of the acquired signal at 500Hz with the amplitude value of the standard data at the same frequency point, and record the differences.

[0088] Step 3: Calculate the difference

[0089] Amplitude difference calculation: Calculate the difference between the amplitude of the acquired signal and the standard data at each frequency point. For example, if the amplitude of the acquired signal at 1kHz is -3dB and the standard data is 0dB, then the amplitude difference is -3dB.

[0090] Phase difference calculation: Similarly, the phase difference is calculated. Phase difference is particularly important in some applications that are sensitive to the phase of the signal (such as audio phase correction).

[0091] Step 4: Parameter Adjustment Confirmation

[0092] Adjustment based on differences: Based on the calculated amplitude and phase differences, and combined with the characteristics and requirements of the audio system, adjustment parameters are determined. For example, if the amplitude of the signals collected at multiple low-frequency points is lower than the standard data, adjustment parameters to increase the gain in the low-frequency bands can be determined.

[0093] Optimization and Verification: After initially determining the adjustment parameters, conduct simulation or actual tests to verify the adjustment effect. If the ideal match is still not achieved after adjustment, the parameters need to be further optimized until the frequency domain signal and the microphone standard data achieve a satisfactory matching degree at each frequency point.

[0094] The acoustic calibration method based on Fast Fourier Transform provided in this application fully considers the characteristics of the audio system and the needs of actual application scenarios during the frequency matching and adjustment parameter determination process, ensuring the accuracy and effectiveness of sound compensation. Through meticulous frequency matching, it can accurately identify the differences between the frequency domain signal and the microphone standard data, providing reliable data support for subsequent adjustment work. When determining the adjustment parameters, it comprehensively considers parameters such as gain, delay, and phase of the audio system, ensuring the comprehensiveness and scientific nature of the adjustment scheme. At the same time, it verifies the adjustment effect through simulation or actual testing, continuously optimizing the adjustment parameters until the best matching effect is achieved, thereby realizing a significant improvement in the sound quality of the audio system.

[0095] In one optional embodiment of the present invention, the adjustment parameters include compensation parameters and constraint parameters;

[0096] The adjustment of the audio system using the adjustment parameters includes:

[0097] The audio system is initially adjusted based on the compensation parameters, and the adjustment range is controlled to conform to the constraint parameters during the initial adjustment process;

[0098] Determine whether the audio characteristics of the adjusted audio system meet the preset requirements;

[0099] If the preset requirements are not met, the adjustment steps are repeated on the adjusted audio system to make the audio characteristics of the audio system meet the preset requirements.

[0100] In this embodiment, the audio system is initially adjusted according to the described compensation parameters, and it is ensured that the adjustment range strictly follows and conforms to the predetermined constraint parameters during the initial adjustment process. After the initial adjustment is completed, the audio characteristics of the audio system are evaluated and judged to determine whether these characteristics have met the preset requirements. If it is found during the evaluation process that the audio characteristics of the audio system do not meet the preset requirements, then the adjustment steps of the audio system need to be executed again. Through repeated adjustments and optimizations, the audio characteristics of the audio system fully meet the preset requirements.

[0101] In an optional embodiment of the present invention, the formula for generating the frequency sweep signal includes:

[0102]

[0103] Where s(t) is the instantaneous amplitude of the swept frequency signal, A is the normalized amplitude, f1 is the lowest frequency, T is the signal duration, f2 is the highest frequency, and t is the time variable.

[0104] In this embodiment, sweep frequency signals with different frequency ranges can be flexibly generated using the sweep frequency signal generation formula.

[0105] In the generation formula, parameter A controls the normalized amplitude of the signal, ensuring consistent intensity across different frequencies for easier subsequent processing. f1 and f2 define the lowest and highest frequencies of the sweep signal, respectively, covering the audible audio range of 20Hz to 20kHz to meet the needs of various applications. T represents the signal duration; adjusting this parameter adapts to audio systems with different sampling rates, ensuring the sweep signal fully covers the entire frequency range. The time variable t describes how the signal changes over time.

[0106] When generating a sweep signal, the signal starts from the lowest frequency f1 and increases linearly or non-linearly over time to the highest frequency f2, forming a signal with a gradually changing frequency. This signal characteristic allows it to excite all frequency components of the audio system in the frequency domain, thereby comprehensively evaluating the frequency response characteristics of the system. The generation of the inverse sweep signal is similar to that of the sweep signal, but the frequency change direction is reversed, that is, it gradually decreases from the highest frequency f2 to the lowest frequency f1. It is used for subsequent convolution processing with the sweep signal to extract the impulse response of the system.

[0107] By precisely controlling the parameters of the sweep frequency signal, this embodiment can generate a sweep frequency signal that meets specific requirements, laying a solid foundation for subsequent signal processing and acoustic calibration work.

[0108] In an optional embodiment of the present invention, the filling process for the frequency sweep signal and the inverse frequency sweep signal includes:

[0109] The signal lengths of the sweep frequency signal and the inverse sweep frequency signal are filled to a preset length, where the preset length is a power of 2.

[0110] In this embodiment, in order to ensure that the signal lengths of the sweep frequency signal and the inverse sweep frequency signal can match and reach a specific standard, the lengths of the two signals are extended or padded to a preset length value. The preset length value is set to an integer power of 2, which helps to simplify the mathematical operations in the signal processing process, especially when performing the Fast Fourier Transform (FFT) algorithm, which can improve processing efficiency.

[0111] Figure 2 A flowchart illustrating another acoustic calibration method based on Fast Fourier Transform provided in this application embodiment is shown below. Figure 2 As shown, the method specifically includes:

[0112] S201. Using Bézier curve interpolation technology, an initial frequency response curve for the frequency domain signal is established, wherein the initial frequency response curve is a third-order continuously differentiable curve.

[0113] S202. Determine whether there is an abnormal region in the initial frequency response curve, wherein the abnormal region includes abnormal fluctuation region and abnormal abrupt change region;

[0114] S203. If there is an abnormal region in the initial frequency response curve, eliminate the abnormal region;

[0115] S204. The initial frequency response curve after elimination is homogenized by fractional octave smoothing algorithm to obtain the frequency response curve of the frequency domain signal.

[0116] In this embodiment, a Bézier curve interpolation technique is used to construct an initial frequency response curve for the frequency domain signal. This initial frequency response curve has the characteristic of being third-order continuously differentiable, ensuring the smoothness and continuity of the curve. After construction, the initial frequency response curve is checked for any abnormal regions. Abnormal regions may manifest as irregular fluctuations or abrupt changes, which can interfere with signal analysis and processing. Once these abnormal regions are identified in the initial frequency response curve, whether they are abnormal fluctuation regions or abnormal abrupt changes, measures need to be taken to eliminate them, which helps to ensure the accuracy and reliability of subsequent processing. After successfully eliminating abnormal regions, the processed initial frequency response curve is homogenized using a fractional octave smoothing algorithm to further smooth the curve, reduce unnecessary details, and thus obtain a more accurate and stable frequency response curve for the frequency domain signal.

[0117] The frequency response curve obtained through the above steps can accurately reflect the frequency response characteristics of the audio system. Furthermore, based on this frequency response curve, a corresponding compensation signal can be generated to correct the frequency response of the audio system, making it closer to the ideal state.

[0118] The generation of the compensation signal takes into account the amplitude and phase information of each frequency point in the frequency response curve, ensuring accurate compensation for the audio system at each frequency point. After generating the compensation signal, it is applied to the audio system. By adjusting parameters such as the equalizer, delay, and phase of the audio system, the goal of acoustic calibration is achieved, ensuring that the audio system can output stable and accurate sound signals at different frequencies, bringing users a high-quality listening experience.

[0119] The frequency response curve construction process provided in this application includes first generating an initial frequency response curve using Bézier curve interpolation technology. This curve is third-order continuously differentiable, ensuring its smoothness and accuracy. Subsequently, the system performs anomaly detection on the initial frequency response curve, identifying and eliminating abnormal fluctuation and abrupt change regions to ensure the authenticity of the frequency response curve. Finally, a fractional octave band smoothing algorithm is used to homogenize the anomaly-eliminating initial frequency response curve, further smoothing the curve to obtain the final frequency domain signal response curve.

[0120] In an optional embodiment of the present invention, the fast Fourier transform formula includes:

[0121]

[0122] Where X[k] is the instantaneous amplitude of the frequency domain signal, and Xeven[k] is the sample with all even indices in X[k]. Xodd[k] is the rotation factor, and Xodd[k] is a sample of all odd indices in X[k].

[0123] In this embodiment, the time-domain signal can be quickly converted into a frequency-domain signal through the efficient calculation of the Fast Fourier Transform formula, and then frequency-domain analysis can be performed.

[0124] For example, the derivation process of the Fast Fourier Transform formula is as follows:

[0125] By employing FFT / IFFT to perform frequency domain filtering and convolution operations, computational efficiency is significantly improved. The Ooura FFT algorithm is used to optimize Fourier transform performance, resulting in a frequency domain convolution processing speed increase of over 30% compared to traditional methods. Real-time processing at a 48kHz sampling rate is supported. The Discrete Fourier Transform converts a discrete signal x[n] of length N into a frequency domain complex number X[k], as shown in the formula:

[0126]

[0127] in k is the frequency index, n is the time-domain sampling point index, and N is the number of signal sampling points. When N = 2M, M represents the number of levels required to decompose the sequence into powers of 2, i.e., when N is a multiple of 2, the FFT (Fast Fourier Transform) formula is:

[0128]

[0129] in, Even-numbered point DFT;

[0130] For odd-numbered points, DFT;

[0131] Define the rotation factor The formula then simplifies to:

[0132]

[0133] The `xeven` function extracts samples with even indices from the original discrete signal X[k], and the `xodd` function extracts samples with odd indices from the original discrete signal X[k]. The complexity can be calculated to be O(N). 2 The value decreases to O(NlogN).

[0134] Specifically, in the Fast Fourier Transform formula, X[k] represents the instantaneous value of the frequency domain signal, which contains the intensity information of the signal at each frequency component. Xeven[k] and Xodd[k] represent the even-indexed and odd-indexed samples in the frequency domain signal, respectively. They are weighted by a rotation factor to reflect the phase information of the signal in the frequency domain. Through decomposition and weighting, the Fast Fourier Transform can efficiently process large-scale data while maintaining the integrity and accuracy of the signal.

[0135] Figure 3 A flowchart illustrating the steps of an acoustic alignment device based on Fast Fourier Transform provided in this application is shown below. Figure 3 As shown, acoustic calibration includes the following steps:

[0136] Step 1: Generate an exponentially swept frequency signal according to the formula, covering a frequency range of 20Hz to 20kHz. Adapt the duration and amplitude to different sampling rates. Simultaneously generate an inverse swept frequency signal for subsequent convolution processing to extract the impulse response.

[0137]

[0138] Where s(t) represents the instantaneous amplitude of the swept frequency signal in the time domain, t is the time variable from 0 to the duration T, f1 = 20Hz and f2 = 20kHz represent the start frequency and cutoff frequency respectively, T is the signal duration, and A is the normalized amplitude. Different sampling rates can be adapted by adjusting T. For example, when the sampling rate is 48kHz, T must satisfy T ≥ 1 / (2*f2) to avoid spectral aliasing. Inverse swept frequency signal s -1 (t) is generated through time reversal and used in subsequent convolution calculations to extract the system's true impulse response. This method ensures the completeness of subsequent frequency response analysis by covering the full-band excitation.

[0139] Step 2: Improve computational efficiency using an FFT acceleration algorithm. After zero-padding the acquired audio signal and the inverse frequency sweep signal to integer powers of 2, a Fast Fourier Transform is performed. Frequency domain complex multiplication replaces time-domain convolution operations, reducing the computational complexity from O(N^2) to O(N^2). 2 The efficiency is reduced to O(NlogN), with a significant improvement, especially in scenarios with large amounts of data. The impulse response recovered by the inverse transform needs to be truncated to the effective segment and the boundary effect needs to be eliminated.

[0140] Step 3: First, match the microphone calibration data with the impulse response spectrum at different frequency points, and calculate the compensation gain point by point at a 1 / 6 octave resolution. To prevent noise amplification caused by overcompensation in the high-frequency band, dynamic range constraints need to be set, such as limiting the maximum gain to no more than +5dB. By using Bezier curve interpolation technology, a third-order continuously differentiable transition curve is constructed between adjacent frequency points, effectively avoiding the phase jump problem caused by traditional linear interpolation, and ensuring that the compensation curve presents a natural and smooth characteristic in auditory perception.

[0141] Step 4: Smooth the frequency response curve using fractional octave bands. Adjust the smoothing factor to reduce high-frequency fluctuations while preserving the main frequency response characteristics. Apply a Blackman window function to the time-domain filter to suppress the Gibbs phenomenon.

[0142] Step 5: Finally, the window function method is used. After truncating the infinite impulse response of the ideal filter into a finite-length sequence, it is multiplied point by point with the Blackman window function to generate filter coefficients with strictly linear phase characteristics.

[0143] Figure 4 This is a schematic diagram of a frequency response curve acquisition step provided in an embodiment of this application, as shown below. Figure 4 As shown, obtaining the frequency response curve includes the following steps:

[0144] Step 1: First, play the sweep frequency signal, fill the audio signal and the inverse sweep frequency signal acquired by the recording to an integer power of 2 in length, and then perform fast Fourier transform on each of them. After frequency domain convolution, obtain the frequency domain impulse response signal, which is the original measurement curve.

[0145] Step 2: Match the microphone calibration data with the original measurement curve at the same frequency point.

[0146] Step 3: Limit the maximum gain to no more than 5dB, construct the transition curve using Bezier curve interpolation, calculate the compensation signal curve, and obtain the initial frequency response curve.

[0147] Step 4: Perform fractional octave smoothing on the initial frequency response curve to finally obtain the frequency response curve of the frequency domain signal.

[0148] In the above steps, according to the convolution theorem, time-domain convolution is equivalent to frequency-domain complex multiplication. This process includes: two forward FFTs (input signal and inverse frequency sweep signal), one frequency-domain complex multiplication (O(N)), and one inverse FFT (recovering the time-domain result);

[0149] After the above processing, the overall complexity is: 3*O(NlogN) + O(N) ≈ O(NlogN)

[0150] Compared to traditional temporal convolution, which requires N multiplication-accumulation operations (N multiplications and N-1 additions for each output point), resulting in a total number of operations: The processing method in this embodiment greatly reduces computational complexity;

[0151] For example, when N=16384: traditional time-domain convolution requires approximately 268,435,456 operations; FFT frequency-domain convolution of the same number of points requires approximately 147,456 operations.

[0152] An example of the process in a specific application is as follows:

[0153] Step 1: Generate an exponential sweep frequency signal according to the formula, with a frequency range covering 20Hz to 20kHz. By adjusting the duration and amplitude to match different sampling rates, an inverse sweep frequency signal is generated synchronously.

[0154] Step 2: Improve computational efficiency using an FFT acceleration algorithm. After zero-padding the acquired audio signal and the inverse frequency sweep signal to an integer power of 2, perform a Fast Fourier Transform.

[0155] Step 3: First, match the microphone calibration data with the impulse response spectrum at different frequency points, calculating the compensation gain point-by-point at a 1 / 6 octave band resolution. To prevent noise amplification caused by overcompensation in the high-frequency band, dynamic range constraints need to be set, such as limiting the maximum gain to no more than +5dB. Using Bezier curve interpolation, a third-order continuously differentiable transition curve is constructed between adjacent frequency points, effectively avoiding the phase jump problem caused by traditional linear interpolation, ensuring that the compensation curve presents a natural and smooth characteristic in auditory perception. Figure 5 , Figure 6As shown, by comparing the schematic diagrams of gain 3dB and gain 5dB, the compensated gain curve after dynamic range constraint avoids overcompensation in the high-frequency band and effectively suppresses noise amplification.

[0156] Step 4: Smooth the frequency response curve using fractional octave bands. Adjust the smoothing factor to reduce high-frequency fluctuations while preserving the main frequency response characteristics. Apply a Blackman window function to the time-domain filter to suppress the Gibbs phenomenon. For example... Figure 7 , Figure 8 As shown, by comparing the 1 / 12 octave band smoothing diagram and the 1 / 13 octave band smoothing diagram, it can be seen that by adjusting the smoothing factor to reduce high-frequency fluctuations, the main frequency response characteristics can be preserved.

[0157] Step 5: Finally, the process is completed using the window function method. The infinite impulse response of the ideal filter is truncated into a finite-length sequence, and then multiplied point-by-point with the Blackman window function to generate filter coefficients with strictly linear phase characteristics. For example... Figure 9 As shown, the left channel displays the FIR filter generated without a window, while the right channel displays the FIR filter generated with the Blackman window function applied; this illustrates the suppression effect of the Blackman window function on the Gibbs phenomenon.

[0158] Figure 10 A schematic diagram of an acoustic calibration device based on Fast Fourier Transform is provided for an embodiment of this application, as shown below. Figure 10 As shown, the device specifically includes:

[0159] The generation module 1001 is used to generate a sweep frequency signal and a corresponding inverse sweep frequency signal.

[0160] The filling module 1002 is used to fill the sweep frequency signal and the inverse sweep frequency signal, and to perform a fast Fourier transform on the filled sweep frequency signal and the inverse sweep frequency signal to obtain the frequency domain signals corresponding to the sweep frequency signal and the inverse sweep frequency signal respectively.

[0161] Module 1003 is established to create the frequency response curve of the frequency domain signal;

[0162] The windowing module 1004 is used to perform windowing processing on the time-domain filter using the Blackman window function based on the frequency response characteristics of the frequency response curve, so as to suppress the Gibbs phenomenon of the time-domain and frequency-domain signals.

[0163] The compensation module 1005 is used to determine the filter coefficients according to the Blackman window function, and to compensate for the influence of the environment on the sound signal according to the filter coefficients, so as to achieve acoustic calibration.

[0164] In one possible implementation, the filling module 1002 is further configured to perform frequency point matching between the frequency domain signal and the microphone standard data to obtain adjustment parameters for the frequency domain signal; and to use the adjustment parameters to adjust the audio system to achieve sound compensation of the audio system.

[0165] In one possible implementation, the filling module 1002 is further configured to perform preliminary adjustments to the audio system based on the compensation parameters, and control the adjustment range to meet the constraint parameters during the preliminary adjustment process; determine whether the audio characteristics of the adjusted audio system meet the preset requirements; and if the preset requirements are not met, re-execute the adjustment steps on the adjusted audio system so that the audio characteristics of the audio system meet the preset requirements.

[0166] In one possible implementation, the formula for generating the frequency sweep signal includes:

[0167] Where s(t) is the instantaneous amplitude of the swept frequency signal, A is the normalized amplitude, f1 is the lowest frequency, T is the signal duration, f2 is the highest frequency, and t is the time variable.

[0168] In one possible implementation, the filling module 1002 is further configured to fill the signal lengths of the sweep frequency signal and the inverse sweep frequency signal to a preset length, wherein the preset length is a power of 2.

[0169] In one possible implementation, the establishment module 1003 is further configured to establish an initial frequency response curve of the frequency domain signal using Bézier curve interpolation technology, wherein the initial frequency response curve is a third-order continuously differentiable curve; determine whether there are abnormal regions in the initial frequency response curve, wherein the abnormal regions include abnormal fluctuation regions and abnormal abrupt change regions; if there are abnormal regions in the initial frequency response curve, eliminate the abnormal regions; and perform homogenization processing on the eliminated initial frequency response curve using a fractional octave band smoothing algorithm to obtain the frequency response curve of the frequency domain signal.

[0170] In one possible implementation, the fast Fourier transform formula includes:

[0171] Where X[k] is the instantaneous amplitude of the frequency domain signal, and Xeven[k] is the sample with all even indices in X[k]. Xodd[k] is the rotation factor, and Xodd[k] is a sample of all odd indices in X[k].

[0172] The acoustic calibration device based on Fast Fourier Transform provided in this embodiment can be as follows: Figure 10 The acoustic calibration device based on Fast Fourier Transform shown can perform, for example... Figure 1-9All steps of acoustic calibration based on Fast Fourier Transform are then implemented to achieve... Figure 1-9 The technical effect of acoustic calibration based on Fast Fourier Transform is shown below. Please refer to [link / reference] for details. Figure 1-9 The relevant descriptions are presented concisely and will not be elaborated upon here.

[0173] The device embodiments described above are merely illustrative. 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 modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0174] Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application, such as... Figure 11 As shown, this application provides an electronic device including a processor 1101, a communication interface 1102, a memory 1103, and a communication bus 1104. The processor 1101, communication interface 1102, and memory 1103 communicate with each other via the communication bus 1104. The memory 1103 stores computer programs. When the processor 1101 executes the program stored in the memory 1103, it implements the steps of the acoustic calibration method based on Fast Fourier Transform provided in any of the aforementioned method embodiments.

[0175] A sweep frequency signal and its corresponding inverse sweep frequency signal are generated; the sweep frequency signal and the inverse sweep frequency signal are padded, and a fast Fourier transform is performed on the padded sweep frequency signal and the inverse sweep frequency signal to obtain the frequency domain signals corresponding to the sweep frequency signal and the inverse sweep frequency signal respectively; the frequency response curve of the frequency domain signal is established; based on the frequency response characteristics of the frequency response curve, a Blackman window function is used to perform windowing processing on the time domain filter to suppress the Gibbs phenomenon of the time domain frequency domain signal; the filter coefficients are determined according to the Blackman window function, and the influence of the environment on the sound signal is compensated according to the filter coefficients to achieve acoustic calibration.

[0176] In one possible implementation, the frequency domain signal is frequency-matched with microphone standard data to obtain adjustment parameters for the frequency domain signal; the audio system is then adjusted using these adjustment parameters to achieve sound compensation for the audio system.

[0177] In one possible implementation, the audio system is initially adjusted according to the compensation parameters, and the adjustment range is controlled to meet the constraint parameters during the initial adjustment process; it is determined whether the audio characteristics of the adjusted audio system meet the preset requirements; if they do not meet the preset requirements, the adjustment steps are re-executed on the adjusted audio system so that the audio characteristics of the audio system meet the preset requirements.

[0178] In one possible implementation, the formula for generating the frequency sweep signal includes: Where s(t) is the instantaneous amplitude of the swept frequency signal, A is the normalized amplitude, f1 is the lowest frequency, T is the signal duration, f2 is the highest frequency, and t is the time variable.

[0179] In one possible implementation, the signal lengths of the sweep frequency signal and the inverse sweep frequency signal are filled to a preset length, the preset length being an integer power of 2.

[0180] In one possible implementation, an initial frequency response curve for the frequency domain signal is established using Bézier curve interpolation, wherein the initial frequency response curve is a third-order continuously differentiable curve; it is determined whether there are abnormal regions in the initial frequency response curve, wherein the abnormal regions include abnormal fluctuation regions and abnormal abrupt change regions; if there are abnormal regions in the initial frequency response curve, the abnormal regions are eliminated; and the eliminated initial frequency response curve is homogenized using a fractional octave band smoothing algorithm to obtain the frequency response curve of the frequency domain signal.

[0181] In one possible implementation, the fast Fourier transform formula includes: Where X[k] is the instantaneous amplitude of the frequency domain signal, and Xeven[k] is the sample with all even indices in X[k]. Xodd[k] is the rotation factor, and Xodd[k] is a sample of all odd indices in X[k].

[0182] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0183] It should be understood that the terminology used herein is for the purpose of describing particular exemplary embodiments only and is not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “described” as used herein may also include the plural forms. The terms “comprising,” “including,” “containing,” and “having” are inclusive and therefore indicate the presence of the stated features, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, elements, components, and / or combinations thereof. The method steps, processes, and operations described herein are not construed as requiring them to be performed in a particular order described or illustrated unless the order of performance is explicitly indicated. It should also be understood that additional or alternative steps may be used.

[0184] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. An acoustic calibration method based on Fast Fourier Transform, characterized in that, include: Generate a sweep frequency signal and a corresponding inverse sweep frequency signal; The frequency sweep signal and the inverse frequency sweep signal are padded, and a fast Fourier transform is performed on the padded frequency sweep signal and the inverse frequency sweep signal to obtain the frequency domain signals corresponding to the frequency sweep signal and the inverse frequency sweep signal respectively. The frequency domain signal is matched with the microphone standard data to obtain the adjustment parameters of the frequency domain signal; the audio system is adjusted using the adjustment parameters to achieve sound compensation of the audio system; Establish the frequency response curve of the frequency domain signal; Based on the frequency response characteristics of the frequency response curve, the Blackman window function is used to perform windowing processing on the time-domain filter to suppress the Gibbs phenomenon in both the time-domain and frequency-domain signals. The filter coefficients are determined based on the Blackman window function, and the environmental influence on the sound signal is compensated based on the filter coefficients to achieve acoustic calibration.

2. The method according to claim 1, characterized in that, The adjustment parameters include compensation parameters and constraint parameters; The adjustment of the audio system using the adjustment parameters includes: The audio system is initially adjusted based on the compensation parameters, and the adjustment range is controlled to conform to the constraint parameters during the initial adjustment process; Determine whether the audio characteristics of the adjusted audio system meet the preset requirements; If the preset requirements are not met, the adjustment steps are repeated on the adjusted audio system to make the audio characteristics of the audio system meet the preset requirements.

3. The method according to claim 1, characterized in that, The formula for generating the frequency sweep signal includes: ； Where s(t) is the instantaneous amplitude of the swept frequency signal, A is the normalized amplitude, f1 is the lowest frequency, T is the signal duration, f2 is the highest frequency, and t is the time variable.

4. The method according to claim 1, characterized in that, The filling process for the frequency sweep signal and the inverse frequency sweep signal includes: The signal lengths of the sweep frequency signal and the inverse sweep frequency signal are filled to a preset length, where the preset length is a power of 2.

5. The method according to claim 1, characterized in that, The establishment of the frequency response curve of the frequency domain signal includes: An initial frequency response curve for the frequency domain signal is established using Bézier curve interpolation technology. The initial frequency response curve is a third-order continuously differentiable curve. Determine whether there are abnormal regions in the initial frequency response curve, including abnormal fluctuation regions and abnormal abrupt change regions; If there are abnormal regions in the initial frequency response curve, eliminate the abnormal regions; The initial frequency response curve after elimination is homogenized by using a fractional octave smoothing algorithm to obtain the frequency response curve of the frequency domain signal.

6. The method according to claim 1, characterized in that, The formula for the Fast Fourier Transform includes: ； Where X[k] is the instantaneous amplitude of the frequency domain signal, X even [k] represents all samples with even indices in X[k]. X is the rotation factor. odd [k] represents all samples with odd indices in X[k].

7. An acoustic calibration device based on Fast Fourier Transform, characterized in that, include: The generation module is used to generate a sweep frequency signal and a corresponding inverse sweep frequency signal. The filling module is used to fill the sweep frequency signal and the inverse sweep frequency signal, and perform a fast Fourier transform on the filled sweep frequency signal and the inverse sweep frequency signal to obtain the frequency domain signals corresponding to the sweep frequency signal and the inverse sweep frequency signal respectively. The frequency domain signal is matched with the microphone standard data to obtain the adjustment parameters of the frequency domain signal; the audio system is adjusted using the adjustment parameters to achieve sound compensation of the audio system; A module is created to generate the frequency response curve of a frequency domain signal; A windowing module is used to perform windowing processing on the time-domain filter using the Blackman window function based on the frequency response characteristics of the frequency response curve, so as to suppress the Gibbs phenomenon of the time-domain and frequency-domain signals. The compensation module is used to determine the filter coefficients based on the Blackman window function, and to compensate for the influence of the environment on the sound signal based on the filter coefficients, so as to achieve acoustic calibration.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the acoustic calibration method based on Fast Fourier Transform as described in any one of claims 1 to 6.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the acoustic calibration method based on fast Fourier transform as described in any one of claims 1 to 6.

Images