Voice signal processing method, device, equipment, medium and computer program product

Abstract

The embodiment of the invention discloses a voice signal processing method and device, equipment, a medium and a computer program product. The method comprises the following steps: acquiring a first voice signal acquired by a first terminal, a second voice signal transmitted to a second terminal by the first terminal, and a third voice signal transmitted to the first terminal by the second terminal; performing adaptive filtering processing on the second voice signal and the third voice signal based on a plurality of initial adaptive filters to obtain an interference signal in the simulated first voice signal; and performing suppression processing on an interference signal in the first voice signal to obtain a target voice signal after suppression processing. According to the method and the device, the howling suppression effect in the call process can be realized, the short-distance call effect is effectively optimized, obviously distorted reverberation trailing caused by howling is avoided, and thus the call voice quality is effectively improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a speech signal processing method, apparatus, device, medium and computer program product. Background Art

[0002] With the continuous development of voice signal processing technology, users have higher and higher requirements for voice quality. However, echo or howling is easy to occur in daily calls, which seriously affects the voice quality, especially when making short-distance calls. The signal of the near-end microphone is received by the near-end speaker through the acoustic path, and then returned from the network side tone or the acoustic path of the other end, thus forming feedback. If the feedback becomes positive feedback, howling will be formed. Howling has a greater impact on the microphone. For example, the delayed feedback of the speaker sound field will cause the entire system to form a series of delayed echoes, and this echo will aggravate the comb filter effect, produce a significantly distorted reverberation tail, and seriously affect the call voice quality. Summary of the invention

[0004] In view of this, the present application provides a method, device, medium and computer program product for processing a speech signal, which are used to solve the problem of deterioration of speech quality caused by howling in existing communication technology. In order to achieve one or part or all of the above purposes or other purposes, the present application proposes a method, device, medium and computer program product for processing a speech signal.

[0005] A first aspect: a speech signal processing method, comprising:

[0006] Acquire a first voice signal collected by a first terminal, a second voice signal transmitted by the first terminal to the second terminal, and a third voice signal transmitted by the second terminal to the first terminal;

[0007] Based on the initial multiple adaptive filters, adaptive filtering is performed on the second voice signal and the third voice signal respectively to obtain a simulated interference signal in the first voice signal; the interference signal is used to update the adaptive filter, and the updated adaptive filter is used to continue filtering the second voice signal and the third voice signal of the subsequent frame;

[0008] Suppress the interference signal in the first speech signal to obtain a target speech signal after suppression processing

[0009] A second aspect: A speech signal processing device, comprising:

[0010] an acquisition module, configured to acquire a first voice signal collected by a first terminal, a second voice signal transmitted by the first terminal to the second terminal, and a third voice signal transmitted by the second terminal to the first terminal;

[0011] A filtering module, configured to perform adaptive filtering processing on the second voice signal and the third voice signal respectively based on the initial multiple adaptive filters to obtain a simulated interference signal in the first voice signal; the interference signal is used to update the adaptive filter, and the updated adaptive filter is used to continue filtering the second voice signal and the third voice signal of subsequent frames;

[0012] The suppression module is used to suppress the interference signal in the first speech signal to obtain a target speech signal after suppression.

[0013] In one embodiment, the first voice signal includes a voice signal sent by the first user, a voice signal corresponding to a first interference path between the first terminal and the second terminal, and a voice signal corresponding to a second interference path of the first terminal itself.

[0014] In one embodiment, the first interference path is a path for transmitting a voice signal through space between the first terminal and the second terminal; and the second interference path is a path for transmitting a voice signal through space between the first terminal and the second terminal.

[0015] In one embodiment, the second voice signal includes a voice signal transmitted from the first terminal to the second terminal through the target path; and the third voice signal includes a voice signal transmitted from the second terminal to the first terminal through the target path.

[0016] In one embodiment, the adaptive filter includes a first adaptive filter and a second adaptive filter; the interference signal includes a first interference signal and a second interference signal;

[0017] The filtering module comprises:

[0018] a first filtering unit, configured to filter the second voice signal through the first adaptive filter to obtain a first interference signal corresponding to a first interference path in the simulated first voice signal;

[0019] The second filtering unit is used to filter the third voice signal through the second adaptive filter to obtain a second interference signal corresponding to the second interference path in the simulated first voice signal.

[0020] In one embodiment, the first filtering unit is specifically used to filter the second voice signal through the first adaptive filter according to the current filter parameters to obtain a first interference signal corresponding to the first interference path in the simulated first voice signal; the first interference signal is a simulated signal having similar characteristics to the second voice signal.

[0021] In one embodiment, the second filtering unit is specifically used to filter the third voice signal through the second adaptive filter according to the current filter parameters to obtain a second interference signal corresponding to the second interference path in the simulated first voice signal; the second interference signal is a simulated signal having similar characteristics to the third voice signal.

[0022] In one embodiment, the device further includes an updating module, wherein the updating module includes:

[0023] A first updating unit, configured to update a first adaptive filter corresponding to the first interference path based on the first interference signal and a speech signal corresponding to a first interference path between the first terminal and the second terminal;

[0024] The second updating unit is used to update the second adaptive filter corresponding to the second interference path based on the second interference signal and the speech signal corresponding to the second interference path of the first terminal itself.

[0025] In one embodiment, the interference signal includes a first interference signal and a second interference signal; the suppression module is specifically used to: suppress the first interference signal and the second interference signal in the first speech signal to remove the first interference signal and the second interference signal from the first speech signal to obtain a target speech signal after suppression processing.

[0026] A third aspect: An electronic device comprises at least one processor and at least one memory, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the speech signal processing method of each embodiment of the present application are implemented.

[0027] A fourth aspect: A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the speech signal processing method of each embodiment of the present application.

[0028] A fifth aspect: A computer program product, comprising a computer program, which, when executed by a processor, implements the steps of the speech signal processing method of each embodiment of the present application.

[0029] Implementing the embodiments of the present application will have the following beneficial effects:

[0030] First, obtain a first voice signal collected by the first terminal, a second voice signal transmitted from the first terminal to the second terminal, and a third voice signal transmitted from the second terminal to the first terminal; then adaptively filter the second voice signal and the third voice signal based on the initial multiple adaptive filters to obtain the interference signal in the simulated first voice signal, thereby updating the adaptive filter through the interference signal, so as to continue to filter the second voice signal and the third voice signal of the subsequent frame through the updated adaptive filter, and finally suppress the interference signal in the first voice signal to obtain the target voice signal after suppression. The present application adds multiple adaptive filters to the duplex call system, adopts the idea of ​​feedback cancellation, calls the adaptive filters at both ends of the duplex call at the same time, estimates the actual feedback path and matches it with the adaptive filter, thereby cutting off the feedback path, realizing the effect of suppressing howling during the call, effectively optimizing the short-distance call effect, and avoiding the reverberation tail with obvious distortion caused by howling, thereby effectively improving the call voice quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0032] in:

[0033] Figure 1 An application environment diagram of the speech signal processing method provided in an embodiment of the present application;

[0034] Figure 2 A flowchart of a speech signal processing method provided in an embodiment of the present application;

[0035] Figure 3 A schematic diagram of a close-range duplex call scenario provided in an embodiment of the present application;

[0036] Figure 4 A schematic diagram of an implementation of a speech signal processing method provided in an embodiment of the present application;

[0037] Figure 5 A schematic diagram of another implementation of the speech signal processing method provided in an embodiment of the present application;

[0038] Figure 6 A schematic diagram of a communication scenario between a door screen inside and a door peephole outside provided in an embodiment of the present application;

[0039] Figure 7 A schematic diagram of the structure of a speech signal processing device provided in an embodiment of the present application;

[0040] Figure 8 This is a diagram of the internal structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0042] In the following description, reference is made to “some embodiments”, which describe a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0043] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0044] The speech signal processing method provided in this application can be applied to Figure 1 In the application environment shown. Figure 1 A speech signal processing system is provided, the target detection system includes a smart device 100, a network device 200 connected to the smart device 100 for communication, a server 300, and a user terminal 400. The smart device 100 may include multiple devices, such as a smart device 1002 and a smart device 1004.

[0045] The smart device 100 accesses the network device 200 in the target detection system and communicates with the network device 200 through its own configured communication module. In one embodiment, the smart device 100 accesses the network device 200 through a local area network path or a wide area network path, and is deployed in the network device 200. The local area network may include ZIGBEE or Bluetooth, and the wide area network may include 2G / 3G / 4G / 5G / WIFI, etc.

[0046] The network device 200 can establish a network connection with the server 300 and the user terminal 400. In one embodiment, the network device 200 and the user terminal 400 can establish a network connection through a local area network or a wide area network path. Through the network connection, the user interacts with the user terminal 400, so that the user can control the smart device 100 connected to the network device 200 to perform corresponding actions with the help of the user terminal 400 or the smart device 100.

[0047] In the specific implementation process, the first terminal is the smart device 1002, and the second terminal is the smart device 1004. For example, a first voice signal collected by the smart device 1002, a second voice signal transmitted from the smart device 1002 to the smart device 1004, and a third voice signal transmitted from the smart device 1004 to the smart device 1002 are obtained, and the second voice signal and the third voice signal are adaptively filtered based on the initial multiple adaptive filters to obtain an interference signal in the simulated first voice signal; the interference signal is used to update the adaptive filter, and the updated adaptive filter is used to continue filtering the second voice signal and the third voice signal of the subsequent frames; the interference signal in the first voice signal is suppressed, so that the smart device 1002 can obtain the target voice signal after suppression.

[0048] In addition, the present application can also be implemented through hardware circuits or hardware circuits combined with software instructions. Therefore, the implementation of the present application is not limited to any specific hardware circuits, software, or a combination of the two.

[0049] It should be noted that the order of description of the following embodiments is not intended to limit the priority order of the embodiments.

[0050] See also Figure 2 A speech signal processing method is provided, and the method is applied to an electronic device as an example for explanation. The electronic device may be Figure 1 The smart device, terminal, gateway, server, terminal, etc. in the present application are not limited to this. The speech signal processing method specifically includes the following steps:

[0051] S1. Acquire a first voice signal collected by the first terminal, a second voice signal transmitted from the first terminal to the second terminal, and a third voice signal transmitted from the second terminal to the first terminal;

[0052] The first terminal is a device with a communication function and can directly interact with a user, and can specifically be a communication device including a sound collection module and a sound playback module. The second terminal is also a device with a communication function and can directly interact with a user, and can specifically be a communication device including a sound collection module and a sound playback module.

[0053] It is understandable that the first terminal and the second terminal may be two independent devices, specifically the same device or different devices. The first terminal and the second terminal may communicate with each other via a wireless connection or a wired connection.

[0054] It can be understood that the voice signal may include sound waves generated by human voices transmitted through a medium such as air. Among them, the first voice signal may refer to the voice signal collected by the first terminal. When a first user makes a close-range call with a second user through a second terminal through a first terminal, the voice signal collected by the first terminal may include the voice signal sent by the first user, and may also include the noise signal collected by the first terminal through the air after the first terminal transmits the voice signal sent by the first user to the second terminal and plays it, and may also include the noise signal collected by the first terminal through the air after the voice signal sent by the second user is transmitted to the first terminal and played.

[0055] The second voice signal may refer to a voice signal transmitted from the first terminal to the second terminal via a wireless network and / or wired transmission, and specifically may be a target voice signal after the first voice signal is suppressed. It is understandable that during the communication between the first terminal and the second terminal, the voice signals collected by each terminal will continue to be suppressed. Among them, the first terminal may not suppress the first voice signal of the first frame collected, but directly transmit it to the second terminal. However, the first voice signal of each frame collected subsequently will be suppressed before being transmitted to the second terminal. Therefore, the second voice signal of the second frame and thereafter transmitted from the first terminal to the second terminal are all target voice signals obtained by suppressing the first voice signal of the previous frame.

[0056] Specifically, in step S1, a first voice signal is collected by a first terminal. Figure 3 As shown, A and B are two users who are talking on the phone, Mic1 and Speaker1 are the microphone and player of the first terminal corresponding to user A, and Mic2 and Speaker2 are the microphone and player of the second terminal corresponding to user B. The first voice signal collected by the first terminal includes the voice signal (A0) of user A speaking directly to the microphone of the first terminal, and may include the interference voice signal A2 collected by the first terminal through the air after the first terminal transmits the voice signal A0 sent by user A to the second terminal and plays it, and may also include the interference voice signal B3 collected by the first terminal through the air after the voice signal B0 sent by user B is transmitted to the first terminal and played.

[0057] Among them, the first terminal is the communication device of user A, which is responsible for collecting the voice of user A and transmitting it to the second terminal corresponding to user B. At the same time, it also receives the voice of user B and eliminates the howling caused by acoustic feedback. The second terminal is the communication device of user B, which is responsible for collecting the voice of user B and transmitting it to user A. At the same time, it also receives the voice of user A and eliminates the howling caused by acoustic feedback. The first voice signal refers to the original voice signal collected by the first terminal in the duplex call system, including the sound of the user speaking directly to the microphone of the first terminal, and the sound that may be transmitted from the speaker to the microphone through space propagation.

[0058] The first terminal transmits the collected voice signal (including the original voice and possible interference) to the second terminal to form a second voice signal. The second voice signal is a voice signal collected by the first terminal and transmitted to the second terminal in a duplex call system. It represents the voice information of the first terminal user (user A). After certain processing and encoding, it is sent to the second terminal through the communication network. At the same time, the voice signal collected by the second terminal is also transmitted to the first terminal to form a third voice signal. The third voice signal is a voice signal collected by the second terminal and transmitted to the first terminal in a duplex call system. This signal represents the voice information of the second terminal user (user B). After processing and encoding, it is sent to the first terminal through the communication network.

[0059] S2. Based on the initial multiple adaptive filters, the second voice signal and the third voice signal are adaptively filtered respectively to obtain the interference signal in the simulated first voice signal; the interference signal is used to update the adaptive filter, and the updated adaptive filter is used to continue filtering the second voice signal and the third voice signal of the subsequent frames.

[0060] Among them, as a filter in the field of digital signal processing, the adaptive filter can automatically adjust its own filtering parameters according to the characteristics of the input signal to achieve the best filtering effect. As a signal processing technology, the adaptive filter mainly uses the adaptive filter to adjust the parameters of the filter so as to automatically optimize the filtering performance in different signal environments. In this embodiment, the adaptive filter can be adaptively updated according to the calculated interference signal to adjust the parameters of the filter.

[0061] Specifically, for step S2, after obtaining the second voice signal transmitted from the first terminal to the second terminal and the third voice signal transmitted from the second terminal to the first terminal, firstly, multiple adaptive filters are initialized on the first terminal and the second terminal, respectively, for subsequent interference signal simulation and suppression. Then, multiple adaptive filters are used to process the second voice signal and the third voice signal, respectively, to simulate the interference signal that may be contained in the first voice signal. The interference signal, as the target of adaptive filtering processing, needs to be accurately identified and eliminated to ensure the clarity and quality of the call.

[0062] For example, the generation path of the interference signal can be simulated by an adaptive filtering algorithm, such as a voice signal A0 sent by user A is transmitted to the second terminal through a wireless network and / or a wired connection, and then broadcasted through the speaker of the second terminal to generate a voice signal A1. Then the voice signal A1 is propagated through the spatial propagation path and can be collected by the microphone Mic1 of the first terminal, thereby generating a voice signal A2; and the voice signal B0 sent by user B is transmitted to the second terminal through a wireless network and / or a wired connection, and then broadcasted through the speaker of the first terminal to generate a voice signal B1. Then the voice signal B1 is propagated through the spatial propagation path and can be collected by the microphone Mic1 of the first terminal, thereby generating a voice signal B3.

[0063] It can be seen that the first voice signal collected by the microphone Mic1 of the first terminal includes: voice signal A0, voice signal A2 and voice signal B3 sent by user A. Among them, voice signal A2 and voice signal B3 are interference signals compared with voice signal A0.

[0064] It can be understood that the voice signal A2 is similar to the second voice signal transmitted from the first terminal to the second terminal, so the interference signal corresponding to the voice signal A2 can be obtained by simulating the second voice signal. The voice signal B3 is similar to the third voice signal transmitted from the second terminal to the first terminal, so the interference signal corresponding to the voice signal B3 can be obtained by simulating the third voice signal.

[0065] Specifically, an initial adaptive filter is used to simulate the second speech signal to obtain an interference signal similar to the second speech signal, and then another initial adaptive filter is used to simulate the third speech signal to obtain an interference signal similar to the third speech signal. The interference signal is used to simulate the noise generated by acoustic feedback.

[0066] In addition, according to the simulated interference signal, the parameters of the adaptive filter are updated to improve the accuracy of the simulation and prepare for the filtering process for the speech signal processing of the subsequent frame.

[0067] S3. Suppress the interference signal in the first speech signal to obtain a target speech signal after suppression.

[0068] Among them, suppression processing refers to removing interference signals from voice signals through technical means such as adaptive filters to improve the clarity and quality of calls.

[0069] It can be understood that the target voice signal is a pure voice signal after the interference signal is filtered out from the first voice signal. As the final voice signal output by the communication system, the target voice signal directly affects the user's call experience and the quality of voice communication. After the howling suppression and noise elimination processing, the target voice signal is the pure voice that the user expects to hear, without any interference components.

[0070] Specifically, for step S3, the updated adaptive filter is used to identify and suppress the interference signal in the first speech signal, so as to extract a clear target speech signal; when the adaptive filter completely matches the actual feedback path, it can simulate a feedback sound that is completely consistent with the original acoustic path. At this time, the actual feedback path can be cut off to fundamentally eliminate the howling.

[0071] It can be understood that the processing of voice signals is time-sequential. The first voice signal of the first frame can be regarded as a voice signal without noise interference and is directly transmitted to the second terminal. At this time, the second voice signal of the second frame transmitted to the second terminal is almost the same as the first voice signal. The first voice signal after the second frame may include a noise interference signal. Then, the second voice signal and the third voice signal after the second frame are adaptively filtered by the initial multiple adaptive filters to obtain the interference signal in the first voice signal of the corresponding frame; then, the interference signal in the first voice signal after the second frame is suppressed to obtain the target voice signal after the suppression processing and transmit it to the second terminal. The second voice signal transmitted to the second terminal is the target voice signal after the suppression processing. Then, the first voice signal of the next frame is processed based on the second voice signal and the third voice signal of the previous frame.

[0072] In summary, this embodiment achieves the fundamental elimination of howling by accurately simulating and suppressing feedback sound, rather than just reducing the loudness of howling, so that users can enjoy a clearer and interference-free voice communication experience during a call; and the adaptive filter can automatically adjust according to changes in the real-time call environment to maintain an efficient howling suppression effect. The entire process does not require manual intervention, and howling suppression is automatically performed to adapt to a variety of communication environments. It can be seen that this embodiment reduces the unstable factors caused by howling by eliminating howling, and improves the stability and reliability of the communication system.

[0073] In some embodiments, the first voice signal in this embodiment includes a voice signal sent by the first user, a voice signal corresponding to a first interference path between the first terminal and the second terminal, and a voice signal corresponding to a second interference path of the first terminal itself.

[0074] Interference paths usually refer to various unwanted signal paths that voice signals are subject to during transmission or recording. These signals will reduce the quality of voice signals and affect the clarity and intelligibility of voice. Interference paths can include the following types: echo, reverberation, environmental noise, electronic equipment noise, interfering sound sources, radio frequency interference, etc.

[0075] The first interference path may be: the sound played by user A through the speaker of the first terminal is propagated through space to the microphone of the first terminal, forming a path from A1 to A2, and the sound played by user B through the speaker of the second terminal is propagated through space to the microphone of the second terminal, forming a path from B1 to B2, such as Figure 3 The two paths in 2.

[0076] The second interference path may be: the sound played by user A through the speaker of the first terminal is propagated through space and collected by the microphone of the first terminal, forming a path from B1 to B3, and the sound played by user B through the speaker of the second terminal is propagated through space and collected by the microphone of the second terminal, forming a path from A1 to A3, such as Figure 3 Path 3 and Path 3' in .

[0077] Specifically, the first voice signal obtained by the first terminal in this embodiment is mainly the sound of the user speaking directly into the microphone of the first terminal, and may also include a voice signal (A2) corresponding to the first interference path between the first terminal and the second terminal, and a voice signal (B3) corresponding to the second interference path of the first terminal itself, wherein A2 is the sound played by the speaker of the second terminal, which is a signal collected by the microphone of the first terminal after propagation through space; B3 is the sound played by the speaker of the first terminal, which is also a signal collected by the microphone of the first terminal after propagation through space.

[0078] This embodiment can effectively reduce or eliminate howling and echo by accurately identifying and suppressing the interference signal in the first voice signal, thereby significantly improving the clarity and quality of the call.

[0079] In some embodiments, the first interference path in this embodiment is a path for transmitting voice signals through space between the first terminal and the second terminal; the second interference path is a path for transmitting voice signals through space between the first terminal and the first terminal.

[0080] The path for transmitting a voice signal through space can be: user A plays sound through the speaker of the first terminal, the sound propagates through the air in the space, and is captured by the microphone of the first terminal; and user B plays sound through the speaker of the second terminal, the sound propagates through the air in the space where user A is located, and is captured by the microphone of the first terminal corresponding to user A.

[0081] Specifically, the first interference path in this embodiment may refer to the path for transmitting voice signals through space between the first terminal and the second terminal, that is, the path for the sound played by the far-end (second terminal) speaker (Speaker2) to propagate through the air to the near-end (first terminal) microphone (Mic1).

[0082] The second interference path may refer to a path for the first terminal itself to transmit a voice signal through space, that is, a path for the sound played by the near-end (first terminal) speaker (Speaker1) to propagate through the air to its own microphone (Mic1).

[0083] This embodiment can more accurately identify and distinguish the source of the interference signal by clearly distinguishing the two interference paths, so as to more effectively perform suppression processing. Since howling and echo may come from different paths, clarifying these interference paths helps the adaptive filter to more accurately simulate and offset these interferences, thereby improving the effect of howling suppression.

[0084] In some embodiments, the second voice signal includes a voice signal transmitted from the first terminal to the second terminal via the target path; and the third voice signal includes a voice signal transmitted from the second terminal to the first terminal via the target path.

[0085] Specifically, the second voice signal in this embodiment may be a voice signal transmitted from the first terminal to the second terminal through the target path, and the third voice signal may be a voice signal transmitted from the second terminal to the first terminal through the target path. The target path may be regarded as a clean voice signal path without interference, and may be a path for transmitting signals between the first terminal and the second terminal through a wired / wireless network.

[0086] This embodiment clarifies the direction of signal flow between the two terminals by distinguishing the second voice signal from the third voice signal, which helps to better understand and control the signal transmission process, thereby adopting different processing strategies for signals in different directions and improving the effect of voice signal processing.

[0087] In some embodiments, the adaptive filter includes a first adaptive filter and a second adaptive filter; the interference signal includes a first interference signal and a second interference signal;

[0088] Step S2 "performing adaptive filtering processing on the second speech signal and the third speech signal based on the initial multiple adaptive filters to obtain the interference signal in the simulated first speech signal" may specifically include:

[0089] S21. Filtering the second speech signal through the first adaptive filter to obtain a first interference signal corresponding to the first interference path in the simulated first speech signal;

[0090] S22. Filter the third speech signal through the second adaptive filter to obtain a second interference signal corresponding to the second interference path in the simulated first speech signal.

[0091] It can be understood that the adaptive filter in this embodiment may include a first adaptive filter and a second adaptive filter. The first adaptive filter is used to simulate the second voice signal to simulate the first interference signal in the first voice signal; the second adaptive filter is used to simulate the third voice signal to simulate the second interference signal in the first voice signal.

[0092] The first interference signal may be: after the first terminal transmits the sound signal A0 emitted by user A to the second terminal and plays the emitted sound signal A1, the sound signal A1 is propagated through the air (such as Figure 3 The first interference signal in the first voice signal collected by the first terminal may be the interference voice signal A2.

[0093] It can be understood that the first terminal will suppress the first voice signal collected by the first terminal, and the second terminal will also suppress the voice signal collected by the second terminal itself. Similarly, for the second terminal, the voice signal collected by the second terminal may also include two interference signals, where the first interference signal can be: after the second terminal transmits the sound signal B0 emitted by user B to the first terminal and plays the emitted sound signal B1, the sound signal B1 is transmitted through the air (such as Figure 3 Path 2 in the figure) is an interfering speech signal B2 collected by the microphone of the second terminal.

[0094] The second interference signal may be: after the second terminal transmits the sound signal B0 sent by user B to the first terminal and plays the sound signal B1 sent, the sound signal B1 is transmitted through the air (such as Figure 3 The second interference signal in the first voice signal collected by the first terminal may be the interference voice signal B3.

[0095] Similarly, for the second terminal, the second interference signal in the voice signal collected by the second terminal may be: after the first terminal transmits the sound signal A0 emitted by user A to the second terminal and plays the emitted voice signal A1, the voice signal A1 is propagated through the air (such as Figure 3 The interfering speech signal A3 is collected by the microphone of the second terminal in the path 3').

[0096] The first adaptive filter is used to eliminate or reduce the interference signal (i.e., the interference voice signal A2) in the voice signal collected by the first terminal, which is caused by the sound emitted by the speaker of the second terminal propagating through the spatial path and being collected by the microphone of the first terminal. The second adaptive filter is used to eliminate or reduce the interference signal (i.e., the interference voice signal B3) in the voice signal collected by the first terminal, which is caused by the sound emitted by the speaker of the first terminal propagating through the spatial path and being collected by the microphone of the same first terminal.

[0097] Specifically, the adaptive filter in this embodiment may include a first adaptive filter of the first terminal and a second adaptive filter of the second terminal. First, the corresponding adaptive filters are initialized on the first terminal and the second terminal, respectively, for simulating and suppressing interference signals. The first adaptive filter is responsible for simulating and suppressing interference signals from the first interference path, and the second adaptive filter is responsible for simulating and suppressing interference signals from the second interference path. When the first terminal receives the second voice signal, the first adaptive filter is used to filter it to simulate the first interference signal. When the first terminal receives the third voice signal, the second adaptive filter is used to filter it to simulate the second interference signal.

[0098] Each adaptive filter in this embodiment simulates an interference signal corresponding to the corresponding interference path according to its parameters. The interference signal refers to the noise component appearing in the original voice signal. By accurately simulating and suppressing the interference signal, the clarity and quality of the call are significantly improved. Subsequently, the parameters of each adaptive filter are updated according to the simulated interference signal to improve the accuracy of the simulation and the filtering effect.

[0099] In some embodiments, step S21 of "filtering the second speech signal through a first adaptive filter to obtain a first interference signal corresponding to the first interference path in the simulated first speech signal" includes:

[0100] The second voice signal is filtered by the first adaptive filter according to the current filter parameters to obtain a first interference signal corresponding to the first interference path in the simulated first voice signal; the first interference signal is a simulated signal having similar characteristics to the second voice signal.

[0101] The filter parameters of an adaptive filter may refer to a series of values ​​that define and adjust the behavior of the filter, such as filter coefficients, step size parameters, and error feedback coefficients, which determine how the filter responds to signals with different frequency components. In an adaptive filter, these parameters can be dynamically adjusted according to the characteristics of the input signal to optimize the filtering effect.

[0102] Among them, the analog signal with similar characteristics refers to an analog signal that has a frequency component similar to the actual interference signal, a time waveform similar to the time waveform of the actual interference signal, and an amplitude change similar to the amplitude change of the actual interference signal.

[0103] Specifically, this embodiment uses a first adaptive filter to filter the second voice signal according to current filter parameters. The simulated first interference signal obtained by filtering is a signal having similar characteristics to the second voice signal. The first interference signal represents the interference transmitted back to the first terminal microphone through the first interference path (i.e., the spatial path).

[0104] In a specific implementation process, the parameters of the first adaptive filter are initialized through the first terminal so that the filter adapts to the expected interference signal characteristics; the second voice signal is received from the second terminal at the first terminal, and the second voice signal is a signal transmitted through the target path; the second voice signal is filtered using the first adaptive filter to simulate the first interference signal, and the first interference signal simulates the echo or noise that may be generated when the second voice signal propagates through the first interference path.

[0105] This embodiment ensures that the first adaptive filter can accurately simulate the interference signal generated by the first interference path. The interference signal after accurate simulation can be suppressed more effectively, thereby significantly reducing howling and improving call quality, so that the user can hardly feel the existence of the interference signal during the call, effectively improving the call experience.

[0106] In some embodiments, step S22 of "filtering the third speech signal through a second adaptive filter to obtain a second interference signal corresponding to the second interference path in the simulated first speech signal" includes:

[0107] The third voice signal is filtered by the second adaptive filter according to the current filter parameters to obtain a second interference signal corresponding to the second interference path in the simulated first voice signal; the second interference signal is a simulated signal having similar characteristics to the third voice signal.

[0108] Filter parameters refer to a series of values ​​that define and adjust the behavior of the filter, such as filter coefficients, step size parameters, and error feedback coefficients. These parameters determine how the filter responds to signals with different frequency components. In adaptive filters, these parameters can be dynamically adjusted according to the characteristics of the input signal to optimize the filtering effect.

[0109] Specifically, this embodiment uses a second adaptive filter to filter the third voice signal according to current filter parameters. The simulated second interference signal obtained by filtering is a signal having similar characteristics to the third voice signal. The second interference signal represents the interference transmitted back to the microphone of the first terminal through the second interference path (i.e., the spatial path inside the first terminal).

[0110] In a specific embodiment, the parameters of the second adaptive filter are initialized through the first terminal so that the filter adapts to the expected internal interference signal characteristics; the first terminal receives a third voice signal from the second terminal, and the third voice signal is a signal transmitted through a target path (which may be wired or wireless); the third voice signal is filtered using the second adaptive filter to simulate a second interference signal, and the second dry signal is used to simulate the echo or noise that may be generated when the third voice signal propagates through a spatial path inside the first terminal.

[0111] This embodiment ensures that the second adaptive filter can accurately simulate the internal interference signal generated by the second interference path, so that the accurately simulated interference signal can be more effectively suppressed, thereby significantly reducing howling and echo and improving call quality. Therefore, the user can hardly feel the existence of the internal interference signal during the call, effectively improving the user's call experience.

[0112] In some embodiments, the speech signal processing method further includes:

[0113] Based on the first interference signal and the speech signal corresponding to the first interference path between the first terminal and the second terminal, updating the first adaptive filter corresponding to the first interference path;

[0114] Based on the second interference signal and the speech signal corresponding to the second interference path of the first terminal itself, the second adaptive filter corresponding to the second interference path is updated.

[0115] Specifically, the voice signal processing method of this embodiment also provides a method for updating each adaptive filter based on an interference signal. First, during a call, a simulated first interference signal is obtained by processing the second voice signal with a first adaptive filter, and a simulated second interference signal is obtained by processing the third voice signal with a second adaptive filter; by analyzing the characteristics corresponding to the first interference signal and the second interference signal, including frequency, amplitude, delay, etc., respectively, to determine the adjustment direction of the parameters of the first adaptive filter and the second adaptive filter; according to the analysis results of the interference signal, the parameters of each adaptive filter are adjusted to improve the performance of subsequent simulation and suppression of interference signals.

[0116] In addition, a real-time update mechanism can be designed to ensure that filter parameters can be dynamically adjusted based on real-time feedback during a call, and filter performance can be evaluated to ensure that the updated filter can effectively simulate and suppress interference signals.

[0117] This embodiment improves the adaptability and howling suppression effect of the filter by ensuring that the adaptive filter can be adjusted according to the real-time call conditions. By accurately updating the filter parameters, the call quality is significantly improved and the howling and echo are reduced. At the same time, dynamic adjustment of the filter parameters helps to optimize the overall performance of the call system and improve the efficiency of signal processing.

[0118] In some embodiments, the interference signal includes a first interference signal and a second interference signal; suppressing the interference signal in the first speech signal to obtain a target speech signal after suppression processing includes:

[0119] Suppression processing is performed on the first interference signal and the second interference signal in the first speech signal to remove the first interference signal and the second interference signal from the first speech signal to obtain a target speech signal after suppression processing.

[0120] Specifically, after filtering the second voice signal through the first adaptive filter to obtain the first interference signal corresponding to the first interference path in the simulated first voice signal, and filtering the third voice signal through the second adaptive filter to obtain the second interference signal corresponding to the second interference path in the simulated first voice signal, it is necessary to suppress the first interference signal and the second interference signal in the first voice signal, and suppress the extracted interference signal, so as to reconstruct the target voice signal, that is, the pure voice that the user wants to transmit, after subtracting the first interference signal and the second interference signal from the first voice signal.

[0121] In addition, this embodiment can also update the parameters of the first adaptive filter and the second adaptive filter according to the result of the suppression process, so as to more accurately simulate and suppress the interference signal in future call frames.

[0122] This embodiment significantly improves the clarity and quality of calls by effectively suppressing interference signals. The adaptive filter can be adjusted according to real-time call conditions, thereby improving the performance of the filter and being able to adapt to a variety of communication scenarios, including short-distance calls, long-distance calls, and different communication environments.

[0123] In order to facilitate understanding of the speech signal processing method provided by this embodiment, this embodiment also provides a specific implementation of the speech signal processing method.

[0124] like Figure 3 As shown, A and B are two users having a conversation, Mic1 and Speaker1 are the microphone and player of the first terminal corresponding to user A, and Mic2 and Speaker2 are the microphone and player of the second terminal corresponding to user B.

[0125] When the first terminal and the second terminal are talking, user A speaks to Mic1, and the voice signal collected by Mic1 is recorded as A0. The voice signal A0 passes through the wired or wireless network path between the first terminal and the second terminal (such as Figure 3 The voice signal A1 broadcast by Speaker 2 passes through the spatial path between user A and user B (such as Figure 3 Path 2) in the figure can be collected by Mic1 of the first terminal, so the first voice signal collected by Mic1 of the first terminal includes voice signal A2.

[0126] User B speaks to Mic2 of the second terminal, and the voice collected by Mic2 is recorded as B0. The voice signal B0 passes through the wired or wireless network path between the first terminal and the second terminal (such as Figure 3 Path 1 in the figure is transmitted to Speaker 1 of the first terminal, and after being broadcast by Speaker 1, it becomes voice signal B1. Voice B1 passes through the spatial path between the first terminal and the second terminal (such as Figure 3 The first voice signal is then collected by Mic1 of the first terminal, so the first voice signal collected by Mic1 of the first terminal includes the voice signal B3.

[0127] In this case, the sound collected by Mic1 of the first terminal includes voice signal A0 (the voice that user A wants to send out) + voice signal A2 (the voice that A1 transmits to Mic1 of the first terminal through path 2 after passing through Speaker2 of the second terminal) + voice signal B3 (the voice B1 sent by user B passes through Speaker1 of the first terminal and is collected by Mic1 of the first terminal). Among them, voice signal A2 and voice signal B3 are interference signals.

[0128] During the voice call between user A and user B, the voice is constantly amplified and oscillated between the Mic and Speaker of the first terminal and the Mic and Speake of the second terminal, which can easily cause howling during the call. At this time, if the interference voice signal A2 and the interference voice signal B3 in the first voice signal collected by Mic1 of the first terminal are removed, a noise-free voice signal A0 can be obtained. The first terminal transmits the voice signal A0 after interference suppression processing to the second terminal through the wired or wireless network path between the first terminal and the second terminal. The voice signal A1 broadcasted by Speaker2 of the second terminal is the purer voice signal corresponding to the voice signal A0 that user A wants user B to hear.

[0129] Similarly, when user A and user B are talking, user B speaks to the Mic2 of the second terminal, and the voice collected by Mic2 is recorded as B0. B0 is transmitted to the Speaker1 of the first terminal through path 1, and broadcasted by the first terminal Speaker1 to become B1. The voice B1 broadcasted by Speaker1 of the first terminal can be collected by Mic2 of the second terminal through path 2. At this time, the voice B2 generated based on voice B1 is collected by Mic2 of the second terminal.

[0130] User A speaks to the microphone Mic1 of the first terminal, and the voice signal collected by the microphone Mic1 of the first terminal is recorded as A0. The voice signal A0 passes through the circuit (such as Figure 3 Path 1 in the figure) is transmitted to the speaker Speaker2 of the second terminal of user B and becomes a voice signal A1; after the voice signal A1 is broadcasted by Speaker2 of the second terminal, it passes through the spatial path between Speaker2 and Mic2 of the second terminal (such as Figure 3 After being transmitted through the path 3' in the figure, it is collected by Mic2 of the second terminal. At this time, the collected voice is the voice signal A3.

[0131] The voice signal collected by Mic2 of the second terminal of user B may include: voice signal B0 (the voice that user B wants to send out) + voice signal B2 (the voice that B1 passes through Speaker1 of the first terminal and is transmitted to Mic2 of the second terminal through the space between users A and B) + voice signal A3 (the voice A1 sent by user A passes through Speaker2 of the second terminal of user B and is collected by Mic2 of the second terminal). Among them, voice signal B2 and voice signal A3 are interference voice signals.

[0132] Similarly, the situation of end B (second terminal) is the same as that of end A (first terminal). The second terminal also removes the interfering voice signal B2 and the interfering voice signal A3 from the voice signal collected by its Mic2, so as to obtain the voice signal B0, which is the voice B1 that user B wants user A to hear. The second terminal transmits the voice signal B0 to the first terminal through the wired or wireless network path between the first terminal and the second terminal. The voice signal B1 broadcasted by Speaker1 of the first terminal is the relatively pure voice signal corresponding to the voice signal B0 that user B wants user A to hear.

[0133] The speech signal processing method provided in this embodiment uses the idea of ​​feedback cancellation to control the acoustic path feedback, mainly by deducting the feedback signal component from the microphone signal to obtain the residual E(ω) / E'(ω), that is, the speech signal after removing the interference signal. The specific process is as follows:

[0134] like Figure 4-Figure 5 As shown, when A and B are talking, taking A as the execution subject, the microphone of A collects voice signal "A", and the microphone of B collects voice signal "B". A will transmit the signal "A1" after interference suppression to B, and A will also receive the voice signal "B1" after interference suppression transmitted by B.

[0135] The voice A collected by the A terminal is the first voice signal, including: A0+A2+B3. Among them, the voice signal A0 [represents the voice that the user A wants to send out (transmitted through the path 1 between the first terminal and the second terminal)], the voice signal A2 [the voice signal A0 is transmitted to the second terminal and then played out through the second terminal as the voice signal A1, and the voice signal A1 passes through the second terminal Speaker2 and then is transmitted through the space path between the first terminal and the second terminal (such as Figure 3 Path 2) in the figure, voice signal A2 collected by Mic1 of the first terminal], and voice signal B3 [voice signal B0 sent by user B is transmitted to the first terminal and then played out as voice signal B1 by the first terminal, and voice signal B3 collected by Mic1 of the first terminal after being played out by Speaker1 of the first terminal].

[0136] At the same time, end A will receive the voice signal B1 (second voice signal) after interference suppression processing transmitted by end B through path 1. (Path 1 can be a wired circuit or a stable wireless network link)

[0137] Specifically, the A end can input the voice signal "A1" into the first adaptive filter (fliter1(ω)) for processing. The adaptive rate filter filters the input voice signal "A1" according to the current filter parameters. Specifically, the voice signal A1 can be simulated and estimated to obtain a filtered output signal (first estimated signal). The output signal is the interference voice signal in the simulated signal A that is similar to A1 and is transmitted to the A end Mic1 through the spatial path (path 2) between AB after passing through the Speaker2 at the B end, that is, the interference voice signal A2 in the first voice signal A.

[0138] At the same time, end A inputs the received voice signal "B1" to the second adaptive filter (fliter2(ω)) for processing. The adaptive rate filter filters the input voice signal "B1" according to the current filter parameters, that is, it simulates and estimates the voice signal B1 to obtain a filtered output signal (second estimated signal). The output signal is the interference voice signal in the simulated signal A that is similar to the voice signal B0, that is, the interference voice signal B3 in the first voice signal A.

[0139] Then, the interference speech signal A2 and the interference speech signal B3 in the first speech signal "A" are deducted (ie, suppressed), thereby obtaining a target speech signal after interference suppression.

[0140] Similarly, for the B end, that is, the second terminal, the second terminal inputs the voice signal B1 into the first adaptive filter (fliter1(ω)) for adaptive filtering processing. The adaptive rate filter filters the input voice signal "B1" according to the current filter parameters. Specifically, the voice signal B1 can be simulated and estimated to obtain a filtered output signal. The output signal is a simulated interference voice signal similar to B1, which passes through the Speaker1 at the A end and is transmitted to the Mic2 at the B end through the spatial path (path 2) between AB, that is, the interference voice signal B2 in the voice signal B.

[0141] At the same time, end B inputs the received voice signal "A1" to the second adaptive filter (fliter2'(ω)) for processing. The adaptive rate filter filters the input voice signal "B1" according to the current filter parameters, that is, it simulates and estimates the voice signal A1 to obtain a filtered output signal. The output signal is the interference voice signal in the simulated signal B that is similar to the voice signal A1, that is, the interference voice signal A3 in the voice signal B.

[0142] Among them, since the first adaptive filter used by the first terminal and the first adaptive filter used by the second terminal both simulate the interference voice signal corresponding to the spatial propagation path between the first terminal and the second terminal, the first adaptive filter used by the first terminal and the first adaptive filter used by the second terminal can be the same, that is, both can use fliter1(ω). The second adaptive filter used by the first terminal simulates the interference voice signal corresponding to the spatial propagation path between the speaker of the first terminal and the microphone of the first terminal (that is, the spatial propagation path around the first terminal itself), and the second adaptive filter used by the second terminal simulates the interference voice signal corresponding to the spatial propagation path between the speaker of the second terminal and the microphone of the second terminal (that is, the spatial propagation path around the second terminal itself). Therefore, the second adaptive filter fliter2(ω) used by the first terminal is different from the second adaptive filter fliter2'(ω) used by the second terminal.

[0143] After the first terminal suppresses the collected first voice signal, it updates the adaptive filters flite1(ω) and flite2(ω), and continues to estimate the actual feedback paths G1(ω) and G2(ω) according to the next frame of voice signal input. When flite1(ω) and G1(ω) are fully matched, and G2(ω) and flite2(ω) are fully matched, the generated re-entry feedback sounds filter_A2 and A2, and filter_B3 and B3 are completely consistent, indicating that passing through the filter is equivalent to passing through the original acoustic path, thus realizing the estimation of the acoustic path. At this time, the feedback path can be cut off to fundamentally eliminate the howling problem caused by the feedback sound. The feedback paths G1(ω) and G2(ω) can be estimated by the filter.

[0144] After the second terminal suppresses the collected voice signal, it updates the adaptive filters flite1(ω) and flite2'(ω), and continues to estimate the actual feedback paths G1(ω) and G2'(ω) according to the next frame of voice signal input. When flite1(ω) and G1(ω) are fully matched, and G2'(ω) and flite2'(ω) are fully matched, the generated re-entry feedback sound filter_B2 is completely consistent with B2, and filter_A3 is completely consistent with A3, indicating that passing through the filter is equivalent to passing through the original acoustic path, and the estimation of the acoustic path is realized. At this time, the feedback path can be cut off to fundamentally eliminate the howling problem caused by the feedback sound. The feedback paths G1(ω) and G2'(ω) can be estimated by the filter.

[0145] The updating process of the filter is as follows: using the evaluation criteria of the adaptive filter, such as the standard deviation, etc., to adjust the parameters of the filter according to the evaluation results, thereby realizing the adaptive updating of the filter, so that the speech signal estimated by the adaptive filter is closer to the interference signal.

[0146] like Figure 6 As shown in the figure, flite1(ω), flite2(ω), and flite2'(ω) are the adaptive filters added to the whole system in this solution to suppress howling. Among them, flite1(ω) [the filter corresponding to end A and path 2], flite2(ω) [the filter corresponding to end A and path 3], and flite2'(ω) [the filter corresponding to end B and path 3]. The filter corresponding to path 2 at end B can be the same as the filter flite1(ω) at end A.

[0147] For example, in the actual application scenario, take the communication scenario between the door screen and the door peephole as an example, as shown below Figure 6 The application scenario of this case is applicable to calls between two devices connected by wire, and is also applicable to devices connected by wireless.

[0148] Taking the inside or outside of the door as an example, the process of howling is as follows: Figure 6 As shown, G1(ω) represents Figure 3 The transfer function corresponding to path 2 in is, Figure 5 The adaptive filter fliter1(ω) in is the reconstruction Figure 3 The transfer function G1(ω) of path 2 is infinitely close to the transfer function G1(ω) through adaptability. Then the speech fliteA2(ω) after passing flite1(ω) is infinitely close to the speech A2(ω) passing through path 2.

[0149] Similarly, Figure 5 The adaptive flite2(ω) in the reconstruction diagram reconstructs the transfer function G2(ω) of path 3. The adaptive filter flite2(ω) is infinitely close to the transfer function G2(ω) through adaptability. Therefore, the speech fliteB3(ω) after passing through flite2(ω) and the speech B3(ω) after passing through path 3 are infinitely close.

[0150] When the adaptive filter is not added, the first speech signal A(ω) collected by the microphone at the A end includes:

[0151] A(ω)=A0(ω)+A2(ω)+B3(ω),

[0152] The first voice signal A(ω) is transmitted to the B terminal through path 1, and after being played through the speaker of the B terminal, howling may be generated through amplification. Wherein, A0(ω) represents the voice signal A0 emitted by user A collected by the first terminal. A2(ω) represents that after the voice signal A0 is transmitted to the second terminal through the wireless network and / or wired, it is broadcasted through the speaker of the second terminal to generate a voice signal A1, and then the voice signal A1 is propagated through the spatial propagation path and can be collected by the microphone Mic1 of the first terminal, thereby generating an interference voice signal A2. B3(ω) represents that after the voice signal B0 emitted by user B is transmitted to the second terminal through the wireless network and / or wired, it is broadcasted through the speaker of the first terminal to generate a voice signal B1, and then the voice signal B1 is propagated through the spatial propagation path and can be collected by the microphone Mic1 of the first terminal, thereby generating an interference voice signal B3.

[0153] After adding the adaptive filters flite1(ω) and flite2(ω), the first speech signal A(ω) is suppressed to obtain the target speech signal E(ω) after suppression:

[0154] E(ω)=A(ω)-filterA2(ω)-filterB3(ω),

[0155] That is, E(ω)=A0(ω)+A2(ω)+B3(ω)-filterA2(ω)-filterB3(ω).

[0156] At this time, E(ω)≈A0(ω), achieving the purpose of transmitting a relatively pure voice signal.

[0157] Wherein, filterA2(ω): represents the output of signal A1(ω) after being suppressed by adaptive filter1(ω);

[0158] filterB3(ω): represents the output of signal B1(ω) after being suppressed by adaptive filter2(ω).

[0159] Similarly, the adaptive filter of the second terminal adopts the same working principle, which will not be described in detail here.

[0160] To summarize, the voice signal processing method provided in the embodiment of the present application adds multiple adaptive filters to the duplex call system, adopts the idea of ​​feedback cancellation, calls the adaptive filters at both ends of the duplex call at the same time, estimates the actual feedback path and matches it with the adaptive filter, thereby cutting off the feedback path, achieving the effect of howling suppression during the call, effectively optimizing the short-distance call effect, avoiding the reverberation tail with obvious distortion caused by howling, and thus effectively improving the call voice quality.

[0161] In order to better implement the speech signal processing method of the embodiment of the present application, the embodiment of the present application also provides a speech signal processing device based on the above speech signal processing method. The meanings of the terms are the same as those in the above speech signal processing method, and the specific implementation details can refer to the description in the method embodiment.

[0162] See also Figure 7 , Figure 7 This is a schematic diagram of the structure of a speech signal processing device provided in an embodiment of the present application. The speech signal processing device includes an acquisition module 201, a filtering module 202 and a suppression module 203. Specifically, it can be as follows:

[0163] An acquisition module 201 is used to acquire a first voice signal collected by a first terminal, a second voice signal transmitted by the first terminal to a second terminal, and a third voice signal transmitted by the second terminal to the first terminal;

[0164] The filtering module 202 is used to perform adaptive filtering processing on the second voice signal and the third voice signal based on the initial multiple adaptive filters to obtain the interference signal in the simulated first voice signal; the interference signal is used to update the adaptive filter, and the updated adaptive filter is used to continue filtering the second voice signal and the third voice signal of the subsequent frame;

[0165] The suppression module 203 is used to suppress the interference signal in the first speech signal to obtain a target speech signal after the suppression process;

[0166] In some embodiments, the first voice signal includes a voice signal sent by the first user, a voice signal corresponding to a first interference path between the first terminal and the second terminal, and a voice signal corresponding to a second interference path of the first terminal itself.

[0167] In some embodiments, the first interference path is a path for transmitting a voice signal through space between the first terminal and the second terminal; and the second interference path is a path for transmitting a voice signal through space between the first terminal and the first terminal.

[0168] In some embodiments, the second voice signal includes a voice signal transmitted from the first terminal to the second terminal via the target path; and the third voice signal includes a voice signal transmitted from the second terminal to the first terminal via the target path.

[0169] In some embodiments, the adaptive filter includes a first adaptive filter and a second adaptive filter; the interference signal includes a first interference signal and a second interference signal;

[0170] Filter module, including:

[0171] A first filtering unit, configured to filter the second speech signal through a first adaptive filter to obtain a first interference signal corresponding to the first interference path in the simulated first speech signal;

[0172] The second filtering unit is used to filter the third voice signal through a second adaptive filter to obtain a second interference signal corresponding to the second interference path in the simulated first voice signal.

[0173] In some embodiments, the first filtering unit is specifically used to filter the second voice signal through a first adaptive filter according to current filter parameters to obtain a first interference signal corresponding to the first interference path in the simulated first voice signal; the first interference signal is an analog signal having similar characteristics to the second voice signal.

[0174] In some embodiments, the second filtering unit is specifically used to filter the third voice signal through a second adaptive filter according to current filter parameters to obtain a second interference signal corresponding to the second interference path in the simulated first voice signal; the second interference signal is an analog signal having similar characteristics to the third voice signal.

[0175] In some embodiments, the apparatus further includes an update module, the update module including:

[0176] A first updating unit, configured to update a first adaptive filter corresponding to the first interference path based on the first interference signal and a speech signal corresponding to the first interference path between the first terminal and the second terminal;

[0177] The second updating unit is used to update the second adaptive filter corresponding to the second interference path based on the second interference signal and the speech signal corresponding to the second interference path of the first terminal itself.

[0178] In some embodiments, the interference signal includes a first interference signal and a second interference signal; the suppression module is specifically used to: suppress the first interference signal and the second interference signal in the first speech signal to remove the first interference signal and the second interference signal from the first speech signal to obtain a target speech signal after suppression processing.

[0179] As described above, the embodiment of the present application provides a voice signal processing device, which acquires a first voice signal collected by a first terminal, a second voice signal transmitted from the first terminal to a second terminal, and a third voice signal transmitted from the second terminal to the first terminal through an acquisition module; the second voice signal and the third voice signal are adaptively filtered based on the initial multiple adaptive filters through a filtering module to obtain an interference signal in a simulated first voice signal; the interference signal is used to update the adaptive filter, and the updated adaptive filter is used to continue filtering the second voice signal and the third voice signal of subsequent frames; the interference signal in the first voice signal is suppressed through a suppression module to obtain a target voice signal after suppression. This embodiment adds multiple adaptive filters to a duplex call system, adopts the idea of ​​feedback cancellation, calls adaptive filters at both ends of the duplex call at the same time, estimates the actual feedback path and matches it with the adaptive filter, thereby cutting off the feedback path, achieving the effect of suppressing howling during the call, effectively optimizing the short-distance call effect, and avoiding reverberation tailing with obvious distortion caused by howling, thereby effectively improving the call voice quality.

[0180] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 8 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a voice signal processing method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or it can be a computer device.

[0181] Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0182] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0183] In one embodiment, a computer-readable storage medium is provided, storing a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0184] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0185] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0186] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0187] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the patent application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent application shall be subject to the attached claims.

Claims

1. A speech signal processing method, characterized in that: include: Acquire a first voice signal collected by a first terminal, a second voice signal transmitted by the first terminal to a second terminal, and a third voice signal transmitted by the second terminal to the first terminal; Based on the initial multiple adaptive filters, adaptive filtering is performed on the second voice signal and the third voice signal respectively to obtain a simulated interference signal in the first voice signal; the interference signal is used to update the adaptive filter, and the updated adaptive filter is used to continue filtering the second voice signal and the third voice signal of the subsequent frame; The interference signal in the first speech signal is suppressed to obtain a target speech signal after the suppression processing.

2. The speech signal processing method according to claim 1, characterized in that: The first voice signal includes a voice signal sent by the first user, a voice signal corresponding to a first interference path between the first terminal and the second terminal, and a voice signal corresponding to a second interference path of the first terminal itself.

3. The speech signal processing method according to claim 2, characterized in that: The first interference path is a path for transmitting a voice signal through space between the first terminal and the second terminal; the second interference path is a path for transmitting a voice signal through space between the first terminal and the second terminal.

4. The speech signal processing method according to claim 1, characterized in that: The second voice signal includes a voice signal transmitted from the first terminal to the second terminal through the target path; and the third voice signal includes a voice signal transmitted from the second terminal to the first terminal through the target path.

5. The speech signal processing method according to claim 1, characterized in that: The adaptive filter includes a first adaptive filter and a second adaptive filter; the interference signal includes a first interference signal and a second interference signal; The method of performing adaptive filtering processing on the second voice signal and the third voice signal based on the initial multiple adaptive filters to obtain a simulated interference signal in the first voice signal includes: Performing filtering processing on the second speech signal through the first adaptive filter to obtain a first interference signal corresponding to a first interference path in the simulated first speech signal; The third voice signal is filtered by the second adaptive filter to obtain a simulated second interference signal corresponding to the second interference path in the first voice signal.

6. The speech signal processing method according to claim 5, characterized in that: The filtering process of the second speech signal by the first adaptive filter to obtain a first interference signal corresponding to a first interference path in the simulated first speech signal includes: The second voice signal is filtered by the first adaptive filter according to the current filter parameters to obtain a first interference signal corresponding to the first interference path in the simulated first voice signal; the first interference signal is a simulated signal having similar characteristics to the second voice signal.

7. The speech signal processing method according to claim 5, characterized in that: The filtering process of the third voice signal by the second adaptive filter to obtain a second interference signal corresponding to the second interference path in the simulated first voice signal includes: The third voice signal is filtered by the second adaptive filter according to the current filter parameters to obtain a second interference signal corresponding to the second interference path in the simulated first voice signal; the second interference signal is a simulated signal having similar characteristics to the third voice signal.

8. The speech signal processing method according to claim 5, characterized in that: The method further comprises: updating a first adaptive filter corresponding to the first interference path based on the first interference signal and a speech signal corresponding to a first interference path between the first terminal and the second terminal; Based on the second interference signal and the speech signal corresponding to the second interference path of the first terminal itself, the second adaptive filter corresponding to the second interference path is updated.

9. The speech signal processing method according to claim 5, characterized in that: The interference signal includes a first interference signal and a second interference signal; and the suppressing process is performed on the interference signal in the first speech signal to obtain a target speech signal after the suppression process, including: Suppression processing is performed on the first interference signal and the second interference signal in the first speech signal to remove the first interference signal and the second interference signal from the first speech signal to obtain a target speech signal after suppression processing.

10. A speech signal processing device, characterized in that: include: An acquisition module, used to acquire a first voice signal collected by a first terminal, a second voice signal transmitted by the first terminal to a second terminal, and a third voice signal transmitted by the second terminal to the first terminal; A filtering module, configured to perform adaptive filtering processing on the second voice signal and the third voice signal respectively based on the initial multiple adaptive filters to obtain a simulated interference signal in the first voice signal; the interference signal is used to update the adaptive filter, and the updated adaptive filter is used to continue filtering the second voice signal and the third voice signal of subsequent frames; The suppression module is used to suppress the interference signal in the first speech signal to obtain a target speech signal after suppression.

11. An electronic device comprising at least one processor and at least one memory, wherein: The memory stores a computer program, wherein the processor implements the steps of the speech signal processing method according to any one of claims 1 to 9 when executing the computer program.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the speech signal processing method according to any one of claims 1 to 9 are implemented.

13. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the speech signal processing method according to any one of claims 1 to 9 are implemented.

Images