CN101964189B - Audio signal switching method and device - Google Patents

Abstract

The invention provides a voice and audio signal switching method and device. The voice and audio signal switching method includes: when the voice and audio signal is switched, weighting the first high frequency band signal of the current frame of voice and audio signal and the second high frequency band signal of the previous M frames of voice and audio signal, so as to be processed The processed first high frequency band signal; wherein, M is greater than or equal to 1; the processed first high frequency band signal and the first low frequency band signal of the current frame speech and audio signal are synthesized into a broadband signal. Process the first high-band signal of the current frame according to the second high-band signal of the speech and audio signal of the previous M frames, so that the second high-band signal of the speech and audio signal of the previous M frames can smoothly transition to the processed The first high-frequency band signal can smoothly switch voice and audio signals of different bandwidths, improving the quality of the user's listening to voice and audio signals.

Description

Speech and audio signal switching method and device

technical field

Embodiments of the present invention relate to the field of communication technologies, and in particular, to a voice and audio signal switching method and device.

Background technique

At present, during the transmission of voice and audio signals in the network state, due to the different network conditions, the network will truncate the code stream of the voice and audio signal transmitted from the encoding end to the network at different bit rates, so that the decoding end will use the truncated The code stream decodes voice and audio signals with different bandwidths.

In the prior art, due to the different bandwidths of voice and audio signals transmitted in the network, during the transmission of voice and audio signals, there are switching from narrowband voice and audio signals to wideband voice and audio signals, and switching from wideband voice and audio signals to narrowband voice signals. A phenomenon in which audio signals switch. The narrow-band signal mentioned in the present invention is through up-sampling and low-pass filtering, is switched to the wide-band signal that has only low-band components and the high-band components are empty, and the wide-band voice and audio signals have both low-band signal components and High frequency band signal components.

In the process of realizing the present invention, the inventor has found that there are at least the following problems in the prior art: due to the difference in the information in the high-band signal between the narrowband speech and audio signal and the wideband speech and audio signal, when switching speech and audio signals of different bandwidths , there will be a phenomenon of sudden change in the energy of the audio signal, which will cause the user to feel uncomfortable in hearing, and cause the quality of the user to listen to the audio signal to deteriorate.

Contents of the invention

Embodiments of the present invention provide a method and device for switching voice and audio signals, so as to realize smooth switching of voice and audio signals of different bandwidths, so as to improve the quality of audio signals listened to by users.

An embodiment of the present invention provides a voice and audio signal switching method, including:

When the voice and audio signal is switched, the first high frequency band signal of the current frame voice and audio signal and the second high frequency band signal of the previous M frames of voice and audio signal are weighted to obtain the processed first high frequency band signal ; Among them, M is greater than or equal to 1;

Combining the processed first high frequency band signal and the first low frequency band signal of the speech and audio signal of the current frame into a broadband signal.

An embodiment of the present invention provides an audio signal switching device, including:

The processing module is used for weighting the first high-band signal of the current frame of speech and audio signals and the second high-band signal of the previous M frames of speech and audio signals when the speech and audio signals are switched, so as to obtain the processed first A high frequency band signal; wherein, M is greater than or equal to 1;

The first synthesizing module is configured to synthesize the processed first high frequency band signal and the first low frequency band signal of the speech and audio signal of the current frame into a broadband signal.

In the speech and audio signal switching method and device of the embodiment of the present invention, the first high frequency band signal of the speech and audio signal of the current frame is processed according to the second high frequency band signal of the speech and audio signal of the previous M frames, so that the previous M frames The second high-frequency band signal of the speech and audio signal can smoothly transition to the processed first high-frequency band signal, and the processed first high-frequency band signal and the first low-frequency band signal are synthesized into a wide-band signal, thereby switching In the process of voice and audio signals of different bandwidths, it can smoothly switch between voice and audio signals of different bandwidths, which reduces the impact of poor subjective auditory quality of voice and audio signals caused by energy shocks, and improves the quality of users listening to voice and audio signals.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description These are some embodiments of the present invention. For those skilled in the art, other drawings can also be obtained according to these drawings without any creative effort.

Fig. 1 is a flow chart of Embodiment 1 of the voice and audio signal switching method of the present invention;

Fig. 2 is the flow chart of Embodiment 2 of the voice and audio signal switching method of the present invention;

Fig. 3 is a flowchart of an embodiment of step 201 in Fig. 2;

Fig. 4 is a flowchart of an embodiment of step 302 in Fig. 3;

FIG. 5 is a second flowchart of another embodiment of step 302 in FIG. 3;

Fig. 6 is a flowchart of an embodiment of step 202 in Fig. 2;

FIG. 7 is a second flowchart of another embodiment of step 201 in FIG. 2;

FIG. 8 is a third flowchart of another embodiment of step 201 in FIG. 2;

FIG. 9 is a schematic structural diagram of Embodiment 1 of the voice and audio signal switching device of the present invention;

FIG. 10 is a schematic structural diagram of Embodiment 2 of the voice and audio signal switching device of the present invention;

FIG. 11 is a first structural schematic diagram of the processing module in Embodiment 2 of the voice and audio signal switching device of the present invention;

12 is a schematic structural diagram of the first module in Embodiment 2 of the voice and audio signal switching device of the present invention;

Fig. 13a is a second structural schematic diagram of the processing module in the second embodiment of the voice and audio signal switching device of the present invention;

Fig. 13b is a third structural schematic diagram of the processing module in the second embodiment of the voice and audio signal switching device of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments It is a part of embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

FIG. 1 is a flow chart of Embodiment 1 of the voice and audio signal switching method of the present invention. As shown in Figure 1, in the voice and audio signal switching method of this embodiment, when the voice and audio signal is switched, each frame after the switching frame is processed in the following manner:

Step 101, when the speech and audio signals switch, weighting the first high frequency band signal of the current frame speech and audio signal and the second high frequency band signal of the previous M frames of speech and audio signals, to obtain the processed first high frequency band signal A frequency band signal; wherein, M is greater than or equal to 1.

Step 102: Synthesize the processed first high frequency band signal and the first low frequency band signal of the speech and audio signal of the current frame into a wide frequency band signal.

The previous M frames of speech and audio signals in the embodiments provided by the present invention refer to M frames of speech and audio signals before the current frame. The L frames of speech and audio signals before switching refer to the L frames of speech and audio signals before the switching frame when the speech and audio signals are switched. The current speech frame is a wideband signal and the previous speech frame is a narrowband signal; or the current speech frame is a narrowband signal and the previous speech frame is a broadband signal, then the speech and audio signals are switched, and the current speech frame is a switching frame .

In the speech and audio signal switching method of the embodiment of the present invention, the first high frequency band signal in the speech and audio signal of the current frame is processed according to the second high frequency band signal in the speech and audio signal of the previous M frames, so that the first M frames The second high frequency band signal in the speech and audio signal can smoothly transition to the processed first high frequency band signal, so that in the process of switching speech and audio signals of different bandwidths, the high frequency band signals of speech and audio signals of different bandwidths Smooth transition switching is possible; finally, the processed first high-frequency signal and the first low-frequency signal synthesize a wide-band signal, and transmit the wide-band signal to the user terminal, so that the user can enjoy high-quality voice and audio signals. The speech and audio signal switching method of this embodiment can smoothly switch speech and audio signals of different bandwidths, reduces the impact of poor subjective auditory quality of speech and audio signals caused by energy shocks, and improves the quality of speech and audio signals received by users.

FIG. 2 is a flow chart of Embodiment 2 of the voice and audio signal switching method of the present invention. As shown in Figure 2, the voice and audio signal switching method of this embodiment includes:

Step 200: When the speech and audio signal does not switch, synthesize the first high frequency band signal and the first low frequency band signal of the speech and audio signal of the current frame into a broadband signal.

Specifically, the first frequency-band speech and audio signal in this embodiment may be a wide-band speech and audio signal or a narrow-band speech and audio signal. In the transmission process of the speech and audio signal, when the first frequency band speech and audio signal does not switch, the following two situations are processed: 1. If the first frequency band speech and audio signal is a wideband speech and audio signal, the wideband The low-band signal and the high-band signal in the voice-audio signal are synthesized into a broadband signal; two, if the first-band voice-audio signal is a narrow-band voice-audio signal, then the low-band signal and the high-frequency signal in the narrow-band voice-audio signal are The band signal is synthesized into a broadband signal. At this time, although it is a broadband signal, the high frequency band is empty and has no information.

Step 201, when the speech and audio signal switches, perform weighting processing on the first high frequency band signal of the speech and audio signal of the current frame and the second high frequency band signal of the previous M frames of speech and audio signals, so as to obtain the first high frequency band signal after processing band signal. Wherein, M is greater than or equal to 1.

Specifically, when voice and audio signals of different bandwidths are switched, the first high frequency band signal of the voice and audio signal of the current frame is processed according to the second high frequency band signal of the voice and audio signal of the previous M frames, so that the first M The second high-band signal of the frame speech and audio signal can smoothly transition to the processed first high-band signal. For example, when the wideband speech and audio signal is switched to the narrowband speech and audio signal, due to the corresponding The high-band signal is empty, so in order to smoothly switch the wide-band speech and audio signal to the narrow-band speech and audio signal, it is necessary to restore the components of the high-band signal corresponding to the narrow-band speech and audio signal, and when the narrow-band speech and audio signal is switched to For wideband voice and audio signals, since the high frequency band signal in wideband voice and audio signals is not empty, in order to smoothly switch narrowband voice and audio signals to wideband voice and audio signals, it is necessary to weaken the continuous multi-frame wideband voice and audio after switching The energy of the high-frequency band signal in the signal makes the high-band signal of the broadband speech and audio signal gradually transition to the real high-band signal. By step 201, the current frame of voice and audio signal is processed, so that the high frequency band signal in the voice and audio signals of different bandwidths can smoothly transition, and when switching between the wideband voice and audio signal and the narrowband voice and audio signal, The user's hearing is uncomfortable due to the sudden energy change, so that the user receives high-quality audio signals. Wherein, in order to simplify the process of obtaining the processed first high-band signal, the first high-band signal and the second high-band signal of the previous M frames of speech and audio signals can be directly weighted, and the obtained result after processing is is the processed first high frequency band signal.

Step 202: Synthesize the processed first high-frequency signal and the first low-frequency signal of the speech and audio signal of the current frame into a wide-band signal.

Specifically, after the current frame speech and audio signal is processed by step 201, the second high frequency band signal of the previous M frames of speech and audio signal can be smoothly transitioned to the processed first high frequency band signal of the current frame, and then by step 202 Synthesize the processed first high-frequency signal and the first low-frequency signal of the current frame voice and audio signal into a broadband signal, so that the voice and audio signals received by the user are all broadband voice and audio signals, realizing voice and audio with different bandwidths The smooth switching of the signal is conducive to improving the quality of the user's listening to the audio signal.

In the speech and audio signal switching method of the embodiment of the present invention, the first high frequency band signal in the speech and audio signal of the current frame is processed according to the second high frequency band signal of the speech and audio signal of the previous M frames, so that the speech of the previous M frames The second high frequency band signal of the audio signal can smoothly transition to the processed first high frequency band signal, so that the high frequency band signals of the voice and audio signals of different bandwidths can be smoothed during the process of switching voice and audio signals of different bandwidths Transition switching; Finally, the processed first high-frequency signal and the first low-frequency signal synthesize a wide-band signal, and the wide-band signal is transmitted to the user terminal, so that the user can enjoy high-quality voice signals. The speech and audio signal switching method of this embodiment can smoothly switch speech and audio signals of different bandwidths, which reduces the impact of poor subjective auditory quality of audio signals caused by energy shocks, and improves the quality of audio signals listened to by users. In addition, by synthesizing the first high frequency band signal and the first low frequency band signal of the current frame voice and audio signal into a broadband signal when switching between voice and audio signals of different bandwidths does not occur, users can obtain high-quality audio signals.

Based on the above technical solution, optionally, when the wideband voice and audio signal is converted to the narrowband voice and audio signal, as shown in Figure 3, step 201 in this embodiment includes:

Step 301. Predict predicted fine structure information and predicted envelope information corresponding to the first high-frequency band signal.

Specifically, the speech and audio signal can be decomposed into fine structure information and envelope information, so that the speech and audio signal can be restored according to the fine structure information and the envelope information. In the process of switching from wideband speech and audio signals to narrowband speech and audio signals, since there are only low-band signals in the narrowband speech and audio signals, the corresponding high-band signals are empty. In order to make the wideband speech and audio signals smooth To switch to the narrow-band speech and audio signals, it is necessary to restore the high-band signals required by the current narrow-band speech and audio signals, so as to realize smooth switching of the speech and audio signals. Step 301 in this embodiment is to predict the predicted fine structure information and the predicted envelope information corresponding to the first high-band signal in the narrow-band speech and audio signal.

In order to more accurately predict the predicted fine structure information and the predicted envelope information corresponding to the current frame speech and audio signal, step 301 can also perform signal classification on the first low frequency band signal of the current frame speech and audio signal; then according to the first low frequency band The signal type prediction corresponding to the signal corresponds to the predicted fine structure information and the predicted envelope information corresponding to the first high frequency band signal. For example, the narrow-band speech and audio signal of the current frame can be information types such as harmonic signals, or non-harmonic signals or transient signals, and then it can be known that the signal of this type should be based on the information type corresponding to the narrow-band speech and audio signal. It has fine structure information and envelope information, so as to more accurately predict the fine structure information and envelope information corresponding to the high frequency band signal of the current frame. Wherein, the speech and audio signal switching method of the present invention does not limit the signal type of the narrowband speech and audio signal.

Step 302: Perform weighting processing on the predicted envelope signal and the envelope information of the previous M frames corresponding to the second high-band signal of the previous M frames of speech and audio signals, so as to obtain the first envelope information corresponding to the first high-band signal .

Specifically, after predicting the predicted fine structure information and predicted envelope information corresponding to the first high-frequency band signal of the current frame in step 301, the second high-frequency The envelope information of the first M frames corresponding to the band signal is used to generate the first envelope information corresponding to the first high frequency band signal.

Specifically, the process of generating the first envelope information corresponding to the first high-frequency band signal in step 302 can be realized in the following two ways, specifically as follows:

Way 1, as shown in Figure 4, an embodiment of obtaining the first envelope information through step 302 may include:

Step 401. Calculate the correlation coefficient between the first low-frequency signal and the low-frequency signal of the speech and audio signal of the previous N frames according to the first low-frequency signal and the low-frequency signal of the speech and audio signal of the previous N frames; wherein, N greater than or equal to 1.

Specifically, compare the first low frequency band signal of the speech and audio signal of the current frame with the low frequency band signal of the speech and audio signal of the previous N frames to obtain the first low frequency band signal of the speech and audio signal of the current frame and the speech and audio signals of the previous N frames. The correlation coefficient between the low-frequency band signals of the audio signal, for example, can be determined by judging a certain frequency band information in the first low-frequency band signal of the current frame speech and audio signal, and the same frequency band of the low-frequency band signal of the speech and audio signal of the previous N frames The energy size of the information or the difference in the type of information, etc., determine the correlation between them, and calculate the desired correlation coefficient. Wherein, the speech and audio signals of the first N frames may be narrowband speech and audio signals, wideband speech and audio signals, or a mixed signal composed of narrowband speech and audio signals and wideband speech and audio signals.

Step 402, judging whether the correlation coefficient is within a given first threshold range.

Specifically, after the correlation coefficient is calculated in step 401, it is judged whether the correlation coefficient is within a given threshold range. The function of calculating the correlation coefficient is to know whether the speech and audio signal of the current frame is gradually changed or mutated from the speech and audio signal of the previous N frames, that is to say, to see whether their characteristics are the same, and then to judge and predict the speech and audio signal of the current frame The weight of the high-band signal in the previous frame when the high-band signal is used. For example, if the first low frequency band signal of the speech and audio signal of the current frame is equivalent to the low frequency signal energy of the speech and audio signal of the previous frame, and the type is the same, then the speech and audio signal of the previous frame is compared with the speech and audio signal of the current frame. High correlation, therefore, in order to accurately restore the first envelope information corresponding to the current frame speech and audio signal, when restoring the first envelope information corresponding to the current frame speech and audio signal, the high frequency corresponding to the speech and audio signal of the previous frame Envelope information with envelope information or transition envelope information takes a larger weight; otherwise, if the energy of the first low-frequency band signal of the current frame of speech and audio signal is very different from that of the speech and audio signal of the previous frame, and the types are different , it means that the speech and audio signal of the previous frame has a low correlation with the speech and audio signal of the current frame. Therefore, in order to accurately restore the first envelope information corresponding to the speech and audio signal of the current frame, restore the corresponding During the first envelope information, the high-frequency band envelope information corresponding to the speech and audio signal of the previous frame or the transitional envelope information accounted for a smaller weight;

Step 403 , if the correlation coefficient is not within the given first threshold range, perform weighting processing according to the set first weight 1 and first weight 2 to calculate the first envelope information. Wherein, the first weight one is the weight value of the envelope information of the previous frame corresponding to the high frequency band signal of the voice and audio signal of the previous frame, and the first weight two is the weight value of the predicted envelope information.

Specifically, when step 402 shows that the correlation coefficient is not within the given first threshold range, it can be known that the speech and audio signal of the current frame has a small correlation with the speech and audio signal of the previous N frames, so the previous M frames The first M frame envelope information corresponding to the speech and audio signal of the first frequency band or the envelope information of the transition or the high frequency band envelope information corresponding to the speech and audio signal of the previous frame have little influence on the first envelope information, and the current When the first envelope information corresponding to the frame speech and audio signal, the first M frame envelope information corresponding to the first frequency band speech and audio signal of the previous M frames or the transitional envelope information or the high frequency band corresponding to the speech and audio signal of the previous frame Envelope information takes less weight. Therefore, according to the set first weight 1 and first weight 2, the first envelope information of the current frame can be calculated. Wherein, the first weight one is the weight value of the envelope information corresponding to the high-band signal of the previous frame of speech and audio signal, and the speech and audio signal of the previous frame can be a wideband speech and audio signal or a processed narrowband speech signal. For the audio signal, when switching for the first time, the speech and audio signal of the previous frame is the broadband speech and audio signal; and the first weight 2 is the weight value of the predicted envelope information. The product of the predicted envelope information and the first weight two is added to the sum of the product of the envelope information of the previous frame and the first weight one, and the obtained weighted sum is the first envelope information of the current frame. In addition, the speech and audio signals to be transmitted later will be restored according to the method and weight, and the first envelope information corresponding to the speech and audio signals will be restored until the speech and audio signals are switched again.

Step 404, if the correlation coefficient is within the first threshold range, perform weighting processing according to the set second weight 1 and second weight 2 to calculate the transition envelope information. Wherein, the second weight one is the weight value of the envelope information before switching, and the second weight two is the weight value of the envelope information of the previous M frames; wherein, L is greater than or equal to 1.

Specifically, when step 402 shows that the correlation coefficient is within a given threshold range, it can be known that the speech and audio signals of the current frame are similar to the speech and audio signals of the previous N consecutive frames, and the speech and audio signals of the current frame correspond to the first The envelope information is greatly affected by the envelope information of the voice and audio signals of the previous N consecutive frames. At the same time, considering the authenticity of the envelope of the previous M frames, it is necessary to solve the problem based on the envelope information of the previous M frames and the envelope information before switching The transition envelope information corresponding to the current frame speech and audio signal, when recovering the first envelope information of the current frame speech and audio signal, allow the envelope information of the previous M frame envelope information and the envelope information of the L frame before switching to account for a larger weight; The first envelope information is solved by the transition envelope information. Wherein, the second weight one is the weight value of the envelope information before switching, and the second weight two is the weight value of the envelope information of the previous M frames. Then, the product of the envelope information before switching and the second weight 1 is added to the sum of the products of the envelope information of the previous M frames and the second weight 2, and the obtained weight value is the transition envelope information.

Step 405: Decrease the second weight one by the unit of the first weight step, and increase the second weight two by the unit of the first weight step.

Specifically, with the transmission of voice and audio signals, the subsequent narrowband voice and audio signals are gradually less affected by the wideband voice and audio signals before switching. In order to make the calculated first envelope information more accurate, it is necessary to The second weight one and the second weight two are adjusted for applicability. Since the follow-up audio signal is gradually reduced by the influence of the broadband speech and audio signal of L frames before the switch, the value of the second weight 1 gradually decreases, while the value of the second weight 2 gradually increases, thereby weakening the value before the switch. The influence of the envelope information on the first envelope information. Among them, step 405 can modify the second weight 1 and the second weight 2 through the following method: the new second weight 1 is equal to the old second weight 1 minus the first weight step size, and the new second weight 2 is equal to the old The second weight of 2 plus the first weight step; wherein, the first weight step is a set value.

Step 406 , judging whether the set third weight one is greater than the first weight one.

Specifically, the third weight one is a weight value of the transition envelope information, and by comparing the third weight one with the second weight one, it can be known how much the first envelope information of the current frame is affected by the transition envelope information. Among them, the transition envelope information is calculated from the envelope information of the previous M frames and the envelope information before switching, therefore, the third weight one actually represents the influence degree of the first envelope information by the envelope information before switching .

Step 407. If the third weight one is not greater than the first weight one, perform weighting processing according to the set first weight one and first weight two to calculate the first envelope information.

Specifically, when it is judged in step 406 that the third weight 1 is less than or equal to the first weight 1, it means that the speech and audio signal of the current frame is far from the speech and audio signal of L frames before switching, and the first envelope information is mainly affected by the envelope of the previous M frames. Therefore, according to the set first weight 1 and first weight 2, the first envelope information of the current frame can be calculated.

Step 408, if the third weight 1 is greater than the first weight 1, perform weighting processing according to the set third weight 1 and third weight 2 to calculate the first envelope information. Wherein, the third weight one is the weight value of the transitional envelope information, and the third weight two is the weight value of the predicted envelope information.

Specifically, when it is judged in step 406 that the third weight one is greater than the first weight one, it means that the current frame speech and audio signal is closer to the L frame speech and audio signal before switching, and the first envelope information is less affected by the envelope information before switching. Therefore, it is necessary to obtain the first envelope information of the current frame according to the transition envelope information. Wherein, the third weight one is the weight value of the transition envelope information, and the third weight two is the weight value of the predicted envelope information. The weighted value obtained by adding the product of the transition envelope information and the third weight one to the sum of the product of the predicted envelope information and the third weight two is the first envelope information.

Step 409: Decrease the third weight 1 by the unit of the second weight step, increase the third weight 2 by the unit of the second weight step, until the third weight 1 is equal to zero.

Specifically, the purpose of modifying the third weight 1 and the third weight 2 in step 409 is the same as the purpose of modifying the second weight 1 and the second weight 2 in step 405. When the influence of the voice and audio signal of the L frame gradually decreases, in order to make the calculated first envelope information more accurate, the third weight 1 and the third weight 2 are adjusted for applicability. Since the subsequent audio signal is gradually reduced by the influence of the L-frame speech and audio signal before the switch, the value of the third weight 1 gradually decreases, and the value of the third weight 2 gradually increases, thereby also reducing the value of the voice signal before the switch. The influence of the envelope information on the first envelope information. Among them, step 409 can modify the third weight 1 and the third weight 2 through the following method: the new third weight 1 is equal to the old third weight 1 minus the second weight step size, and the new third weight 2 is equal to the old The third weight of 2 plus the second weight step; wherein, the second weight step is a set value.

The sum of the first weight one and the first weight two is one, the sum of the second weight one and the second weight two is one, the sum of the third weight one and the third weight two is one; the initial value of the third weight one is greater than The initial value of the first weight one; the first weight one and the first weight two are fixed constants. Specifically, weight 1 and weight 2 in this embodiment actually represent the percentages when the envelope information before switching and the envelope information of the previous M frames form the first envelope information of the current frame. For the speech and audio signal of the current frame, the closer the speech and audio signal of L frames before switching and the greater the correlation, the higher the percentage of the envelope information before switching is, while the envelope information of M frames before the opposite. the lower the percentage. When the voice and audio signal of the current frame is far from the voice and audio signal of L frames before switching, it means that the voice and audio signal has been transmitted stably in the network, or when the correlation between the voice and audio signal of the current frame and the voice and audio signal of L frames before switching is low, It shows that the characteristics of the speech and audio signal of the current frame have changed. Therefore, the speech and audio signal of the current frame is less affected by the speech and audio signal of L frames before switching, and the percentage of the envelope information before switching is lower.

In addition, the execution order of step 404 and step 405 in this embodiment can be interchanged, that is, the second weight 1 and the second weight 2 can be modified first, and then the transition envelope can be calculated according to the second weight 1 and the second weight 2 information. Similarly, the execution order of step 408 and step 409 in this embodiment can be interchanged, that is, the third weight 1 and the third weight 2 can be modified first, and then the third weight 1 and the third weight 2 can be modified to calculate First envelope information.

Way 2, as shown in Figure 5, another embodiment of obtaining the first envelope information through step 302 may also include:

Step 501, according to the first low frequency band signal of the speech and audio signal of the current frame and the low frequency signal of the speech and audio signal of the previous frame, calculate the correlation coefficient between the first low frequency band signal and the low frequency signal of the speech and audio signal of the previous frame .

Specifically, in order to obtain the first envelope information more accurately, the relationship between the energy of a certain frequency band of the first low frequency band signal of the speech and audio signal of the current frame and the same frequency band of the low frequency signal of the speech and audio signal of the previous frame is solved. In this embodiment, "corr" can be used to represent the correlation coefficient, and the current frame speech and audio signal can be obtained through the energy relationship of the first low frequency band signal of the current frame speech and audio signal and the low frequency band signal of the speech and audio signal of the previous frame in the same frequency band The correlation coefficient corr between the first low-frequency band signal of and the low-frequency band signal of the speech audio signal of the previous frame, the smaller the energy difference is, the larger the corr is, otherwise, the smaller the corr is. For the specific process, please refer to the introduction about the calculation of the correlation of speech and audio signals of the previous N frames in step 401 .

Step 502, judging whether the correlation coefficient is within a given second threshold range.

Specifically, after the value of corr is calculated in step 501, it is determined whether the calculated corr is within a given second threshold range. For example, in this embodiment, the second threshold range may be represented by c1-c2.

Step 503 , if the correlation coefficient is not within the second threshold range, perform weighting processing according to the set first weight 1 and first weight 2 to calculate the first envelope information. Wherein, the first weight one is the weight value of the previous frame envelope information corresponding to the high frequency band signal of the speech audio signal of the previous frame, and the first weight two is the weight value of the predicted envelope information; the first weight one and The first weight two is a fixed constant.

Specifically, when it is obtained in step 502 that corr is less than c1 or greater than c2, it is known that the first envelope information corresponding to the speech and audio signal of the current frame is less affected by the envelope information of the speech and audio signal of the previous frame before switching. Therefore, according to the After setting the first weight 1 and the first weight 2, the first envelope information of the current frame can be calculated. The product of the predicted envelope information and the first weight 2 is added to the sum of the product of the envelope information of the previous frame and the first weight 1, and the obtained weighted sum is the first envelope information of the current frame. In addition, the narrowband speech and audio signals to be transmitted later will restore the first envelope information corresponding to the narrowband speech and audio signals in this manner and weight until the speech and audio signals of different bandwidths are switched again. For example: in this embodiment, the first weight one can be represented by a1, the first weight two can be represented by b1, the envelope information of the previous frame can be represented by pre_fenv, the predicted envelope information can be represented by fenv, the first envelope Information can be represented by cur_fenv. Then step 503 can be expressed by the following formula: cur_fenv=pre_fenv*a1+fenv*b1.

Step 504, if the correlation coefficient is within the second threshold range, determine whether the set second weight one is greater than the first weight one. Wherein, the second weight one is the weight value of the envelope information before switching corresponding to the high frequency band signal of the voice and audio signal of the previous frame,

Specifically, if c1<corr<c2, by comparing the size of the second weight one with the first weight one, it can be known that the first envelope information of the current frame is affected by the envelope information before switching and the envelope information of the previous frame degree of influence.

Step 505. If the second weight one is not greater than the first weight one, calculate the first envelope information according to the set first weight one and first weight two.

Specifically, when it is judged in step 504 that the second weight one is less than the first weight one, it means that the speech and audio signal of the current frame is far from the speech and audio signal of the frame before switching, and the first envelope information is influenced by the envelope information before switching. Therefore, according to the set first weight 1 and first weight 2, the first envelope information of the current frame can be calculated. Then step 505 can be expressed by the following formula: cur_fenv=pre_fenv*a1+fenv*b1.

Step 506, if the second weight one is greater than the first weight one, perform weighting processing according to the second weight one and the set second weight two to calculate the first envelope information. Wherein, the second weight two is the weight value of the predicted envelope information. For example: the second weight one can be represented by a2, and the second weight two can be represented by b2.

Specifically, when it is judged in step 504 that the second weight one is greater than the first weight one, it means that the speech and audio signal of the current frame is closer to the speech and audio signal of the first frequency band of the previous frame before switching, and the first envelope information is affected by the previous frame of switching. The envelope information corresponding to the speech and audio signals before switching has a greater influence. Therefore, according to the set second weight 1 and second weight 2, the first envelope information of the current frame can be calculated. Then the product of the predicted envelope information and the second weight 2 is added to the sum of the products of the envelope information before switching and the second weight 1, and the obtained weighted sum is the first envelope information of the current frame. Wherein, the envelope information before switching can be represented by con_fenv, and step 506 can be expressed by the following formula: cur_fenv=con_fenv*a2+fenv*b2.

Step 507: Decrease the second weight one by the unit of the second weight step, and increase the second weight two by the unit of the second weight step.

Specifically, with the transmission of speech and audio signals, the subsequent speech and audio signals of the current frame are gradually less affected by the speech and audio signals before switching. In order to make the calculated first envelope information more accurate, the second weight One and the second weight two are adjusted for applicability. Since the subsequent audio signal is gradually less affected by the voice and audio signal of the previous frame after switching; and the influence of the previous frame of voice and audio signal close to the current frame of the voice and audio signal becomes greater. Therefore, the value of the second weight 1 gradually decreases, while the value of the second weight 2 gradually increases, thereby weakening the influence of the envelope information before switching on the first envelope information, and enhancing the predicted envelope information on the first envelope information. The impact of envelope information. Among them, step 507 can modify the second weight 1 and the second weight 2 through the following method: the new second weight 1 is equal to the old second weight 1 minus the first weight step size, and the new second weight 2 is equal to the old The second weight of 2 plus the first weight step; wherein, the first weight step is a set value.

Wherein, the sum of the first weight one and the first weight two is one, and the sum of the second weight one and the second weight two is one; the initial value of the second weight one is greater than the initial value of the first weight one.

Step 303: Generate a processed first high frequency band signal according to the first envelope information and the predicted fine structure information.

Specifically, after obtaining the first envelope information of the current frame through step 302, the required processed first high-frequency band signal can be generated according to the first envelope information and the predicted fine structure information, so that the second The high frequency band signal can smoothly transition to the processed first high frequency band signal.

In the speech and audio signal switching method of this embodiment, in the process of switching the speech and audio signal from the broadband speech and audio signal to the narrowband speech and audio signal, the processed information of the current frame is obtained through the predicted fine structure information and the first envelope information. The first high-band signal, so that the second high-band signal of the wide-band speech and audio signal before switching can smoothly transition to the processed first high-band signal corresponding to the narrow-band speech and audio signal, and more It is beneficial to improve the quality of the user listening to the audio signal.

Based on the above technical solution, optionally, as shown in FIG. 6, step 202 in this embodiment includes:

Step 601 : According to the speech and audio signal of the current frame and the speech and audio signal of the frame before switching, determine whether the processed first high frequency band signal needs to be attenuated.

Specifically, since the first high-band signal of the narrowband speech and audio signal is empty, in the process of switching the wideband speech and audio signal to the narrowband speech and audio signal, in order to prevent the recovered narrowband speech and audio signal from processing The first high-frequency band signal has a bad influence. After the number of frames extended from the narrow-band speech and audio signal to the wide-band signal reaches a given number of frames, the energy of the processed first high-band signal is frame by frame. Decay until the decay coefficient reaches the given threshold. Through the voice and audio signal of the current frame and the voice and audio signal of the previous frame, the interval time between the voice and audio signal of the current frame and the voice and audio signal of the previous frame can be known. For example, the narrowband voice and audio signal can be recorded by the counter The number of frames to be transmitted, which can be a predetermined value greater than or equal to 0.

Step 602. If no attenuation is required, synthesize the processed first high-frequency signal and the first low-frequency signal into a wide-band signal.

Specifically, if it is determined in step 601 that the processed first high-frequency band signal does not need to be attenuated, the processed first high-frequency band signal and the first low-frequency band signal are directly synthesized into a wide-band signal.

Step 603. If attenuation is required, determine whether the attenuation factor corresponding to the processed first high frequency band signal is greater than a threshold.

Specifically, the initial value of the attenuation factor is one; the threshold value is less than one and greater than or equal to zero. If it is determined in step 601 that the processed first high frequency band signal needs to be attenuated, then in step 603 it is determined whether the attenuation factor corresponding to the processed first high frequency band signal is greater than a given threshold.

Step 604: If the attenuation factor is not greater than a given threshold, multiply the processed first high-frequency signal by the threshold, and then synthesize a broadband signal with the first low-frequency signal.

Specifically, if it is obtained in step 603 that the value of the attenuation factor is not greater than the given threshold, it means that the energy of the processed first high-frequency band signal has been attenuated to a certain extent, and the processed first high-frequency band signal has no longer It will bring bad effects, and this attenuation ratio can be maintained in the future. Then the processed first high frequency band signal is multiplied by a threshold, and then synthesized with the first low frequency band signal into a broadband signal.

Step 605: If the attenuation factor is greater than a given threshold, multiply the processed first high-frequency signal by the attenuation factor, and then synthesize a broadband signal with the first low-frequency signal.

Specifically, if step 603 shows that the value of the attenuation factor is greater than the given threshold, it means that at this attenuation factor, the processed first high-frequency band signal may cause bad auditory effects, and further attenuation is required. , up to a given threshold. After the processed first high frequency band signal is multiplied by the attenuation factor, the broadband signal is synthesized with the first low frequency band signal.

Step 606. Modify the attenuation factor to reduce the attenuation factor.

Specifically, as the voice and audio signals are transmitted, the subsequent narrowband voice and audio signals are gradually less affected by the voice and audio signals before switching, and correspondingly, the attenuation factor should also be gradually reduced.

Based on the above technical solution, optionally, when the narrowband speech and audio signal is switched to the wideband speech and audio signal, as shown in Figure 7, an implementation of obtaining the processed first high frequency band signal through step 201 in this embodiment Examples include:

Step 701. Perform weighting processing according to the set fourth weight 1 and fourth weight 2 to calculate the processed first high frequency band signal. Wherein, the fourth weight 1 is the weight value of the second high frequency band signal, and the fourth weight 2 is the weight value of the first high frequency band signal of the current frame audio signal.

Specifically, in the process of switching from the narrowband speech and audio signal to the wideband speech and audio signal, since the high frequency signal in the wideband speech and audio signal is not empty, the high frequency signal corresponding to the narrowband speech and audio signal is Empty or processed high-frequency signal, in order to make the narrow-band speech and audio signal can be smoothly switched to the wide-band speech and audio signal, it is necessary to attenuate the energy of the high-band signal in the wide-band speech and audio signal to realize speech and audio signal to switch smoothly. Through the product of the second high frequency band signal and the fourth weight one, plus the sum of the product of the first high frequency band signal and the fourth weight two, the obtained weighted value is the processed first high frequency band signal .

Step 702: Decrease the fourth weight 1 by the unit of the third weight step, increase the fourth weight 2 by the unit of the third weight step, until the fourth weight 1 is equal to zero. Wherein, the sum of the fourth weight one and the fourth weight two is one.

Specifically, with the transmission of the voice and audio signals, the subsequent wideband voice and audio signals are gradually less affected by the narrowband voice and audio signals before switching. Therefore, the fourth weight one gradually becomes smaller, and the fourth weight two gradually increases until the fourth weight one becomes zero, and the fourth weight two becomes one, that is, the transmitted voice and audio signal is always a broadband voice and audio signal .

Similarly, as shown in FIG. 8, another embodiment of obtaining the processed first high-frequency band signal through step 201 in this embodiment may also include:

Step 801: Perform weighting processing according to the set fifth weight 1 and fifth weight 2 to calculate the processed first high frequency band signal. Wherein, the fifth weight 1 is the weight value of the fixed parameter that has been set, and the fifth weight 2 is the weight value of the first high frequency band signal of the speech and audio signal of the current frame.

Specifically, since the first high-band signal of the narrow-band speech and audio signal is empty, a fixed parameter can be set to replace the high-band signal of the narrow-band speech and audio signal, wherein the fixed parameter is one greater than or equal to zero and less than the first A constant for high-band signal energy. The weighted value obtained by adding the product of the fixed parameter and the fifth weight one to the sum of the product of the first high frequency band signal and the fifth weight two is the processed first high frequency band signal.

Step 802: Decrease the fifth weight 1 by the unit of the fourth weight step, increase the fifth weight 2 by the unit of the fourth weight step, until the fifth weight 1 is equal to zero; wherein, the fifth weight 1 and the fifth weight The sum of two is one.

Specifically, with the transmission of the voice and audio signals, the subsequent wideband voice and audio signals are gradually less affected by the narrowband voice and audio signals before switching. Therefore, the fifth weight 1 gradually decreases, and the fifth weight 2 gradually increases until the fifth weight 1 becomes zero, and the fifth weight 2 becomes 1, that is, the transmitted voice and audio signal is always a real broadband voice signal. audio signal.

In the speech and audio signal switching method of this embodiment, in the process of switching the speech and audio signal from the narrowband speech and audio signal to the wideband speech and audio signal, the processed signal is obtained by attenuating the high frequency band signal of the wideband speech and audio signal. High-band signal, so that the high-band signal corresponding to the narrow-band voice and audio signal before switching can smoothly transition to the processed high-band signal corresponding to the wide-band voice and audio signal, which is more conducive to improving the user's ability to listen to audio The quality of the signal.

Wherein, the envelope information in this embodiment may also be replaced by other parameters that can represent high-frequency band signals, such as: Linear Predictive Coding (Linear Predictive Coding, hereinafter referred to as: LPC) parameters, or amplitude parameters, etc.

Those of ordinary skill in the art can understand that all or part of the steps for realizing the above-mentioned method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the It includes the steps of the above method embodiments; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other various media that can store program codes.

FIG. 9 is a schematic structural diagram of Embodiment 1 of the speech and audio signal switching device of the present invention. As shown in FIG. 9 , the audio signal switching device of this embodiment includes: a processing module 91 and a first synthesis module 92 .

The processing module 91 is used to perform weighting processing on the first high-band signal of the current frame of speech and audio signal and the second high-frequency signal of the previous M frames of speech and audio signal when the speech and audio signal is switched, so as to obtain the processed first high frequency band signal. A high frequency band signal. Wherein, M is greater than or equal to 1.

The first synthesizing module 92 is used for synthesizing the processed first high frequency band signal and the first low frequency band signal of the speech and audio signal of the current frame into a broadband signal.

In the speech and audio signal switching device of the embodiment of the present invention, the first high frequency band signal in the speech and audio signal of the current frame is processed by the processing module according to the second high frequency band signal in the speech and audio signal of the previous M frames, so that the first The second high frequency band signal can smoothly transition to the processed first high frequency band signal, so that in the process of switching voice and audio signals of different bandwidths, the high frequency band signals of voice and audio signals of different bandwidths can be switched smoothly; Finally, the processed first high-frequency signal and the first low-frequency signal are synthesized into a wide-band signal through the first synthesis module, and the wide-band signal is transmitted to the user terminal, so that the user can enjoy high-quality voice and audio signals. The speech and audio signal switching method of this embodiment can smoothly switch speech and audio signals of different bandwidths, reduces the impact of poor subjective auditory quality of speech and audio signals caused by energy shocks, and improves the quality of speech and audio signals received by users.

FIG. 10 is a schematic structural diagram of Embodiment 2 of the speech and audio signal switching device of the present invention. As shown in FIG. 10 , the audio signal switching device of this embodiment is based on the first embodiment of the audio signal switching device above, and the difference is that the audio signal switching device of this embodiment further includes: a second synthesis module 103 .

The second synthesizing module 103 is used for synthesizing the first high-frequency signal and the first low-frequency signal into a wide-band signal when the voice and audio signals are not switched.

In the audio signal switching device of this embodiment, by setting the second synthesis module, the second synthesis module can combine the first low-frequency audio signal in the first frequency band audio and video signal of the current frame without switching audio and audio signals of different bandwidths. The broadband signal is synthesized from the band signal and the first high frequency band signal, thereby improving the quality of the user listening to the voice and audio signals.

Based on the above technical solution, optionally, when the wideband voice and audio signal is switched to the narrowband voice and audio signal, as shown in Figure 10 and Figure 11, the processing module 101 in this embodiment includes:

The prediction module 1011 is configured to predict the predicted fine structure information and the predicted envelope information corresponding to the first high frequency band signal.

The first generating module 1012 is used to perform weighting processing according to the predicted envelope information and the envelope information of the previous M frames corresponding to the second high-band signal of the previous M frames of speech and audio signals, so as to obtain the first high-band signal corresponding to the first An envelope of information.

The second generating module 1013 is configured to generate the processed first high frequency band signal according to the first envelope information and the predicted fine structure information.

Furthermore, the audio signal switching device in this embodiment may further include: a classification module 1010, configured to perform signal classification on the first low-frequency band signal of the current frame of speech and audio signals; and the prediction module 1011 is also configured to The corresponding signal type predicts the predicted fine structure information and the predicted envelope information corresponding to the first low frequency band signal of the speech and audio signal of the current frame.

The audio signal switching device in this embodiment uses the prediction module to predict the predicted fine structure information and predicted envelope information corresponding to the first high-frequency band signal, so that the first generation module and the second generation module can accurately generate the processed The first high-frequency band signal, so that the first high-frequency band signal can transition to the processed first high-frequency band signal more smoothly, which is more conducive to improving the quality of the user's listening to the voice and audio signals. In addition, signal classification is performed on the first low-frequency band signal of the current frame speech and audio signal through the classification module, and then the prediction module obtains the predicted fine structure information and the predicted envelope information according to the signal type, so that the predicted fine structure information and the predicted The envelope information is more accurate, so that the quality of the voice and audio signals received by the user is higher.

Based on the above technical solution, optionally, as shown in Figure 10 and Figure 12, the first synthesis module 102 in this embodiment includes:

The first judging module 1021 is configured to judge whether the processed first high frequency band signal needs to be attenuated according to the speech and audio signal of the current frame and the speech and audio signal of the previous frame before switching.

The third synthesis module 1022 is used for synthesizing the processed first high frequency signal and the first low frequency signal into a broadband signal if the first judging module 1021 finds that the processed first high frequency signal does not need to be attenuated.

The second judging module 1023 is configured to judge whether the attenuation factor corresponding to the processed first high-frequency signal is greater than a given threshold if the first judging module 1021 finds that the processed first high-frequency signal needs to be attenuated.

The fourth synthesizing module 1024 is used for if the second judging module 1023 obtains that the attenuation factor is not greater than a given threshold value, then multiply the processed first high frequency band signal by the threshold value, and then synthesize the broadband signal with the first low frequency band signal .

The fifth synthesizing module 1025 is used for if the second judging module 1023 obtains that the attenuation factor is greater than a given threshold, then multiply the processed first high-frequency band signal by the attenuation factor, and then synthesize the broadband with the first low-frequency band signal Signal.

The first modification module 1026 is used for modifying the attenuation factor, so as to reduce the attenuation factor.

Wherein, the initial value of the attenuation factor is one; the threshold value is less than one and greater than or equal to zero.

The audio signal switching device of this embodiment, by attenuating the processed first high-frequency band signal, can make the broadband signal obtained after the current frame speech and audio signal is processed more accurate, which is more conducive to improving the user's ability to listen to the audio signal. quality.

Based on the above technical solution, optionally, when the narrowband speech and audio signal is switched to the wideband speech and audio signal, as shown in Figure 10 and Figure 13a, the processing module 101 in this embodiment includes:

The first calculation module 1011a is used to perform weighting processing according to the set fourth weight one and fourth weight two to calculate the processed first high frequency band signal; wherein, the fourth weight one is the second high frequency The weight value of the band signal, the fourth weight two is the weight value of the first high frequency band signal;

The second modifying module 1012a is used to reduce the fourth weight 1 with the third weight step as a unit, and increase the fourth weight 2 with the third weight step as a unit, until the fourth weight 1 is equal to zero; wherein, the fourth weight 1 and The sum of the fourth weight two is one.

Similarly, when the narrowband speech and audio signal is switched to the wideband speech and audio signal, as shown in Figure 10 and Figure 13b, the processing module 101 in this embodiment may further include:

The second calculation module 1011b is used to perform weighting processing according to the set fifth weight 1 and the fifth weight 2 to calculate the processed first high-frequency band signal; wherein, the fifth weight 1 is set The weight value of the fixed parameter, the fifth weight two is the weight value of the first high frequency band signal;

The third modification module 1012b is used to reduce the fifth weight 1 in units of the fourth weight step size, increase the fifth weight 2 in units of the fourth weight step size, until the fifth weight 1 is equal to zero; wherein, the fifth weight 1 The sum of the fifth weight 2 and the fifth weight 2 is one; wherein, the fixed parameter is a constant greater than or equal to zero and less than the energy value of the first high frequency band signal.

The speech and audio signal switching device of this embodiment, in the process of switching the speech and audio signal from the narrowband speech and audio signal to the wideband speech and audio signal, obtains the processed signal by attenuating the high frequency band signal of the wideband speech and audio signal High-band signal, so that the high-band signal corresponding to the narrow-band voice and audio signal before switching can smoothly transition to the processed high-band signal corresponding to the wide-band voice and audio signal, which is more conducive to improving the user's ability to listen to audio The quality of the signal.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still be Modifications are made to the technical solutions described in the foregoing embodiments, or equivalent replacements are made to some of the technical features; 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 various embodiments of the present invention.

Claims (16)

1. A voice and audio signal switching method, characterized in that, comprising: When the voice and audio signal is switched, the first high frequency band signal of the current frame voice and audio signal and the second high frequency band signal of the previous M frames of voice and audio signal are weighted to obtain the processed first high frequency band signal ; Among them, M is greater than or equal to 1; Combining the processed first high frequency band signal and the first low frequency band signal of the speech and audio signal of the current frame into a broadband signal. 2. The speech and audio signal switching method according to claim 1, is characterized in that, also comprises: When the speech and audio signals do not switch, the first high frequency band signal and the first low frequency band signal are synthesized into the broadband signal. 3. The voice and audio signal switching method according to claim 1 or 2, wherein, when the broadband voice and audio signal is switched to the narrowband voice and audio signal; The first high-band signal of the current frame speech and audio signal and the second high-band signal of the previous M frames of speech and audio signal are weighted to obtain the processed first high-band signal as follows: Predicting the predicted fine structure information and predicted envelope information corresponding to the first high-frequency band signal of the speech and audio signal of the current frame; performing weighting processing on the predicted envelope information and the envelope information of the previous M frames corresponding to the second high-band signal of the speech and audio signal of the previous M frames, so as to obtain the first envelope information; The processed first high frequency band signal is generated according to the first envelope information and the predicted fine structure information. 4. The voice-audio signal switching method according to claim 3, wherein the predicted fine structure information and the predicted envelope information corresponding to the first high frequency band signal are specifically: performing signal classification on the first low frequency band signal of the speech and audio signal of the current frame; Predicting the predicted fine structure information and the predicted envelope information according to the signal type corresponding to the first low-frequency signal. 5. voice signal switching method according to claim 3, is characterized in that, the envelope signal of described prediction and the first M frame corresponding to the second high frequency band signal of described preceding M frame voice signal The envelope information is weighted to obtain the first envelope information corresponding to the first high-frequency band signal, specifically: According to the first low frequency band signal and the low frequency band signal of the speech audio signal of the previous N frames, calculate the correlation coefficient between the first low frequency band signal and the low frequency band signal of the speech audio signal of the previous N frames; Among them, N is greater than or equal to 1; judging whether the correlation coefficient is within a given first threshold range; If the correlation coefficient is not within the first threshold range, perform weighting processing according to the set first weight one and first weight two to calculate the first envelope information; wherein, the The first weight one is the weight value of the envelope information of the previous frame corresponding to the high frequency band signal of the voice and audio signal of the previous frame, and the first weight two is the weight value of the envelope information; If the correlation coefficient is within the first threshold range, perform weighting processing according to the set second weight 1 and second weight 2 to calculate transition envelope information; wherein, the second weight 1 It is the weight value of the envelope information before the switch corresponding to the high frequency band signal of the L frame voice and audio signal before the switch, and the second weight two is the weight value of the envelope information of the M frame before; wherein, L is greater than equal to 1; Decreasing the second weight one by the unit of the first weight step, and increasing the second weight two by the unit of the first weight step; judging whether the set third weight one is greater than the first weight one; If the third weight 1 is not greater than the first weight 1, perform weighting processing according to the set first weight 1 and the first weight 2 to calculate the first envelope information ; If the third weight 1 is greater than the first weight 1, perform weighting processing according to the set third weight 1 and third weight 2 to calculate the first envelope information; wherein, the first 3. Weight 1 is the weight value of the transition envelope information, and the third weight 2 is the weight value of the predicted envelope information; Decreasing the third weight 1 in units of the second weight step, increasing the third weight 2 in units of the second weight step, until the third weight 1 is equal to zero; Wherein, the sum of the first weight one and the first weight two is one, the sum of the second weight one and the second weight two is one, the third weight one and the third weight The sum of two is one; the initial value of the third weight one is greater than the initial value of the first weight one; the first weight one and the first weight two are fixed constants. 6. The speech and audio signal switching method according to claim 3, characterized in that, the envelope signal of said prediction and the corresponding front M frame envelope of the second high frequency band signal of the previous M frame speech and audio signal The information is weighted to obtain the first envelope information corresponding to the first high-frequency band signal, which is specifically: Calculate the correlation coefficient between the first low frequency band signal and the low frequency signal of the speech and audio signal of the previous frame according to the first low frequency band signal of the current frame and the low frequency signal of the speech and audio signal of the previous frame ; judging whether the correlation coefficient is within a given second threshold range; If the correlation coefficient is not within the second threshold range, perform weighting processing according to the set first weight 1 and first weight 2 to calculate the first envelope information; wherein, the The first weight one is the weight value of the envelope information of the previous frame corresponding to the high frequency band signal of the speech and audio signal of the previous frame, and the first weight two is the weight value of the predicted envelope information; The first weight one and the first weight two are fixed constants; If the correlation coefficient is within the second threshold range, it is judged whether the set second weight one is greater than the first weight one; wherein, the second weight one is the speech and audio signal of the previous frame before switching The weight value of the envelope information before switching corresponding to the high frequency band signal, If the second weight one is not greater than the first weight one, performing weighting processing according to the set first weight one and first weight two to calculate the first envelope information; If the second weight 1 is greater than the first weight 1, perform weighting processing according to the second weight 1 and the set second weight 2 to calculate the first envelope information; wherein, The second weight two is the weight value of the predicted envelope information; Decreasing the second weight one by the unit of the second weight step, increasing the second weight two by the unit of the second weight step; Wherein, the sum of the first weight one and the first weight two is one, and the sum of the second weight one and the second weight two is one; the initial value of the second weight one is greater than the The initial value of the first weight one. 7. The voice-audio signal switching method according to claim 3, wherein the first low-band signal of the processed first high-band signal and the current frame voice-audio signal is synthesized into a broadband The signals are specifically: According to the voice and audio signal of the current frame and the voice and audio signal of the previous frame before switching, it is judged whether the processed first high frequency band signal needs to be attenuated; If attenuation is not required, combining the processed first high-frequency signal with the first low-frequency signal to synthesize the broadband signal; If attenuation is required, judging whether the attenuation factor corresponding to the processed first high-frequency band signal is greater than a given threshold; If the attenuation factor is not greater than the given threshold, multiplying the processed first high-frequency signal by a threshold, and then synthesizing the broadband signal with the first low-frequency signal; If the attenuation factor is greater than the given threshold, multiplying the processed first high-frequency signal by the attenuation factor, and then synthesizing the broadband signal with the first low-frequency signal; modifying the attenuation factor such that the attenuation factor is reduced; Wherein, the initial value of the attenuation factor is one; the threshold is less than one and greater than or equal to zero. 8. The voice and audio signal switching method according to claim 1 or 2, characterized in that, When the narrowband speech and audio signal is switched to the wideband speech and audio signal; The first high frequency band signal of the current frame speech and audio signal and the second high frequency band signal of the previous M frame speech and audio signals are weighted to obtain the processed first high frequency band signal as follows: Perform weighting processing according to the set fourth weight 1 and fourth weight 2 to calculate the processed first high-frequency band signal; wherein, the fourth weight 1 is the second high-frequency band A weight value of the signal, the fourth weight two being the weight value of the first high frequency band signal; Decreasing the fourth weight 1 with the third weight step as a unit, increasing the fourth weight 2 with the third weight step as a unit, until the fourth weight 1 is equal to zero; wherein, the fourth weight 1 and the sum of the fourth weight two is one. 9. The voice and audio signal switching method according to claim 1 or 2, wherein, when the narrowband voice and audio signal is switched to the wideband voice and audio signal; The first high frequency band signal of the current frame speech and audio signal and the second high frequency band signal of the previous M frame speech and audio signals are weighted to obtain the processed first high frequency band signal as follows: Perform weighting processing according to the set fifth weight 1 and fifth weight 2 to calculate the processed first high-frequency band signal; wherein, the fifth weight 1 is a set fixed parameter The weight value of the fifth weight two is the weight value of the first high frequency band signal; Decreasing the fifth weight 1 with the fourth weight step as a unit, increasing the fifth weight 2 with the fourth weight step as a unit, until the fifth weight 1 is equal to zero; wherein, the fifth weight 1 and the sum of the fifth weight two is one; Wherein, the fixed parameter is a constant greater than or equal to zero and less than the energy value of the first high frequency band signal. 10. A voice and audio signal switching device, characterized in that, comprising: The processing module is used for weighting the first high-band signal of the current frame of speech and audio signals and the second high-band signal of the previous M frames of speech and audio signals when the speech and audio signals are switched, so as to obtain the processed first A high frequency band signal; wherein, M is greater than or equal to 1; The first synthesizing module is configured to synthesize the processed first high frequency band signal and the first low frequency band signal of the speech and audio signal of the current frame into a broadband signal. 11. The voice and audio signal switching device according to claim 10, further comprising: The second synthesizing module is used for synthesizing the wideband signal with the first high frequency band signal and the first low frequency band signal when the voice and audio signal is not switched. 12. The speech and audio signal switching device according to claim 10 or 11, wherein, when the broadband speech and audio signal is switched to the narrowband speech and audio signal; The processing modules include: A prediction module, configured to predict the predicted fine structure information and predicted envelope information corresponding to the first high-frequency band signal of the current frame speech and audio signal; The first generation module is configured to perform weight processing on the predicted envelope information and the envelope information of the previous M frames corresponding to the second high frequency band signal of the speech and audio signal of the previous M frames, so as to obtain the first high First envelope information corresponding to the frequency band signal; A second generating module, configured to generate the processed first high frequency band signal according to the first envelope information and the predicted fine structure information. 13. The speech and audio signal switching device according to claim 12, further comprising: a classification module for performing signal classification on the first low frequency band signal of the current frame speech and audio signal; The prediction module is further configured to predict the predicted fine structure information and the predicted envelope information according to the signal type corresponding to the first low frequency band signal. 14. The speech and audio signal switching device according to claim 12, wherein the first synthesis module comprises: The first judging module is used to judge whether the processed first high frequency band signal needs to be attenuated according to the speech and audio signal of the current frame and the speech and audio signal of the previous frame before switching; A third synthesis module, configured to combine the processed first high frequency signal with the first low frequency signal if the first judging module obtains that the processed first high frequency signal does not need to be attenuated signal synthesizing the broadband signal; The second judging module is used for judging whether the attenuation factor corresponding to the processed first high-frequency band signal is greater than a given value if the first judging module obtains that the processed first high-frequency band signal needs to be attenuated the threshold; A fourth synthesis module, configured to multiply the processed first high frequency band signal by a threshold if the attenuation factor obtained by the second judging module is not greater than the given threshold, and then combine it with the combining the first low-band signal with the wide-band signal; The fifth synthesis module is configured to multiply the processed first high-frequency band signal by the attenuation factor if the attenuation factor is greater than the given threshold value by the second judging module, and then combine with the attenuation factor Synthesizing the broadband signal with the first low-band signal; A first modification module, configured to modify the attenuation factor, so that the attenuation factor is reduced; Wherein, the initial value of the attenuation factor is one; the threshold is less than one and greater than or equal to zero. 15. The speech and audio signal switching device according to claim 10 or 11, wherein when the narrowband speech and audio signal is switched to the wideband speech and audio signal; The processing modules include: The first calculation module is configured to perform weighting processing according to the set fourth weight one and fourth weight two, so as to calculate the processed first high-frequency band signal; wherein, the fourth weight one is The weight value of the second high frequency band signal, the fourth weight two is the weight value of the first high frequency band signal; The second modifying module is configured to reduce the fourth weight one by the unit of the third weight step, and increase the fourth weight two by the unit of the third weight step until the fourth weight one is equal to zero; wherein , the sum of the fourth weight one and the fourth weight two is one. 16. The speech and audio signal switching device according to claim 13, wherein, when the narrowband speech and audio signal is switched to the wideband speech and audio signal; the processing module includes: The second calculation module is configured to perform weighting processing according to the set fifth weight 1 and fifth weight 2 to calculate the processed first high-frequency band signal; wherein, the fifth weight 1 is The weight value of the fixed parameter that has been set, the fifth weight 2 is the weight value of the first high frequency band signal; A third modifying module, configured to decrease the fifth weight 1 by the unit of the fourth weight step, increase the fifth weight 2 by the unit of the fourth weight step, until the fifth weight 1 is equal to zero; Wherein, the sum of the fifth weight 1 and the fifth weight 2 is one; wherein, the fixed parameter is a constant greater than or equal to zero and less than the energy value of the first high frequency band signal.

Images