CN106057220B - High-frequency extension method of audio signal and audio player - Google Patents

Abstract

The invention discloses a high-frequency expansion method and an audio player of an audio signal, which are used to realize the high-frequency expansion of an audio signal missing a high-frequency part and meet the user's requirement for high-quality audio. An embodiment of the present invention provides a high-frequency extension method for an audio signal, including: up-sampling an original first audio signal to obtain a second audio signal; acquiring a low-frequency spectrum of the second audio signal, and according to the The low-frequency spectrum estimates the high-frequency spectrum of the second audio signal to obtain a high-frequency spectrum envelope; copy the low-frequency spectrum to the high-frequency spectrum of the second audio signal according to the high-frequency spectrum envelope. frequency band to obtain a third audio signal; and performing energy adjustment on the third audio signal to obtain a fourth audio signal.

Description

A high frequency expansion method of audio signal and audio player

technical field

The invention relates to the technical field of audio processing, and in particular, to a high-frequency expansion method of an audio signal and an audio player.

Background technique

The frequency range of the sound that the human ear can hear is about 2-20 kHz (Kilo Hertz, KHz). Ordinary people are not very sensitive to the high-frequency part of the audio signal. Although the human ear cannot hear the sound with a frequency exceeding 20KHz, Some people can feel it. When storing and transmitting audio files, in order to save space and improve transmission efficiency, it is often necessary to remove high-frequency components that are not very sensitive to the human ear. Although this loses a certain sound quality, it can greatly improve the compression rate of audio files. This trade-off To meet the needs of the market, for example, most of the popular MP3 audio files will remove the high frequency part higher than 16KHz. Although the sound quality has dropped a little, it does not affect the appreciation of audio files by ordinary people, and the audio files take up less storage. Space, MP3 has also become popular with the east wind of the Internet, which greatly meets the needs of ordinary people for music appreciation.

With the improvement of people's consumption level, people's requirements for the quality of audio files are also increasing, and people are more and more fond of high-resolution, complete frequency domain components, and high-fidelity high-quality music. However, such high-quality audio files are often expensive, so it has become a great demand for ordinary users to reasonably expand most of the free audio files that lack high-frequency components to improve the audio playback effect. The audio player in can only play audio files that lack high-frequency components, and cannot perform high-frequency expansion on these audio files.

SUMMARY OF THE INVENTION

Embodiments of the present invention provide a high-frequency expansion method and an audio player for an audio signal, which are used to implement high-frequency expansion of an audio signal missing a high-frequency part to meet user requirements for high-quality audio.

In order to solve the above-mentioned technical problems, the embodiments of the present invention provide the following technical solutions:

In a first aspect, an embodiment of the present invention provides a high-frequency expansion method for an audio signal, including:

Upsampling the original first audio signal to obtain a second audio signal;

acquiring the low-frequency spectrum of the second audio signal, and estimating the high-frequency spectrum of the second audio signal according to the low-frequency spectrum, to obtain a high-frequency spectrum envelope;

Copy the low-frequency spectrum to the high-frequency band of the second audio signal according to the high-frequency spectrum envelope to obtain a third audio signal;

Energy adjustment is performed on the third audio signal to obtain a fourth audio signal.

In a second aspect, an embodiment of the present invention further provides an audio player, including:

an upsampling module for upsampling the original first audio signal to obtain a second audio signal;

a high-frequency estimation module, configured to obtain a low-frequency spectrum of the second audio signal, and perform high-frequency spectrum estimation on the second audio signal according to the low-frequency spectrum to obtain a high-frequency spectrum envelope;

A spectrum copying module, configured to copy the low-frequency spectrum to the high-frequency band of the second audio signal according to the high-frequency spectrum envelope to obtain a third audio signal;

An energy adjustment module, configured to perform energy adjustment on the third audio signal to obtain a fourth audio signal.

As can be seen from the above technical solutions, the embodiments of the present invention have the following advantages:

In the embodiment of the present invention, first up-sampling the original first audio signal to obtain the second audio signal, then acquiring the low-frequency spectrum of the second audio signal, and performing high-frequency spectrum analysis on the second audio signal according to the low-frequency spectrum Then, copy the low-frequency spectrum to the high-frequency band of the second audio signal according to the high-frequency spectrum envelope to obtain the third audio signal, and finally adjust the energy of the third audio signal, A fourth audio signal is obtained. In the embodiment of the present invention, after the first audio signal is up-sampled as the original signal, the high-frequency spectrum envelope of the second audio signal is estimated, and the second audio signal is copied in the high-frequency band and subjected to the low-frequency spectrum before performing energy adjustment, A fourth audio signal can be obtained. The fourth audio signal carries spectrum information of a high frequency band. When the audio player plays the fourth audio signal, the playback effect of the audio signal can be improved.

Description of drawings

In order to illustrate the technical solutions in the embodiments of the present invention more clearly, the following briefly introduces the accompanying drawings used in the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those skilled in the art, other drawings can also be obtained from these drawings.

1 is a schematic block diagram of a flowchart of a method for high-frequency expansion of an audio signal according to an embodiment of the present invention;

2 is a schematic block diagram of a flowchart of another high-frequency expansion method of an audio signal provided by an embodiment of the present invention;

3 is a schematic diagram of an application scenario of up-sampling a first audio signal in an embodiment of the present invention;

4 is a schematic block diagram of a flowchart of another high-frequency expansion method of an audio signal provided by an embodiment of the present invention;

5 is a schematic diagram of a spectral envelope of a second audio signal according to an embodiment of the present invention;

6 is a schematic diagram of a spectrum fitting process according to an embodiment of the present invention;

FIG. 7-a is a schematic diagram of the composition and structure of an audio player provided by an embodiment of the present invention;

FIG. 7-b is a schematic diagram of the composition and structure of an upsampling module according to an embodiment of the present invention;

FIG. 7-c is a schematic diagram of the composition and structure of another audio player provided by an embodiment of the present invention;

7-d is a schematic diagram of the composition and structure of a high-frequency estimation module provided by an embodiment of the present invention;

FIG. 7-e is a schematic structural diagram of another audio player provided by an embodiment of the present invention.

Detailed ways

Embodiments of the present invention provide a high-frequency expansion method and an audio player for an audio signal, which are used to implement high-frequency expansion of an audio signal missing a high-frequency part to meet user requirements for high-quality audio.

In order to make the purpose, features and advantages of the present invention more obvious and understandable, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the following The described embodiments are only some, but not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art fall within the protection scope of the present invention.

The terms "first", "second" and the like in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the terms used in this way can be interchanged under appropriate circumstances, and this is only a way of distinguishing objects with the same attributes when describing the embodiments of the present invention. Furthermore, the terms "comprising" and "having" and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, product or device comprising a series of elements is not necessarily limited to those elements, but may include no explicit or other units inherent to these processes, methods, products, or devices.

Each of them will be described in detail below.

An embodiment of the high-frequency expansion method of an audio signal of the present invention can be applied to the expansion of an audio signal with a missing high-frequency part by an audio player. Please refer to FIG. 1 . The frequency expansion method may include the following steps:

101. Perform up-sampling on the original first audio signal to obtain a second audio signal.

In this embodiment of the present invention, the first audio signal is an original audio file input to the audio player, and the first audio signal may be an audio signal lacking high-frequency parts. If the audio player directly plays the first audio signal, because The frequency domain is incomplete and the playback effect is poor. Therefore, the high-frequency part can be expanded according to the high-frequency expansion method of the audio signal provided by the embodiment of the present invention. In the embodiment of the present invention, the first audio signal is a time domain signal, and the first audio signal may be up-sampled to obtain the second audio signal. symbol, so that a second audio signal can be obtained, and the second audio signal is expanded in the time domain with respect to the original first audio signal.

102. Acquire a low-frequency spectrum of the second audio signal, and perform high-frequency spectrum estimation on the second audio signal according to the low-frequency spectrum to obtain a high-frequency spectrum envelope.

In the embodiment of the present invention, after the first audio signal is converted into the second audio signal, a low-frequency spectrum is first extracted from the second audio signal, where the low-frequency spectrum refers to the existing frequency spectrum in the second audio signal. Spectrum component, the low-frequency spectrum carries spectrum energy information, and content information carried in the second audio signal can be played through the low-frequency spectrum. Usually, the natural logarithm value of the absolute value of the amplitude of the low-frequency spectrum of the second audio signal exhibits a linear decrease from low frequency to high frequency. Therefore, according to the low-frequency spectrum, the second audio signal can be subjected to high-frequency analysis. Estimation of the frequency spectrum, so that the high frequency band spectral envelope of the second audio signal can be generated, and the high frequency frequency spectral envelope is the natural logarithm value of the absolute value of the amplitude of the missing high frequency part in the second audio signal. By estimating the high frequency spectral envelope of the second audio signal, the approximate area of the missing high frequency signal in the second audio signal can be determined.

103. Copy the low-frequency spectrum to the high-frequency frequency of the second audio signal according to the high-frequency spectrum envelope to obtain a third audio signal.

In this embodiment of the present invention, after the high-frequency spectrum envelope is generated from the second audio signal, the low-frequency spectrum can be copied to the high-frequency band of the second audio signal according to the high-frequency spectrum envelope. The second audio signal filled with spectral information may be referred to as a third audio signal. Therefore, in the embodiment of the present invention, filling the low-frequency spectrum in the high-frequency band of the second audio signal can effectively expand the missing high-frequency part in the second audio signal, so that the third audio signal contains both the low-frequency part and the low-frequency part. Including the high frequency part, such a third audio signal can present a high-quality sound quality effect during playback.

104. Perform energy adjustment on the third audio signal to obtain a fourth audio signal.

In this embodiment of the present invention, after the third audio signal is acquired, energy adjustment can also be performed on the third audio signal, because through the spectrum filling of the high frequency band in the preceding steps, the high frequency band of the third audio signal is filled with a low frequency band spectrum, and it is known that the low-frequency spectrum often has greater energy. If the high-energy spectrum continues to be used in the high-frequency band, it will cause harsh sound when the third audio signal is played, and the user feels uncomfortable. Therefore, it is necessary to reduce the first frequency. The energy of the high frequency part of the three audio signals to meet the general needs of users. There are various specific ways of energy adjustment. For example, directly reducing the energy of the middle and high frequency bands of the third audio signal according to a preset proportional relationship, or setting an energy adjustment coefficient, and adjusting the third audio signal according to the energy adjustment coefficient, thereby A fourth audio signal may be generated. Both the fourth audio signal and the third audio signal carry spectrum information of a high frequency band. When the audio player plays the fourth audio signal, the playback effect of the audio signal can be improved, so that the user can enjoy high-quality audio files. In the embodiment of the present invention, the high-frequency extension of the audio signal lacking the high-frequency part can be realized by the coding technology on the audio player side, so that the user does not need to pay expensive fees and only needs to use the audio player provided by the embodiment of the present invention. Enjoying high-quality audio files makes the audio player provided by the embodiment of the present invention highly competitive.

It can be seen from the detailed description of the present invention by the foregoing embodiments that firstly, the original first audio signal is up-sampled to obtain the second audio signal, and then the low-frequency spectrum of the second audio signal is obtained, and the second audio signal is analyzed according to the low-frequency spectrum. Estimate the high-frequency spectrum to obtain the high-frequency spectrum envelope, and then copy the low-frequency spectrum to the high-frequency band of the second audio signal according to the high-frequency spectrum envelope to obtain the third audio signal, and finally compare the third audio signal. Perform energy adjustment to obtain a fourth audio signal. In the embodiment of the present invention, after the first audio signal is up-sampled as the original signal, the high-frequency spectrum envelope of the second audio signal is estimated, and the second audio signal is copied in the high-frequency band and subjected to the low-frequency spectrum before performing energy adjustment, A fourth audio signal can be obtained. The fourth audio signal carries spectrum information of a high frequency band. When the audio player plays the fourth audio signal, the playback effect of the audio signal can be improved.

The foregoing embodiment introduces a high-frequency expansion method of an audio signal provided by the present invention. Next, another embodiment is used to introduce the high-frequency expansion method of an audio signal provided by the present invention. The high-frequency expansion method of an audio signal provided by an embodiment of the invention may include the following steps:

201 . Divide the first audio signal into frames to obtain first audio signals of multiple frames, where the first audio signals of two adjacent frames of the first audio signals of the multiple frames overlap.

In some embodiments of the present invention, the original first audio signal is a time-domain signal, and the first audio signal is first divided into frames. For example, every N bits of the first audio signal can be divided into a frame, so that multiple frames can be obtained. The first audio signals of 1 frame, the first audio signals of multiple frames overlap between the first audio signals of two adjacent frames, and it is assumed that the length of the overlapping area between the first audio signals of two adjacent frames is 2L .

202. Perform discrete cosine transform (Discrete Cosine Transform, DCT) processing on the first audio signal of each frame, to obtain a first audio signal after DCT processing.

In some embodiments of the present invention, the first audio signal for each frame may be sequentially converted to the DCT domain. For example, when DCT transform is performed on each frame, DCT processing can be performed according to matlab. For example, the following formula can be used to calculate the DCT:

Where: ω(k) represents the transformation factor, X(k) represents the first audio signal after the kth DCT processing, x(n) represents the first audio signal before the DCT processing, and N represents the first audio signal of each frame length of the signal.

In some embodiments of the present invention, the transformation factor ω(k) can be calculated by the following formula:

在k为其它取值的情况下，变换因子ω(k)的取值为 Among them, for the first audio signal processed by the first DCT, the value of the transformation factor ω(k) is

In the case where k is other values, the value of the transformation factor ω(k) is

It is not limited that the above transformation factor may also have other value implementation manners, as long as it can be used to calculate X(k).

203. Add 0 to the end of the first audio signal of each frame after DCT processing according to the sampling rate conversion ratio to obtain the first audio signal after 0-filling, and the sampling rate conversion ratio is the ratio of the target sampling rate to the original sampling rate.

In some embodiments of the present invention, after the DCT-processed first audio signal of each frame is obtained, the number of 0s that need to be added to the end of the first audio signal of each frame can be calculated according to a preconfigured sampling rate conversion ratio , so as to obtain the first audio signal after 0-filling. Specifically, the sampling rate conversion ratio can be the ratio of the target sampling rate and the original sampling rate. For example, the sampling rate conversion ratio can be expressed as ^I / _D , where I represents the target sampling rate, D represents the original sampling rate. For example, if an audio file with a sampling rate of 48KHz is upsampled to 192KHz, the original sampling rate is 48KHz, and the target sampling rate is 192KHz.

In some embodiments of the present invention, adding 0 to the end of the first audio signal of each frame after DCT processing can realize the expansion of the first audio signal in the DCT domain, and N spectral lines are obtained after DCT transformation of N sample points. , assuming that the sample rate conversion ratio is ^I / _D , then add N ₁ -N zeros after X(k) to get the upsampled spectrum Y(k):

Wherein N ₁ /N=I/D, for example, the audio with a sampling rate of 44.1KHz is upsampled to 192KHz audio, then N ₁ /N=192/44.1=640/147≈4.35374.

204 . Perform inverse discrete cosine transformation (Inverse Discrete Cosine Transformation, IDCT) processing on the first audio signal after 0-filling, to obtain a first audio signal after IDCT processing.

In some embodiments of the present invention, after 0 is added to the first audio signal, since DCT processing is performed on the first audio signal in step 202, IDCT processing needs to be performed on the first audio signal after 0 is added, so that the first audio signal can be processed by IDCT. The IDCT-processed first audio signal is obtained.

可以得到上采样的时域信号y(n)：An example is as follows, the first audio signal after 0-filling is Y(k), which can be processed by IDCT according to matlab, and Y(k) can be IDCT transformed with the IDCT function, and multiplied by the scaling factor

The upsampled time domain signal y(n) can be obtained:

205. Connect the IDCT-processed first audio signals of all frames in a head-to-tail splicing manner to obtain a second audio signal.

In some embodiments of the present invention, after the IDCT-processed first audio signal of each frame is calculated in the foregoing manner, the IDCT-processed first audio signals of all frames can be obtained, and then two adjacent frames can be analyzed for The first audio signal is spliced end-to-end, and after the IDCT-processed first audio signals of all frames are connected, the second audio signal can be obtained.

It should be noted that, in the foregoing steps of the present invention, the manner of upsampling the original first audio signal to obtain the second audio signal is described in detail. It is not limited that other upsampling may also be adopted in this embodiment of the present invention. way to generate the second audio signal. For example, when performing high-frequency expansion, audio upsampling should be performed first, for example, to upsample audio with a sampling rate of 48KHz to audio of 192KHz, you can use a finite unit impulse response filter (Finite Impulse Response, FIR) method in the time domain. However, compared with the foregoing embodiments of the present invention, this method consumes hardware resources. In the foregoing example, the present invention adopts the method of adding 0 to the DCT frequency domain for upsampling, which saves hardware resources. In addition, the frame overlapping technology is used when up-sampling the long sequence audio signal in the frequency domain. Next, the present invention illustrates how to select an appropriate frame length and overlapping area length so that the total sequence length error is reduced to 0.

In some embodiments of the present invention, step 204 performs inverse discrete cosine transform (IDCT) processing on the first audio signal after 0-filling, and after obtaining the IDCT-processed first audio signal, the high frequency of the audio signal provided by the embodiment of the present invention The extension method can also include the following steps:

A1. Cut off the head end and the tail end of the IDCT-processed first audio signal to obtain the first audio signal with the head and tail ends cut off.

Wherein, when the sequence length of the IDCT-processed first audio signal is very long, in order to avoid the compression or stretching of the up-sampled time-domain signal caused by the overlapping area, the IDCT-processed first audio signal may also be subjected to head-end processing. The first audio signal with the head and tail ends cut off can be obtained, and the head and tail ends of the IDCT processed first audio signal are cut off to avoid compression or stretching of the audio signal.

Further, in some embodiments of the present invention, in step A1, the lengths of the head end and the tail end cut out in the first audio signal after IDCT processing are both L ₁ , and the value of L ₁ is calculated by the following formula get:

Wherein, the value of L ₁ is a positive integer, the length of the overlapping region between the first audio signals of two adjacent frames in the first audio signals of multiple frames is 2L, I represents the target sampling rate, and D represents the original sampling rate . Please refer to FIG. 3 , which is a schematic diagram of an application scenario of up-sampling the first audio signal in an embodiment of the present invention. For example, the first audio signal is used as the original audio sequence. In frame P, frame P+1, and frame P+2, each frame includes N points, and the head and tail of each frame are adjacent to each other. The overlapping area between frames is 2L, the first audio signal after up-sampling includes N ₁ points, the length of the discarded area at the head and tail ends in each frame is L ₁ , and the head and tail of each frame y(n) The L ₁ samples at both ends are removed, where L ₁ =I/D×L, and then each frame is connected end to end to form the final up-sampled long sequence signal.

It should be noted that when choosing to discard the length L ₁ and the overlapping area 2L, no decimals are allowed, even if I×L needs to be an integer multiple of D, for example, when the audio with the sampling rate of 44.1KHz is upsampled to the audio of 192KHz , because the upsampling ratio is I/D=192/44.1=640/147, so to ensure that N ₁ and L ₁ are integers, N and L must be multiples of 147, so you can take the frame length N=2352, L=147 As a set of suitable values, if there is a decimal, the up-sampled time-domain signal will be slightly compressed or stretched, and the longer the sequence, the more serious it is.

In the implementation scenario of performing step A1 in this embodiment of the present invention, step 205 connects the first audio signals processed by IDCT of all frames in a head-to-tail splicing manner to obtain a second audio signal, specifically:

B1. Connect the first audio signals from which the first and last ends of all frames are cut off in a head-to-tail splicing manner to obtain a second audio signal.

That is to say, if the head and tail ends of the first audio signal are cut, then when the first audio signals of all frames are spliced together, the newly generated head and tail ends of the first audio signal need to be spliced. After the first audio signal is spliced, the second audio signal can be generated.

206. Obtain a low-frequency spectrum of the second audio signal, and perform high-frequency spectrum estimation on the second audio signal according to the low-frequency spectrum, to obtain a high-frequency spectrum envelope.

207. Copy the low-frequency spectrum to the high-frequency band of the second audio signal according to the high-frequency spectrum envelope to obtain a third audio signal.

208. Perform energy adjustment on the third audio signal to obtain a fourth audio signal.

In some embodiments of the present invention, after step 205 is performed, steps 206 to 208 may be performed. The implementation of steps 206 to 208 is similar to the implementation of steps 102 to 104 in the foregoing embodiments, and is not repeated here. Repeat.

From the detailed description of the present invention in the foregoing embodiments, it can be seen that in the embodiment of the present invention, the first audio signal is used as the original signal to perform DCT processing, and then 0 is added to realize the expansion of the first audio signal. An audio signal is spliced end-to-end to obtain a second audio signal. The entire upsampling process occupies less hardware resources, and the total sequence length error of the obtained second audio signal can be reduced to 0. The high-frequency spectrum envelope of the second audio signal is estimated, and the second audio signal is copied in the high-frequency band to perform the energy adjustment of the low-frequency spectrum, and a fourth audio signal can be obtained. The fourth audio signal carries the high-frequency spectrum. Spectrum information, when the audio player plays the fourth audio signal, the playback effect of the audio signal can be improved.

The foregoing embodiment introduces a high-frequency expansion method of an audio signal provided by the present invention. Next, another embodiment is used to introduce the high-frequency expansion method of an audio signal provided by the present invention. The high-frequency expansion method of an audio signal provided by an embodiment of the invention may include the following steps:

401. Perform up-sampling on the original first audio signal to obtain a second audio signal.

In some embodiments of the present invention, the implementation of step 401 is similar to the implementation of step 101 in the foregoing embodiments, and details are not described herein again.

402. Divide the second audio signal into frames to obtain second audio signals of multiple frames.

In some embodiments of the present invention, after the second audio signal is obtained, the second audio signal is divided into frames, for example, the second audio signal may be divided into multiple frames, and the second audio signal of each frame is separately executed. The following steps 403 to 405 describe the generation method of the high frequency band spectral envelope.

403. Perform modified discrete cosine transform (Modified Discrete Cosine Transform, MDCT) processing on the second audio signal of each frame to obtain a second audio signal after MDCT processing.

In some embodiments of the present invention, the second audio signal of each frame is converted to the MDCT frequency domain, so as to obtain the second audio signal processed by MDCT, and the conversion from the time domain to the frequency domain can be realized through the transformation method of MDCT. , the second audio signal after MDCT processing is a frequency domain signal including a plurality of spectral information.

404. Use a random sampling test (RANdom Sample Consensus, RANSAC) algorithm to perform straight line fitting on the natural logarithm of the absolute value of the amplitude of the low-frequency spectrum of the second audio signal processed by the MDCT, to obtain a low-frequency spectrum envelope, and a low-frequency spectrum envelope. The spectrum includes: a spectrum segment before the effective cutoff frequency of the MDCT-processed second audio signal.

In some embodiments of the present invention, after obtaining the second audio signal processed by MDCT, the RANSAC algorithm may be used to perform linear fitting on the natural logarithm value of the absolute value of the amplitude of the low frequency spectrum of the second audio signal processed by MDCT, Get the low-band spectral envelope. The low-frequency spectrum can be obtained by calculating the effective cutoff frequency of the second audio signal, and the spectrum segment before the effective cutoff frequency of the MDCT-processed second audio signal can constitute the low-frequency spectrum.

405. Estimate the high-frequency spectrum of the second audio signal processed by the MDCT according to the straight line equation corresponding to the low-frequency spectrum envelope, to obtain the high-frequency spectrum envelope.

In some embodiments of the invention, after calculating the straight line equation corresponding to the spectrum envelope of the low frequency band, the high frequency band spectrum of the second audio signal after MDCT processing can be estimated according to the straight line equation, so as to generate the line equation in the second audio signal. A high-frequency spectrum envelope, the high-frequency spectrum envelope may include a high-frequency part component that needs to be expanded in the second audio signal.

An example is as follows. The audio coding and compression scheme can be completed based on MDCT. As shown in FIG. 5, FIG. 5 is a schematic diagram of the spectral envelope of the second audio signal provided by the embodiment of the present invention. In the MDCT frequency domain, the ln| The X[n]| envelope is roughly linearly decreasing. According to this feature, the missing high-frequency part in the second audio signal also follows this rule. Specifically, the second audio signal obtained after up-sampling (for Time domain signal) is divided into frames, and each frame is converted to the MDCT frequency domain to obtain X[n]. In X[n], a smaller positive value 10 ^-9 can be used to replace 0, so that the subsequent process can calculate the corresponding ln| X[n]|, and then obtain the effective cutoff frequency, and then select the ln|X[n]| values of some spectral lines before the cutoff frequency to perform straight line fitting as the low frequency spectral envelope. In Figure 5, the vertical dotted line represents the spectrum lines of the expanded high-frequency spectrum at multiple frequency points, and the corresponding value of each spectrum line is the natural logarithm of the absolute value of the amplitude corresponding to the frequency point of the spectrum line. The dashed line represents the spectral envelope obtained after straight line fitting. When fitting a straight line, first divide the spectral line before the effective cut-off frequency of the ln|X[n]| curve into several segments, find the maximum value of each segment, and then use the points corresponding to these maximum values to fit the line. The resulting straight line equation is y=a*x+b, where x is the spectral line number, y is the natural logarithm of the absolute value of the spectral line, a is the slope, and b is the intercept. In most cases, the fitted straight line The slope is inverse, i.e. a<0. When using the maximum point of each segment for straight line fitting, the RANSAC algorithm can be used to effectively cope with the influence of outliers. Please refer to FIG. 6, which is a schematic diagram of a spectrum fitting process provided by an embodiment of the present invention. In FIG. 6, the point marked with a "*" sign represents the maximum point in the spectrum grouping, and the point marked with a circle It is the boundary point of the grouping, and the dotted line represents the straight line fitted by the RANSAC method.

In some embodiments of the present invention, step 404 uses the RANSAC algorithm to perform straight line fitting on the low-frequency spectrum of the second audio signal processed by MDCT, and after obtaining the low-frequency spectrum envelope, the audio signal provided by the embodiment of the present invention has a The high-frequency expansion method may further include the following steps:

C1. Determine the correction parameter of the straight line equation according to the preset minimum value point in the spectral coordinate system and the effective value point of the effective cut-off frequency in the spectral coordinate system;

C2. Adjust the parameters of the straight line equation corresponding to the spectrum envelope of the low frequency band according to the correction parameters.

In the above-mentioned embodiment of the present invention, when the slope of the straight line equation corresponding to the spectral envelope of the low frequency band is a positive value, the parameters of the straight line equation can also be corrected. The correction parameter is determined by the minimum point in the spectrum coordinate system and the effective value point of the effective cutoff frequency in the spectrum coordinate system. A slope value can be calculated as the correction parameter through the minimum point and the effective value point. An example is as follows. If the fitted straight line equation is y=a*x+b, in most cases, the fitted straight line has a negative slope, that is, a<0, but in some frames, there will be a>0 In the case of , a and b need to be corrected at this time. The method of correction is to set up a minimum value point (kNum,ln(2^(-MBits))) at the far right end of the spectrum coordinate system, and find the difference between this point and the effective value. The correction parameter determined by the cut-off frequency point (k _c ,ln|X[kc]|), which is used to replace the original parameter a, where kNum is the total number of spectral lines, and MBits is the number of sampling bits, for example, generally 16 or 24 .

Further, in some embodiments of the present invention, the effective cutoff frequency of the second audio signal after MDCT processing can be obtained in the following manner:

D1, read the second audio signal of the pre-buffered T frames, and obtain the spectral lines corresponding to multiple frequency points of the second audio signal of each frame in the T frames, and the value of T is a natural number;

D2. For each of the T frames, determine the effective cut-off frequency of each frame according to the following processing method for the first frame: start the search from the last spectral line of the first frame forward, and find the first spectral line The corresponding frequency is used as the effective cutoff frequency of the first frame, and the first spectral line is the natural logarithm value of the absolute amplitude value of the first one of the spectral lines corresponding to multiple frequency points of the second audio signal of the first frame, which is greater than a predetermined value. Thresholded spectrum line, the first frame is any one of the T frames;

D3. After obtaining the effective cut-off frequency of each of the T frames, determine the maximum value of the effective cut-off frequencies of the T frames as the global cut-off frequency;

D4. Search the spectrum line of the second audio signal processed by MDCT forward from the global cutoff frequency, and find the frequency point corresponding to the first spectrum line whose amplitude absolute value of the natural log value is greater than another preset threshold as the MDCT process Effective cutoff frequency of the second audio signal after.

Among them, the second audio signal of T frames is pre-buffered, for example, T is 25, then the buffered second audio signal of 25 frames can be read first, and the second audio signal of each of the 25 frames can be obtained by MDCT processing. Spectral lines of the second audio signal corresponding to a plurality of frequency points. As shown in FIG. 5 , for the second audio signal of one frame, each vertical line (including the dotted line and the solid line) is a spectral line, and for each of the 25 frames, the effective cutoff of the frame can be determined Frequency, in step D2, take the calculation of the effective cutoff frequency of the first frame as an example, as shown in Figure 5, start the search from the last spectral line of the first frame forward, and find the frequency corresponding to the first spectral line as the first frequency. The effective cutoff frequency of one frame, the first spectral line is the first spectral line whose natural logarithm value of the absolute value of the amplitude is greater than the preset threshold among the spectral lines of the second audio signal of the first frame corresponding to multiple frequency points, That is to say, for all the spectral lines of the first frame, search from the back to the front, and find the spectral line whose natural logarithm value of the absolute value of the amplitude is greater than a preset threshold as the first spectral line, and the frequency point of the first spectral line is the first spectral line. The effective cutoff frequency of one frame, and then select the maximum value from the effective cutoff frequency of 25 frames as the global cutoff frequency. The above preset threshold can be calculated in the following way: ln(2^(-MBits))+q, where q is an adjustable parameter, and MBits is the number of sampling bits. In addition, the preset threshold can also be combined with specific applications The specific value is determined according to the scene, and will not be repeated here. In step D3, an example is given as follows. When calculating the effective cut-off frequency of the second audio signal after MDCT processing of each frame, the effective cut-off frequency of each frame is uniformly calculated from the first frame of all frames, and the calculated The method is to start searching forward from the global cutoff frequency, and find the frequency point of the first spectral line whose amplitude absolute value of the natural logarithm value is greater than another preset threshold as the effective cutoff frequency k _c of the frame, wherein the above Another preset threshold can be calculated as follows: ln(2^(-MBits))+p, where p is an adjustable parameter. The purpose of this processing is to avoid searching forward from the last spectral line in each frame, so as to save search time and improve speed.

It should be noted that, in the foregoing steps of the present invention, the acquisition of the low-frequency spectrum of the second audio signal is performed, and the high-frequency spectrum of the second audio signal is estimated according to the low-frequency spectrum to obtain the high-frequency spectrum envelope. For detailed description, it is not limited that the embodiment of the present invention may also adopt other manners to generate the high-frequency frequency spectrum envelope of the second audio signal. For example, after up-sampling the original audio signal, the audio sequence x[n] is obtained, and then converted to the MDCT frequency domain to obtain X[n], and then _ln |X[n]| is obtained. The spectral envelopes are grouped, and a straight line is then fitted using the least squares method. However, in the foregoing embodiment of the present invention, the RANSAC algorithm is used to perform straight line fitting on the maximum point of each group to obtain a straight line parameter, and then the missing high-frequency part is expanded according to this parameter. Due to the diversity of waveform changes in the frequency domain, the least squares method is easily affected by outliers when fitting a straight line, which leads to a large deviation in the fitted straight line, resulting in poor sound quality after high-frequency expansion. Effectively deal with the interference of outliers and improve the sound quality.

406. Copy the low-frequency spectrum to the high-frequency band of the second audio signal according to the high-frequency spectrum envelope to obtain a third audio signal.

407. Perform energy adjustment on the third audio signal to obtain a fourth audio signal.

In some embodiments of the present invention, after step 405 is performed, steps 406 to 407 may be performed. The implementation of steps 406 to 407 is similar to the implementation of steps 103 to 104 in the foregoing embodiments, and is not repeated here. Repeat.

In some embodiments of the present invention, step 406 copies the low-frequency spectrum to the high-frequency band of the second audio signal according to the high-frequency spectrum envelope to obtain the third audio signal, which may specifically include the following steps:

E1. Divide the low frequency spectrum into a plurality of spectral line segments according to the effective cutoff frequency of the second audio signal;

E2. Copy the plurality of spectral line segments to the high frequency band of the second audio signal in sequence to obtain a third audio signal.

Wherein, the low-frequency spectrum needs to be copied when performing high-frequency expansion. In the embodiment of the present invention, the cyclic mirroring method is adopted to copy the frequency band, so that the junction of the copied frequency band is relatively smooth, which is conducive to the improvement of sound quality. For example, mirror copy is used when copying the low frequency spectrum, and the mirror symmetrical spectral lines are copied as k _c +n*U (U is the number of spectral lines to be copied, n represents the label of the symmetry axis, n=0/1/2 /3...), for example the k _c +1, kc+2,..., k _c +U spectral lines are copied from k _c -1, k _c -2, k _c -U in turn, and then the k _c +U+ 1. k _c +U+2,…,k _c +2U spectral lines are copied from k _c +(U-1), k _c +(U-2),…,k _c +1,k _c , and so on . The purpose of the mirror copy is to make the spectral lines at the junction of the two sides of the copy segment not jump as much as possible, so as to improve the sound quality. It is not limited that, in other embodiments of the present invention, a segment of frequency spectrum may be selected from the low frequency band first, and then copied to the high frequency area multiple times by means of translation until it is filled, but if the difference between the spectral lines at both ends of the copied frequency band is If it is larger, then a jump phenomenon will occur at the junction of the copied frequency bands, which will lead to a decrease in sound quality. In the application scenarios provided by steps E1 to E2, the cyclic mirroring method is used to copy the frequency bands, so that the junction of the copied frequency bands is relatively smooth. , which helps to improve the sound quality.

In some embodiments of the present invention, step 407 performs energy adjustment on the third audio signal to obtain a fourth audio signal, which may specifically include the following steps:

F1. Divide the third audio signal into S spectral line segments according to the effective cutoff frequency and the signal termination frequency of the third audio signal, wherein each spectral line segment includes w spectral lines, and S and w are natural numbers;

The energy of each spectral line in the F2 and S spectral line segments is adjusted as follows:

X'[n]=X[n]×α _i ,

n=k _c +i×w～k _c +(i+1)×w-1, i=0～S-1,

Among them, X'[n] represents the fourth audio signal, X[n] represents the third audio signal, α _i represents the energy adjustment coefficient, E _i represents the energy of each spectral line before adjustment, and P _i represents the second audio signal The pseudo energy filled in the high frequency band of , k _c represents the effective cutoff frequency of the third audio signal, and a and b represent the straight line equation parameters of the third audio signal.

Specifically, after the spectral lines are copied in the aforementioned manner, the effective cutoff frequency k _c to the signal termination frequency end can be divided into S spectral line segments, each spectral line segment includes w spectral lines, and energy adjustment is performed for each segment respectively , it is necessary to first calculate the pre-adjustment energy E _i and pseudo energy P _i of each spectral line, then calculate the energy adjustment coefficient α _i of each section, and finally multiply each spectral line by the corresponding adjustment coefficient to obtain the adjusted spectral line value . It should be noted that the above implementation manner is only a specific and feasible implementation scenario of energy adjustment. It is not limited that the energy adjustment coefficient in the embodiment of the present invention may also adopt other manners, such as further modifying the above energy adjustment coefficient. , set the correction factor. In addition, the implementation manner of the fourth audio signal X'[n] in the embodiment of the present invention may not be limited to the above-mentioned exemplary manner, and the third audio signal may also be modified according to a fixed energy adjustment value. The settings are not limited here.

It should be noted that, in some embodiments of the present invention, MDCT processing is performed on the second audio signal of each frame in step 403 to obtain the second audio signal after MDCT processing, and the first audio signal obtained after performing energy adjustment in step 407 is obtained. The fourth audio signal may also be a frequency domain signal, and the fourth audio signal may be further processed by Inverse Modified Discrete Cosine Transform (IMDCT), and the fourth audio signal after IMDCT processing of each frame is processed. Frame-combining processing is performed to obtain a complete fourth audio signal in the time domain, and the fourth audio signal can be output to the user for audio file playback, so that the user can experience high-quality audio effects.

It can be seen from the detailed description of the present invention in the foregoing embodiments that the RANSAC algorithm can be used when estimating the high frequency spectral envelope of the second audio signal in the embodiment of the present invention, which can effectively deal with the interference of outliers and improve the sound quality. The second audio signal is copied in the high frequency band, and then the energy adjustment is performed on the low frequency spectrum, and a fourth audio signal can be obtained. The fourth audio signal carries the spectrum information of the high frequency band. When the audio player plays the fourth audio signal, it can be Improve the playback of audio signals.

It should be noted that, for the sake of simple description, the foregoing method embodiments are all expressed as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described action sequence. As in accordance with the present invention, certain steps may be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

In order to better implement the above solutions of the embodiments of the present invention, related devices for implementing the above solutions are also provided below.

Referring to FIG. 7-a, an audio player 700 provided by an embodiment of the present invention may include: an up-sampling module 701, a high-frequency estimation module 702, a spectrum copy module 703, and an energy adjustment module 704, wherein,

an upsampling module 701 for upsampling the original first audio signal to obtain a second audio signal;

A high-frequency estimation module 702, configured to acquire a low-frequency spectrum of the second audio signal, and perform high-frequency spectrum estimation on the second audio signal according to the low-frequency spectrum to obtain a high-frequency spectrum envelope;

A spectrum copy module 703, configured to copy the low-frequency spectrum to the high-frequency band of the second audio signal according to the high-frequency spectrum envelope to obtain a third audio signal;

The energy adjustment module 704 is configured to perform energy adjustment on the third audio signal to obtain a fourth audio signal.

In some embodiments of the present invention, as shown in FIG. 7-b, the upsampling module 701 includes:

The first framing module 7011 is used for framing the first audio signal to obtain the first audio signal of multiple frames, the first audio signal of two adjacent frames of the first audio signal of the multiple frames there is overlap between;

The DCT processing module 7012 is used to perform discrete cosine transform DCT processing on the first audio signal of each frame, to obtain the first audio signal after the DCT processing;

The expansion module 7013 is used to add 0 to the tail end of the first audio signal of each frame after the DCT process according to the sampling rate conversion ratio to obtain the first audio signal after the 0 is filled, and the sampling rate conversion ratio is the target sampling rate and the ratio of the original sampling rate;

The ICDT processing module 7014 is configured to perform inverse discrete cosine transform IDCT processing on the first audio signal after 0-filling, to obtain the first audio signal after IDCT processing;

The splicing module 7015 is configured to connect the IDCT-processed first audio signals of all frames in a head-to-tail splicing manner to obtain the second audio signal.

In some embodiments of the present invention, as shown in FIG. 7-c, the audio player 700 may further include the following modules:

The cutting module 705 is used for the IDCT processing module to perform inverse discrete cosine transform (IDCT) processing on the first audio signal after 0-filling, and after obtaining the IDCT-processed first audio signal, the IDCT-processed first audio signal is processed. The head and tail of the signal are cut off to obtain the first audio signal with the head and tail cut off;

In this case, the splicing module 7015 is specifically configured to connect the first audio signals with the first and last ends of all frames cut off in a head-to-tail splicing manner to obtain the second audio signal.

In some embodiments of the present invention, the lengths of the head end and the tail end cut out in the IDCT-processed first audio signal are both L ₁ , and the value of L ₁ is calculated by the following formula:

Wherein, the value of L ₁ is a positive integer, the length of the overlapping region between the first audio signals of two adjacent frames in the first audio signals of the multiple frames is 2L, and the I represents the target sampling rate , the D represents the original sampling rate.

In some embodiments of the present invention, as shown in FIG. 7-d, the high frequency estimation module 702 includes:

The second framing module 7021 is used for framing the second audio signal to obtain the second audio signal of multiple frames;

The MDCT processing module 7022 is used to perform modified discrete cosine transform MDCT processing on the second audio signal of each frame to obtain the second audio signal after MDCT processing;

The straight-line fitting module 7023 is used to perform straight-line fitting on the natural logarithm value of the absolute value of the amplitude of the low-frequency spectrum of the second audio signal processed by the MDCT using a random sampling inspection algorithm, so as to obtain the low-frequency spectrum envelope. The low-frequency spectrum includes: a spectrum segment before the effective cutoff frequency of the second audio signal processed by the MDCT;

The envelope generating module 7024 is configured to estimate the high-frequency spectrum of the second audio signal processed by the MDCT according to the straight line equation corresponding to the low-frequency spectrum envelope, and obtain the high-frequency spectrum envelope.

In some embodiments of the present invention, the effective cutoff frequency of the MDCT-processed second audio signal is obtained in the following manner:

reading the pre-buffered second audio signals of the T frames, and acquiring spectral lines corresponding to multiple frequency points of the second audio signal of each of the T frames, where the value of T is a natural number;

For each of the T frames, the effective cut-off frequency of each frame is determined according to the following processing method for the first frame: start searching from the last spectral line of the first frame forward, and find the first The frequency corresponding to the spectral line is used as the effective cutoff frequency of the first frame, and the first spectral line is the absolute amplitude of the first one of the spectral lines corresponding to multiple frequency points of the second audio signal of the first frame The natural logarithm value of the value is greater than a spectral line of a preset threshold, and the first frame is any one of the T frames;

After obtaining the effective cut-off frequency of each of the T frames, determine the maximum value among the T effective cut-off frequencies of the T frames as the global cut-off frequency;

Starting from the global cutoff frequency, the frequency spectrum line of the second audio signal processed by the MDCT is searched forward, and the frequency point corresponding to the first spectrum line whose natural logarithm value of the absolute value of the amplitude is greater than another preset threshold is found as the frequency point. The effective cutoff frequency of the MDCT-processed second audio signal.

In some embodiments of the present invention, as shown in FIG. 7-e, compared to that shown in FIG. 7-a, the audio player 700 may further include the following modules:

The correction module 706 is used for the straight-line fitting module to perform straight-line fitting on the natural logarithm value of the absolute value of the amplitude of the low-frequency spectrum of the second audio signal processed by the MDCT using a random sampling check algorithm to obtain a low-frequency spectrum envelope After that, the correction parameter of the straight line equation is determined according to the minimum value point in the preset spectral coordinate system and the effective value point of the effective cut-off frequency in the spectral coordinate system; Adjust the parameters of the straight line equation corresponding to the network line.

In some embodiments of the present invention, the spectrum copy module 703 is specifically configured to divide the low frequency spectrum into multiple spectrum line segments according to the effective cutoff frequency of the second audio signal; The third audio signal is obtained by sequentially copying to the high frequency band of the second audio signal.

In some embodiments of the present invention, the energy adjustment module 704 is specifically configured to divide the third audio signal into S spectral line segments according to the effective cutoff frequency and the signal termination frequency of the third audio signal, wherein each One spectral line segment includes w spectral lines, and the S and w are natural numbers; each spectral line in the S spectral line segments is energy adjusted in the following manner:

X'[n]=X[n]×α _i ,

n=k _c +i×w～k _c +(i+1)×w-1, i=0～S-1,

Wherein, the X'[n] represents the fourth audio signal, the X[n] represents the third audio signal, the α _i represents the energy adjustment coefficient, and the E _i represents the adjustment of each spectral line The P _i represents the pseudo energy filled in the high frequency band of the second audio signal, the k _c represents the effective cut-off frequency of the third audio signal, the a and the b represent the The linear equation parameters of the third audio signal are described.

It can be seen from the detailed description of the present invention by the foregoing embodiments that firstly, the original first audio signal is up-sampled to obtain the second audio signal, and then the low-frequency spectrum of the second audio signal is obtained, and the second audio signal is analyzed according to the low-frequency spectrum. Estimate the high-frequency spectrum to obtain the high-frequency spectrum envelope, and then copy the low-frequency spectrum to the high-frequency band of the second audio signal according to the high-frequency spectrum envelope to obtain the third audio signal, and finally compare the third audio signal. Perform energy adjustment to obtain a fourth audio signal. In the embodiment of the present invention, after the first audio signal is up-sampled as the original signal, the high-frequency spectrum envelope of the second audio signal is estimated, and the second audio signal is copied in the high-frequency band and subjected to the low-frequency spectrum before performing energy adjustment, A fourth audio signal can be obtained. The fourth audio signal carries spectrum information of a high frequency band. When the audio player plays the fourth audio signal, the playback effect of the audio signal can be improved.

It should be noted that the information exchange, execution process and other contents between the modules/units of the above device are based on the same concept as the method embodiments of the present invention, and the technical effects brought by them are the same as those of the method embodiments of the present invention, and the specific content can be Refer to the descriptions in the foregoing method embodiments of the present invention, which will not be repeated here.

In addition, it should be noted that the device embodiments described above are only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be A physical unit, which can be located in one place or distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution in this embodiment. In addition, in the drawings of the apparatus embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. Those of ordinary skill in the art can understand and implement it without creative effort.

From the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus necessary general-purpose hardware. Special components, etc. to achieve. Under normal circumstances, all functions completed by a computer program can be easily implemented by corresponding hardware, and the specific hardware structures used to implement the same function can also be various, such as analog circuits, digital circuits or special circuit, etc. However, in many cases a software program implementation is the preferred embodiment for the present invention. Based on such understanding, the technical solutions of the present invention can be embodied in the form of software products in essence or the parts that make contributions to the prior art. The computer software products are stored in a readable storage medium, such as a floppy disk of a computer. , U disk, mobile hard disk, Read-Only Memory (ROM, Read-Only Memory), Random Access Memory (RAM, Random Access Memory), magnetic disk or CD, etc., including several instructions to make a computer device (which can be A personal computer, a server, or a network device, etc.) executes the methods described in the various embodiments of the present invention.

To sum up, the above embodiments are only used to illustrate the technical solutions of the present invention, but not to limit them; although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that it can still be used for The technical solutions described in the above embodiments are modified, or some technical features thereof are equivalently replaced; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims (12)

1. a high frequency expansion method of audio signal, is characterized in that, comprises: Upsampling the original first audio signal to obtain a second audio signal; acquiring the low-frequency spectrum of the second audio signal, and estimating the high-frequency spectrum of the second audio signal according to the low-frequency spectrum, to obtain a high-frequency spectrum envelope; Copy the low-frequency spectrum to the high-frequency band of the second audio signal according to the high-frequency spectrum envelope to obtain a third audio signal; Adjusting the energy of the third audio signal to reduce the energy of the high frequency part of the third audio signal to obtain a fourth audio signal; The acquiring the low-frequency spectrum of the second audio signal, and estimating the high-frequency spectrum of the second audio signal according to the low-frequency spectrum, to obtain a high-frequency spectrum envelope, including: Framing the second audio signal to obtain multiple frames of the second audio signal; The second audio signal of each frame is subjected to modified discrete cosine transform MDCT processing to obtain the second audio signal processed by the MDCT; A random sampling check algorithm is used to perform straight line fitting on the natural logarithm value of the absolute value of the amplitude of the low-frequency spectrum of the second audio signal processed by the MDCT, so as to obtain a low-frequency spectrum envelope, where the low-frequency spectrum includes: the frequency spectrum segment before the effective cut-off frequency of the second audio signal after MDCT processing; The high-frequency spectrum of the second audio signal processed by the MDCT is estimated according to the straight line equation corresponding to the low-frequency spectrum envelope, so as to obtain the high-frequency spectrum envelope. 2. The method according to claim 1, wherein the upsampling of the original first audio signal to obtain the second audio signal comprises: Framing the first audio signal to obtain a first audio signal of multiple frames, where there is overlap between the first audio signals of two adjacent frames in the first audio signals of the multiple frames; Discrete cosine transform DCT processing is performed on the first audio signal of each frame respectively to obtain the first audio signal after the DCT processing; Add 0 to the end of each frame of the first audio signal after the DCT processing according to the sampling rate conversion ratio to obtain the first audio signal after 0-filling, and the sampling rate conversion ratio is the ratio of the target sampling rate and the original sampling rate ; Inverse discrete cosine transform (IDCT) processing is performed on the first audio signal after the complement of 0 to obtain the first audio signal after IDCT processing; The IDCT-processed first audio signals of all frames are connected in a head-to-tail splicing manner to obtain the second audio signal. 3. The method according to claim 2, wherein the described first audio signal after complementing 0 is subjected to inverse discrete cosine transform (IDCT) processing, after obtaining the first audio signal after IDCT processing, the method Also includes: The head end and the tail end of the first audio signal processed by the IDCT are cut off to obtain the first audio signal with the head and tail ends cut off; The first audio signals processed by the IDCT of all frames are connected in a head-to-tail splicing manner to obtain the second audio signals, specifically: The second audio signal is obtained by connecting the first audio signals with the first and last ends of all frames cut off in a head-to-tail splicing manner. 4. The method according to claim 3, wherein the lengths of the head end and the tail end cut off in the first audio signal after the IDCT process are both L ₁ , and the value of the L ₁ is as follows The formula calculates:

Wherein, the value of L ₁ is a positive integer, the length of the overlapping region between the first audio signals of two adjacent frames in the first audio signals of the multiple frames is 2L, and the I represents the target sampling rate , the D represents the original sampling rate. 5. The method according to claim 1, wherein the effective cutoff frequency of the second audio signal processed by the MDCT is obtained by the following manner: reading the pre-buffered second audio signals of the T frames, and acquiring spectral lines corresponding to multiple frequency points of the second audio signal of each of the T frames, where the value of T is a natural number; For each of the T frames, the effective cut-off frequency of each frame is determined according to the following processing method for the first frame: start searching from the last spectral line of the first frame forward, and find the first The frequency corresponding to the spectral line is used as the effective cutoff frequency of the first frame, and the first spectral line is the absolute amplitude of the first one of the spectral lines corresponding to multiple frequency points of the second audio signal of the first frame The natural logarithm value of the value is greater than a spectral line of a preset threshold, and the first frame is any one of the T frames; After obtaining the effective cut-off frequency of each of the T frames, determine the maximum value among the T effective cut-off frequencies of the T frames as the global cut-off frequency; Starting from the global cutoff frequency, the frequency spectrum line of the second audio signal processed by the MDCT is searched forward, and the frequency point corresponding to the first spectrum line whose natural logarithm value of the absolute value of the amplitude is greater than another preset threshold is found as the frequency point. The effective cutoff frequency of the MDCT-processed second audio signal. 6. The method according to claim 1, wherein the random sampling test RANSAC algorithm is used to perform straight line fitting on the low-frequency spectrum of the second audio signal processed by the MDCT to obtain a low-frequency spectrum envelope Afterwards, the method further includes: Determine the correction parameter of the straight line equation according to the minimum value point in the preset spectral coordinate system and the effective value point of the effective cut-off frequency in the spectral coordinate system; Parameter adjustment is performed on the linear equation corresponding to the spectrum envelope of the low frequency band according to the correction parameter. 7. The method according to any one of claims 1 to 6, wherein the copying the low frequency spectrum to the high frequency band of the second audio signal according to the high frequency frequency spectrum envelope, Get the third audio signal, including: dividing the low frequency spectrum into a plurality of spectral line segments according to the effective cutoff frequency of the second audio signal; The plurality of spectral line segments are sequentially copied to the high frequency band of the second audio signal to obtain the third audio signal. 8. The method according to any one of claims 1 to 6, wherein the performing energy adjustment on the third audio signal to obtain a fourth audio signal, comprising: The third audio signal is divided into S spectral line segments according to the effective cutoff frequency and the signal termination frequency of the third audio signal, wherein each spectral line segment includes w spectral lines, and the S and w are natural numbers ; The energy of each spectral line in the S spectral line segments is adjusted in the following manner: X'[n]=X[n]×α _i , n=k _c +i×w～k _c +(i+1)×w-1, i=0～S-1,

Wherein, the X'[n] represents the fourth audio signal, the X[n] represents the third audio signal, the α _i represents the energy adjustment coefficient, and the E _i represents the adjustment of each spectral line The P _i represents the pseudo energy filled in the high frequency band of the second audio signal, the k _c represents the effective cut-off frequency of the third audio signal, the a and the b represent the The linear equation parameters of the third audio signal are described. 9. An audio player, characterized in that, comprising: an upsampling module for upsampling the original first audio signal to obtain a second audio signal; A high-frequency estimation module, configured to acquire the low-frequency spectrum of the second audio signal, and perform high-frequency spectrum estimation on the second audio signal according to the low-frequency spectrum to obtain a high-frequency spectrum envelope; spectrum copy a module, configured to copy the low-frequency spectrum to the high-frequency band of the second audio signal according to the high-frequency spectrum envelope to obtain a third audio signal; an energy adjustment module, configured to perform energy adjustment on the third audio signal, reduce the energy of the high frequency part in the third audio signal, and obtain a fourth audio signal; The high frequency estimation module includes: The second framing module is used for framing the second audio signal to obtain the second audio signal of multiple frames; The MDCT processing module is used to perform modified discrete cosine transform MDCT processing on the second audio signal of each frame to obtain the second audio signal after the MDCT processing; A straight-line fitting module, configured to perform straight-line fitting on the natural logarithm value of the absolute value of the amplitude of the low-frequency spectrum of the second audio signal processed by the MDCT using a random sampling check algorithm to obtain a low-frequency spectrum envelope, the The low-frequency spectrum includes: a spectrum segment before the effective cut-off frequency of the MDCT-processed second audio signal; The envelope generating module is configured to estimate the high-frequency spectrum of the second audio signal processed by the MDCT according to the straight line equation corresponding to the low-frequency spectrum envelope, and obtain the high-frequency spectrum envelope. 10. The audio player according to claim 9, wherein the upsampling module comprises: The first framing module is used for framing the first audio signal to obtain the first audio signal of multiple frames, and the first audio signal of the two adjacent frames of the first audio signal of the multiple frames is between the first audio signals. overlap; The DCT processing module is used to perform discrete cosine transform DCT processing on the first audio signal of each frame, to obtain the first audio signal after the DCT processing; The expansion module is used to add 0 to the tail end of the first audio signal of each frame after the DCT process according to the sampling rate conversion ratio to obtain the first audio signal after the 0 is filled, and the sampling rate conversion ratio is the target sampling rate and The ratio of the original sampling rate; The ICDT processing module is used to perform inverse discrete cosine transform (IDCT) processing on the first audio signal after 0-filling, to obtain the first audio signal after IDCT processing; A splicing module, configured to connect the IDCT processed first audio signals of all frames in a head-to-tail splicing manner to obtain the second audio signal. 11. The audio player of claim 10, wherein the audio player further comprises: The cutting module is used for the IDCT processing module to perform inverse discrete cosine transform IDCT processing on the first audio signal after 0-filling, and after obtaining the IDCT-processed first audio signal, the IDCT-processed first audio signal The head and tail ends are cut off, and the first audio signal with the head and tail ends is cut off; The splicing module is specifically configured to connect the first audio signals with the first and last ends of all frames cut out in the manner of end-to-end splicing to obtain the second audio signal. 12. The audio player of claim 9, wherein the audio player further comprises: The correction module is used for the straight-line fitting module to perform straight-line fitting on the natural logarithm value of the absolute value of the amplitude of the low-frequency spectrum of the second audio signal processed by the MDCT using a random sampling check algorithm, and after obtaining the low-frequency spectrum envelope , determine the correction parameter of the straight line equation according to the minimum value point in the preset spectral coordinate system and the effective value point of the effective cut-off frequency in the spectral coordinate system; Adjust the parameters of the line equation corresponding to the line.

Images