A method, device, chip and module device for processing voice echo

Abstract

The present application discloses a method, device, chip and module device for processing voice echo, the method comprising: playing a first sound signal of a first user sent by a remote device through a loudspeaker; collecting sound signals through a microphone to obtain a second sound signal, the second sound signal comprising a third sound signal of the second user and an echo signal corresponding to the first sound signal; eliminating the echo signal in the second sound signal based on the first sound signal to obtain a residual signal; performing sub-band division processing on the residual signal to obtain a first signal on one or more sub-bands; determining the sound field state of the first signal on the one or more sub-bands; processing the first signal on the one or more sub-bands based on the sound field state to obtain a third sound signal; and sending the third sound signal to the remote device. Based on the method described in the present application, the quality of voice communication can be improved and the voice call experience can be enhanced.

Description

Technical Field

[0001] The present invention relates to the field of computers, and in particular to a method, device, chip and module equipment for processing voice echo. Background Art

[0002] During real-time voice communication, the far-end voice signal is played back through the near-end sound reinforcement system, picked up directly or indirectly by the microphone after reflection, and transmitted to the far-end, resulting in acoustic echo. If this echo is not processed, the other end will always hear their own voice while speaking, seriously affecting the call experience. Therefore, how to process echo in voice communication to improve voice communication quality is an unresolved problem. Summary of the Invention

[0003] The present application provides a method, device, chip, and module device for processing voice echo, which can improve the quality of voice communication and enhance the voice call experience.

[0004] In a first aspect, the present application provides a method for processing a voice echo, the method comprising: playing a first sound signal of a first user sent by a remote device through a loudspeaker; collecting the sound signal through a microphone to obtain a second sound signal, the second sound signal including a third sound signal of the second user and an echo signal corresponding to the first sound signal; eliminating the echo signal in the second sound signal based on the first sound signal to obtain a residual signal; performing sub-band division processing on the residual signal to obtain a first signal on one or more sub-bands; determining a sound field state of the first signal on the one or more sub-bands; processing the first signal on the one or more sub-bands based on the sound field state to obtain a third sound signal; and sending the third sound signal to the remote device.

[0005] Based on the method described in the first aspect, after the residual signal is divided into sub-bands, the sound field state is accurately judged based on the first signal on the sub-band. At the same time, different echo suppression processing methods are used for different sound field states to more effectively suppress echoes, which is conducive to improving the quality of voice communication and enhancing the voice call experience.

[0006] In one possible implementation, canceling the echo signal in the second sound signal based on the first sound signal to obtain a residual signal includes: filtering the first sound signal using an adaptive filter to obtain an estimated echo signal; and canceling the echo signal in the second sound signal based on the estimated echo signal to obtain a residual signal. This approach can effectively cancel linear echo signals.

[0007] In a possible implementation, the method further includes: adjusting an update rate of the adaptive filter based on the sound field state. Based on this approach, the accuracy of linear echo suppression can be improved.

[0008] In one possible implementation, the sound field state includes a double talk state, an abnormal echo state, and a far-end single talk state. Determining the sound field state of the first signal on the one or more sub-bands includes: if a speech signal exists in the second sound signal and the first sound signal, calculating a leakage coefficient based on the residual signal and the estimated echo signal, the leakage coefficient being used to indicate the degree of signal leakage echo; if the leakage coefficient is greater than a first threshold, determining that the sound field state of all the first signals on the one or more sub-bands is the abnormal echo state; if the leakage coefficient is less than or equal to the first threshold, determining whether an echo suppression ratio of the first signal on a first sub-band is greater than a second threshold, the first sub-band being any one of the one or more sub-bands, the echo suppression ratio being used to indicate the degree of suppression of the echo signal in the sound signal; if the echo suppression ratio of the first signal on the first sub-band is greater than the second threshold, determining that the sound field state of the first signal on the first sub-band is the far-end single talk state; and if the echo suppression ratio of the first signal on the first sub-band is less than or equal to the second threshold, determining that the sound field state of the first signal on the first sub-band is the double talk state. Based on this approach, the sound field state of the first signal in each sub-band can be effectively determined.

[0009] In one possible implementation, if a speech signal is present in the residual signal, the second threshold is equal to the third threshold; if a speech signal is not present in the residual signal, the second threshold is equal to the fourth threshold, and the third threshold is greater than the fourth threshold. This approach can improve the accuracy of sound field state determination.

[0010] In one possible implementation, processing the first signal on one or more subbands based on the sound field state to obtain a third sound signal includes: performing residual echo suppression on the first signal on one or more subbands based on the sound field state to obtain a second signal on one or more subbands; performing nonlinear echo suppression on the second signal on one or more subbands based on the sound field state to obtain a third signal on one or more subbands; and combining the third signals on one or more subbands to obtain a third sound signal. This approach enables effective echo suppression.

[0011] In one possible implementation, residual echo suppression is performed on a first signal on one or more subbands based on the sound field state to obtain a second signal on one or more subbands, including: determining an initial residual echo estimate corresponding to the first signal on the one or more subbands based on the leakage coefficient and the estimated echo signal, or determining an initial residual echo estimate corresponding to the first signal on the one or more subbands based on the leakage coefficient and the first sound signal; determining a residual echo gain corresponding to the first signal on the one or more subbands based on the sound field state; determining a residual echo estimate corresponding to the first signal on the one or more subbands based on the residual echo gain and the initial residual echo estimate; and performing residual echo suppression on the first signal on the one or more subbands based on the residual echo estimate to obtain the second signal on the one or more subbands. Based on this approach, the accuracy of the residual echo estimate corresponding to the first signal on each subband can be improved, thereby improving the effectiveness of residual echo suppression.

[0012] In one possible implementation, determining an initial residual echo estimate corresponding to a first signal on one or more subbands based on the leakage coefficient and the estimated echo signal includes: determining a fourth sound signal based on the estimated echo signal, the first sound signal, and the second sound signal; and determining the initial residual echo estimate corresponding to the first signal on one or more subbands based on the leakage coefficient and the fourth sound signal. This approach is more conducive to improving the accuracy of the residual echo estimate corresponding to the first signal on each subband.

[0013] In one possible implementation, performing residual echo suppression on a first signal on one or more subbands based on the residual echo estimate to obtain a second signal on the one or more subbands includes: determining a first suppression coefficient corresponding to the first signal on the one or more subbands based on the residual echo estimate; and performing residual echo suppression on the first signal on the one or more subbands based on the first suppression coefficient to obtain the second signal on the one or more subbands. In this approach, the first suppression coefficient corresponding to the first signal on the one or more subbands is directly determined based on the magnitude of the residual echo estimate, which helps improve the efficiency of echo suppression.

[0014] In one possible implementation, performing residual echo suppression on a first signal on one or more subbands based on the residual echo estimate to obtain a second signal on one or more subbands includes: determining a noise estimate corresponding to the first signal on the one or more subbands; determining a fused sound signal on the one or more subbands based on the noise estimate and the residual echo estimate; and performing residual echo suppression on the first signal on the one or more subbands based on the fused sound signal to obtain the second signal on the one or more subbands. In this approach, incorporating residual echo into the noise estimate for fused noise reduction further improves the quality of voice communication.

[0015] In one possible implementation, performing residual echo suppression on a first signal in one or more subbands based on the sound field state to obtain a second signal in one or more subbands includes: determining a second suppression coefficient corresponding to the first signal in the one or more subbands based on the sound field state; and performing residual echo suppression on the first signal in the one or more subbands based on the second suppression coefficient to obtain the second signal in the one or more subbands. Based on this approach, the degree of residual echo suppression is adjusted according to the sound field state to improve the accuracy of residual echo suppression.

[0016] In one possible implementation, the second suppression coefficient for the dual-talk state is greater than that for the far-end single-talk state, and the second suppression coefficient for the far-end single-talk state is greater than that for the abnormal echo state. This approach ensures the continuity of near-end speech in the dual-talk state while ensuring clean echo suppression, and effectively identifies and suppresses abnormal echoes caused by sudden path changes.

[0017] In one possible implementation, performing nonlinear echo suppression on the second signal in one or more subbands based on the sound field state to obtain a third signal in the one or more subbands includes: determining a third suppression coefficient corresponding to the second signal in the one or more subbands based on the sound field state; and performing nonlinear echo suppression on the second signal in the one or more subbands based on the third suppression coefficient to obtain the third signal in the one or more subbands. Based on this approach, the degree of nonlinear echo suppression is adjusted according to the sound field state to improve the accuracy of nonlinear echo suppression.

[0018] In one possible implementation, the third suppression coefficient for the dual-talk state is greater than that for the far-end single-talk state, and the third suppression coefficient for the far-end single-talk state is greater than that for the abnormal echo state. This approach ensures near-end speech continuity in the dual-talk state while ensuring clean echo suppression, and effectively identifies and suppresses abnormal echoes caused by sudden path changes.

[0019] In a second aspect, the present application provides a device for processing voice echo, the processing device comprising: a processing unit for playing a first voice signal of a first user sent by a remote device through a loudspeaker; the processing unit is also used to collect a voice signal through a microphone to obtain a second voice signal, the second voice signal including a third voice signal of the second user and an echo signal corresponding to the first voice signal; the processing unit is also used to eliminate the echo signal in the second voice signal based on the first voice signal to obtain a residual signal; the processing unit is also used to perform sub-band division processing on the residual signal to obtain a first signal on one or more sub-bands; a determination unit is used to determine a sound field state of the first signal on one or more sub-bands; the processing unit is also used to process the first signal on one or more sub-bands based on the sound field state to obtain a third voice signal; and a sending unit is used to send the third voice signal to the remote device.

[0020] In a third aspect, the present application provides a chip comprising a processor and a communication interface, wherein the processor is configured to enable the chip to execute the method in the above-mentioned first aspect or any possible implementation thereof.

[0021] In a fourth aspect, the present application provides a module device, which includes a communication module, a power module, a storage module and a chip, wherein: the power module is used to provide power to the module device; the storage module is used to store data and instructions; the communication module is used to conduct internal communication within the module device, or for the module device to communicate with external devices; the chip is used to execute the method in the above-mentioned first aspect or any possible implementation thereof.

[0022] In a fifth aspect, an embodiment of the present invention discloses a device for processing voice echoes, which includes a memory and a processor, the memory being used to store a computer program, the computer program including program instructions, and the processor being configured to call the program instructions to execute the method of the first aspect or any possible implementation thereof.

[0023] In a sixth aspect, the present application provides a computer-readable storage medium storing computer-readable instructions. When the computer-readable instructions are executed on a speech echo processing device, the speech echo processing device executes the method of the first aspect or any possible implementation thereof.

[0024] In a seventh aspect, the present application provides a computer program or computer program product, comprising codes or instructions, which, when executed on a computer, enable the computer to execute the method in the first aspect or any possible implementation thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0026] Figure 1 This is a system architecture diagram of a communication system provided by an embodiment of the present application;

[0027] Figure 2 is a schematic diagram of an acoustic echo in voice communication provided by an embodiment of the present application;

[0028] Figure 3 This is a flow chart of a method for processing voice echo provided in an embodiment of the present application;

[0029] Figure 4 It is a spectrogram provided in an embodiment of the present application;

[0030] Figure 5 1 is a flow chart of another method for processing voice echo provided in an embodiment of the present application;

[0031] Figure 6 This is a structural diagram of a speech echo processing device provided in an embodiment of the present application;

[0032] Figure 7 1 is a schematic structural diagram of another apparatus for processing speech echoes provided in an embodiment of the present application;

[0033] Figure 8 It is a structural diagram of a module device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0034] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this invention.

[0035] The terms used in the following examples of the present application are for the purpose of describing specific embodiments only and are not intended to limit the present application. As used in the specification and appended claims of the present application, the singular expressions "a," "an," "said," "above," "the," and "this" are intended to include plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in the present application refers to and encompasses any or all possible combinations of one or more of the listed items.

[0036] It should be noted that the terms "first", "second", "third", etc. in the specification and claims of the present application and in the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the term "comprising" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products, or devices.

[0037] To better understand the embodiments of the present application, the following first introduces the system architecture involved in the embodiments of the present application:

[0038] The technical solutions of the embodiments of the present application can be applied to various communication systems, such as: global system of mobile communication (GSM) system, code division multiple access (CDMA) system, wideband code division multiple access (WCDMA) system, general packet radio service (GPRS), long term evolution (LTE) system, LTE frequency division duplex (FDD) system, LTE time division duplex (TDD), universal mobile telecommunication system (UMTS), world-wide interoperability for microwave access (WiMAX) communication system, fifth generation (5G) system or new radio (NR) and future communication systems.

[0039] Figure 1 This is a system architecture diagram of a communication system provided by an embodiment of the present application. The solution in this application is applicable to the communication system. The communication system may include at least one near-end device and at least one far-end device. Figure 1 Taking a communication system including one near-end device and one far-end device as an example, the communication system may also include a network device (such as a base station).

[0040] Among them, the proximal device and the distal device can be a terminal device, and the terminal device includes a device that provides voice and / or data connectivity to the user. For example, the terminal device is a device with wireless transceiver function, which can be deployed on land, including indoors or outdoors, handheld, wearable or vehicle-mounted; it can also be deployed on the water surface (such as a ship, etc.); it can also be deployed in the air (such as an airplane, a balloon, and a satellite, etc.). The terminal device in the embodiment of the present application can be a device equipped with a microphone and a speaker, such as a mobile phone, a headset, a tablet computer (Pad), a computer with wireless transceiver function, a virtual reality (VR) terminal device, an augmented reality (AR) terminal device, a wireless terminal in industrial control, a vehicle-mounted terminal device, a wireless terminal in self-driving, a wireless terminal in remote medical care, a wireless terminal in a smart grid, a wireless terminal in transportation safety, a wireless terminal in a smart city, a wireless terminal in a smart home, a wearable terminal device, etc. The embodiment of the present application does not limit the application scenario. In addition, the terminal device may sometimes also be referred to as a terminal, user equipment (UE), access terminal device, vehicle-mounted terminal, industrial control terminal, UE unit, UE station, mobile station, mobile station, remote station, remote terminal device, mobile device, UE terminal device, terminal device, wireless communication device, UE agent or UE device, etc. The terminal may also be fixed or mobile. In the embodiments of the present application, the device for realizing the function of the near-end device or the far-end device may be a terminal device, or may be a device that can support the terminal device to realize the function, such as a chip system or a combination device or component that can realize the function of the terminal device, and the device may be installed in the terminal device.

[0041] To facilitate understanding of the solution provided in the embodiments of the present application, the following describes acoustic echo in voice communication:

[0042] like Figure 2 As shown, the far-end device collects the voice signal of the first user through microphone 1, and transmits the voice signal of the first user to the near-end device through network transmission. After the near-end device receives the voice signal of the first user, it can play the voice signal of the first user through speaker 2. At this time, the voice signal of the first user will be picked up directly or indirectly by the microphone 2 of the near-end device after reflection, thereby forming an acoustic echo signal. If the acoustic echo signal is not processed, the near-end device transmits the acoustic echo signal to the far-end device through network transmission, and the first user will always be able to hear his own voice when speaking, which seriously affects the call experience.

[0043] In addition, if the speaker 2 of the near-end device is playing the voice signal of the first user while the second user is also speaking (i.e., in a dual-talk state), the microphone 2 of the near-end device will also pick up the voice signal of the second user. The near-end device will then transmit the second user's voice signal and the acoustic echo signal together to the far-end device through the network, causing the first user to not only be able to hear his own voice when speaking, but also be unable to hear the second user's voice clearly, which seriously reduces the effectiveness of the voice call and the quality of voice communication.

[0044] In order to improve the quality of voice communication and enhance the voice call experience, the present application provides a method, device, chip, and module device for processing voice echo. The following further describes in detail the method, device, chip, and module device for processing voice echo provided in the embodiments of the present application.

[0045] Figure 3 FIG. 1 is a flow chart of a method for processing voice echo provided in an embodiment of the present application. Figure 3 As shown, the method for processing voice echo includes the following steps S301 to S307. Figure 3 The method shown can be performed by a proximal device (such as Figure 1 the near-end device in the ). Alternatively, Figure 3 The execution subject of the method shown may be a chip in a proximal device, which is not limited here. Figure 3 The method is described by taking the proximal device as the execution subject as an example.

[0046] S301: A local device plays a first voice signal of a first user sent by a remote device through a loudspeaker.

[0047] In an embodiment of the present application, the near-end device and the far-end device can use direct communication to transmit data, that is, the far-end device directly sends the first sound signal of the first user collected by the microphone to the near-end device; the near-end device and the far-end device can also use indirect communication to transmit data, for example, the far-end device sends the first sound signal of the first user collected by the microphone to the network device, and the network device then sends the first sound signal of the first user to the near-end device. Accordingly, after the near-end device receives the first sound signal of the first user, it can play it through the speaker. The embodiment of the present application takes the near-end device and the far-end device using direct communication to transmit data as an example.

[0048] S302: The near-end device collects a sound signal through a microphone to obtain a second sound signal.

[0049] In this embodiment of the present application, the second sound signal includes the second user's third sound signal and an echo signal corresponding to the first sound signal. The near-end device can capture the second user's third sound signal through a microphone; at the same time, it can also capture the first user's first sound signal played through a speaker, which can be considered the echo signal corresponding to the first sound signal. Optionally, the second sound signal may also include a noise signal.

[0050] It should be noted that the echo signal corresponding to the first sound signal can be directly collected by the microphone after the first sound signal of the first user is played, or it can be indirectly collected by the microphone after the first sound signal of the first user is played and reflected, which is not limited here.

[0051] S303: The near-end device cancels the echo signal in the second sound signal based on the first sound signal to obtain a residual signal.

[0052] In an embodiment of the present application, the residual signal includes the second user's third sound signal and a residual echo signal, etc. For example, the residual signal may be a signal obtained after the near-end device performs echo cancellation processing using an adaptive filter. It should be noted that the purpose of the near-end device performing echo cancellation processing is to retain the second user's third sound signal in the second sound signal and to eliminate the echo signal in the second sound signal as much as possible. However, it is not necessarily possible to completely eliminate the echo. In this case, the residual signal includes not only the second user's third sound signal but also the residual echo signal, etc. Therefore, the near-end device will need to further process the residual signal.

[0053] In one possible implementation, when the near-end device eliminates the echo signal in the second sound signal based on the first sound signal to obtain a residual signal, the specific implementation method may be: filtering the first sound signal based on an adaptive filter to obtain an estimated echo signal; and eliminating the echo signal in the second sound signal based on the estimated echo signal to obtain a residual signal. The adaptive filter (AF) may be a normalized least mean square adaptive filter (NLMS), a recursive least squares adaptive filter (RLS), a frequency domain adaptive filter (FDAF), etc., without limitation herein. The adaptive filter is primarily used to eliminate linear echoes in the first sound signal. Here, linear echoes refer to echo signals that are linearly correlated with the first sound signal. Based on this approach, linear echo signals can be effectively eliminated. It should be noted that the adaptive filter here only needs to achieve a preset robustness and be able to eliminate linear echoes greater than a preset threshold, and its specific type is not limited.

[0054] Specifically, the near-end device can input the first sound signal into an adaptive filter for filtering processing and output an estimated echo signal; subtract the estimated echo signal from the second sound signal (including the third sound signal and the echo signal corresponding to the first sound signal) to obtain a residual signal.

[0055] S304: The near-end device performs sub-band division processing on the residual signal to obtain first signals on one or more sub-bands.

[0056] In the embodiment of the present application, the so-called sub-band refers to a sub-frequency band. The near-end device can perform sub-band division processing on the residual signal in the frequency domain. For example, a fast Fourier transform (FFT) can be performed on the residual signal to convert the residual signal from the time domain to the frequency domain, and then divide it into one or more sub-bands. Each sub-band corresponds to a signal frequency point (that is, the Fourier coefficient after the FFT of the time-domain signal). Therefore, each sub-band will correspond to a first signal, that is, the sub-band division processing of the residual signal is realized, and the first signal on one or more sub-bands is obtained. In other words, the near-end device can divide the residual signal into the first signal on one or more sub-bands, and can subsequently process the first signal on each sub-band separately to improve the accuracy and effectiveness of the echo processing.

[0057] Of course, the near-end device can also perform sub-band processing on the residual signal in the time domain. For example, the residual signal can be output to a corresponding bandpass filter for bandpass filtering, thereby obtaining a first signal on one or more sub-bands. The embodiment of the present application takes the near-end device performing sub-band processing on the residual signal in the frequency domain as an example.

[0058] S305: The near-end device determines the sound field state of the first signal on the one or more sub-bands.

[0059] In the embodiments of the present application, the sound field state can be a double-talk state, an abnormal echo state, a far-end single-talk state, a silent state, a near-end single-talk state, etc. The double-talk state refers to a state where both parties are speaking simultaneously; the abnormal echo state refers to a sound state caused by a sudden change in the echo path, such as a sudden occlusion of the microphone, which causes the echo to gather and amplify; the far-end single-talk state refers to a state where only the far-end sound signal is present and no near-end sound signal is present, i.e., a pure echo state; the silent state refers to a state where neither party is speaking; the near-end single-talk state refers to a state where only the near-end sound signal is present and no far-end sound signal is present, i.e., a no-echo state. The silent state and the near-end single-talk state do not have any echo, so no subsequent echo suppression is required.

[0060] In one possible implementation, the sound field state includes a double-talk state, an abnormal echo state, and a far-end single-talk state. When the near-end device determines the sound field state of the first signal on the one or more sub-bands, a specific implementation method may include the following steps s11 to s15. Based on this method, the sound field state of the first signal on each sub-band can be effectively determined.

[0061] s11. If a speech signal exists in the second sound signal and the first sound signal, the near-end device calculates a leakage coefficient based on the residual signal and the estimated echo signal.

[0062] In a specific implementation, if there is a voice signal in both the second sound signal and the first sound signal, proving the presence of an echo, the silent state and the near-end single-talk state can be excluded. Furthermore, the near-end device can distinguish the abnormal echo state by calculating the leakage coefficient. The leakage coefficient is used to indicate the degree of signal leakage echo and is an intermediate variable that describes the similarity between the two frame signals. The larger the leakage coefficient, the more echo components the signal leaks. The calculation is performed in the full band of the time domain or frequency domain and is completed by correlation calculation or quasi-correlation calculation. For example, the leakage coefficient can be calculated using formula (1), using the conjugate multiplication of the frequency domain signal instead of the cross-correlation of the time domain signal to reduce the amount of calculation. Formula (1) is as follows:

[0063]

[0064] Where n represents the signal frame number, k represents the frequency point number, η(n) represents the leakage coefficient, and P ey (n), P ee (n) and P yy (n) are all the results of multiplying the complex spectrum conjugate pairs and summing the real parts, as follows:

[0065]

[0066]

[0067]

[0068] Among them, R ey (n,k), R ee (n,k) and R yy (n,k) is represented as follows:

[0069] R ey (n,k)=real(E(n,k)*Y(n,k) * ) (5)

[0070] R ee (n,k)=real(E(n,k)*E(n,k) * ) (6)

[0071] R yy (n,k)=real(Y(n,k)*Y(n,k) * ) (7)

[0072] Where E(n,k) represents the spectrum of the residual signal, Y(n,k) represents the spectrum of the estimated echo signal, (·) * Indicates conjugation.

[0073] It should be noted that, if the calculation of the leakage coefficient is not performed in the frequency domain, it can also be completed through the cross-correlation of time domain signals.

[0074] s12. If the leakage coefficient is greater than the first threshold, the near-end device determines that the sound field state of all first signals in the one or more sub-bands is an abnormal echo state.

[0075] In a specific implementation, when the leakage coefficient is greater than the first threshold, it indicates that the residual signal and the estimated echo signal are more similar, that is, the residual signal leaks echo to a greater extent, and it is more likely that a sudden change in the echo path causes a large amount of echo to leak without being eliminated. Therefore, the sound field state of all first signals in the first signals on one or more sub-bands can be determined as an abnormal echo state.

[0076] For example, assuming the first threshold is 0.5, the residual signal is divided into the first signal on subband 1, the first signal on subband 2, and the first signal on subband 3. If the leakage coefficient is 0.6 (i.e., greater than the first threshold), the sound field state of the first signal on subband 1, the sound field state of the first signal on subband 2, and the sound field state of the first signal on subband 3 are all abnormal echo states.

[0077] s13. If the leakage coefficient is less than or equal to the first threshold, the near-end device determines whether the echo suppression ratio of the first signal on the first sub-band is greater than a second threshold.

[0078] In a specific implementation, the first subband is any one of one or more subbands, and the echo suppression ratio indicates the degree of suppression of echo signals in the sound signal. When the leakage coefficient is less than or equal to a first threshold, it indicates that the residual signal and the estimated echo signal are less similar, that is, the residual signal has less echo leakage, most of the echo has been eliminated, and the sound field state of the first signal in the first subband is more likely to be a double talk state or a far-end single talk state. Therefore, the near-end device needs to further determine whether the sound field state of the first signal in the first subband is a double talk state or a far-end single talk state based on the echo suppression ratio.

[0079] s14. If the echo suppression ratio of the first signal on the first sub-band is greater than the second threshold, the near-end device determines that the sound field state of the first signal on the first sub-band is a far-end single-talk state.

[0080] s15. If the echo suppression ratio of the first signal on the first sub-band is less than or equal to the second threshold, the near-end device determines that the sound field state of the first signal on the first sub-band is a double-talk state.

[0081] In a specific implementation, when the echo suppression ratio of the first signal on the first sub-band is greater than the second threshold, it indicates that most echo signals have been eliminated by the adaptive filter, that is, there are relatively more echo signals linearly correlated with the first sound signal, then it can be determined that the sound field state of the first signal on the first sub-band is a far-end single-talk state, that is, a pure echo; when the echo suppression ratio of the first signal on the first sub-band is less than or equal to the second threshold, it indicates that most signals have not been eliminated by the adaptive filter, that is, there are relatively few echo signals linearly correlated with the first sound signal, then it can be determined that the sound field state of the first signal on the first sub-band is a double-talk state.

[0082] For example, assuming the first threshold is 0.5 and the second threshold is 30dB, the residual signal is divided into the first signal on subband 1, the first signal on subband 2, and the first signal on subband 3. If the leakage coefficient is 0.4 (i.e., less than the first threshold), then it is further determined whether the echo suppression ratio of the first signal on each subband is greater than 30dB. Among them, the echo suppression ratio of the first signal on subband 1 is 28dB, the echo suppression ratio of the first signal on subband 2 is 29dB, and the echo suppression ratio of the first signal on subband 3 is 32dB. Therefore, the sound field state of the first signal on subband 1 is a dual-talk state, the sound field state of the first signal on subband 2 is a dual-talk state, and the sound field state of the first signal on subband 3 is a far-end single-talk state.

[0083] Optionally, if a speech signal is present in the residual signal, the second threshold is the third threshold; if no speech signal is present in the residual signal, the second threshold is the fourth threshold, and the third threshold is greater than the fourth threshold. That is, when a speech signal is present in the residual signal, it indicates that the second user is speaking. Furthermore, step s11 has already determined that speech signals are present in both the second and first sound signals, indicating that the first user is also speaking. Therefore, when a speech signal is present in the residual signal, the sound field state of the first signal in the current subband is more likely to be a double-talk state. Conversely, when no speech signal is present in the residual signal, the sound field state of the first signal in the current subband is more likely to be a far-end single-talk state. Therefore, when a speech signal is present in the residual signal, the second threshold is set to a larger value, making it more likely that the sound field state of the first signal in the current subband is a double-talk state. When no speech signal is present in the residual signal, the second threshold is set to a smaller value, making it more likely that the sound field state of the first signal in the current subband is a far-end single-talk state. This approach can improve the accuracy of sound field state determination.

[0084] For example, assuming the third threshold is 35 dB and the fourth threshold is 28 dB, if a speech signal exists in the residual signal, the second threshold is 35 dB; if no speech signal exists in the residual signal, the second threshold is 28 dB.

[0085] In one possible implementation, the sound field state also includes a silent state, and the method further includes: if no speech signal is present in the second sound signal, determining by the near-end device that the sound field state of all first signals in the one or more sub-bands is a silent state. In other words, the microphone of the near-end device does not capture any sound, i.e., is in a silent state, and therefore, no subsequent echo suppression is required.

[0086] In one possible implementation, the sound field state also includes a near-end single-talk state. The method further includes: if a voice signal is present in the second sound signal and no voice signal is present in the first sound signal, the near-end device determining that the sound field state of all first signals in the one or more sub-bands is a near-end single-talk state. In other words, the microphone of the near-end device only captures the third sound signal of the second user and does not capture an echo signal corresponding to the first sound signal of the first user, i.e., a near-end single-talk state, and therefore, no subsequent echo suppression is required.

[0087] It should be noted that the near-end device may use voice activity detection (VAD) to determine whether a voice signal exists in the second sound signal, the first sound signal, or the residual signal. Specifically, energy-based VAD, correlation-based VAD, etc. may be used, which are not limited here.

[0088] S306: The near-end device processes the first signal on one or more sub-bands based on the sound field state to obtain a third sound signal.

[0089] In an embodiment of the present application, the near-end device can further perform residual echo suppression, nonlinear echo suppression, and merging processing on the first signal on one or more sub-bands. Specifically, according to different sound field states, different methods are used to process the first signal on one or more sub-bands, which can more effectively suppress the echo. For example, when the sound field state of the first signal on the sub-band is a double-talk state, the echo suppression coefficient is a; when the sound field state of the first signal on the sub-band is an abnormal echo state, the echo suppression coefficient is b; when the sound field state of the first signal on the sub-band is a far-end single-talk state, the echo suppression coefficient is c, and then the first signal on one or more sub-bands is processed using the echo suppression coefficient corresponding to the sound field state.

[0090] It should be noted that the near-end device may process the first signal on one or more sub-bands based on the sound field state in the time domain or in the frequency domain, which is not limited here.

[0091] In one possible implementation, the method further includes: the near-end device adjusting an update rate of the adaptive filter based on the sound field state. In other words, the update rate of the adaptive filter needs to be adjusted based on the sound field state to improve the accuracy of linear echo suppression. The specific strategy for adjusting the update rate of the adaptive filter based on the sound field state is not limited; for example, the update rate of the adaptive filter may be adjusted based on the sound field state of the first signal in the first subband of the current frame.

[0092] For example, if the sound field state of the first signal on the first sub-band of the current frame is a double-talk state, the update rate of the current adaptive filter can be reduced and applied to the processing of the next frame signal; if the sound field state of the first signal on the first sub-band of the current frame is an abnormal echo state, the update rate of the current adaptive filter can be increased and applied to the processing of the next frame signal; if the sound field state of the first signal on the first sub-band of the current frame is a far-end single-talk state, the update rate of the current adaptive filter does not need to be changed and is applied to the processing of the next frame signal.

[0093] S307: The local device sends a third sound signal to the remote device. Correspondingly, the remote device may receive the third sound signal.

[0094] In an embodiment of the present application, the near-end device can directly send the third sound signal obtained after echo suppression to the far-end device; alternatively, the near-end device can send the third sound signal obtained after echo suppression to the network device, which then sends it to the far-end device, without limitation. The far-end device can play the received third sound signal through a speaker, effectively preventing the first user from hearing his or her own voice during the voice call.

[0095] See Figure 4 , Figure 4 This is a spectrogram provided in the embodiment of the present application. Figure 4 As shown, the horizontal axis is time and the vertical axis is frequency. Figure 4 (a) shows the processing result of the sound signal in the silent field state, with a sampling rate of 16kHz. (a) includes the far-end single-talk state (far-end simplex, i.e., pure echo) and the double-talk state (duplex). Figure 4 Figure (b) shows the sound signal processed by this method. Compared to the case without sound field state judgment, this method can retain more near-end speech in the dual-talk state, especially the mid- and high-frequency parts, while ensuring that the linear and nonlinear echo components of the pure echo segment are completely suppressed. This method can preserve more details and clear sounds, improving the listening experience.

[0096] It can be seen that based on Figure 3 The described method divides the residual signal into subbands and then accurately determines the sound field state based on the first signal in the subband. Different echo suppression processing methods are used for different sound field states to more effectively suppress echoes, thereby improving voice communication quality and enhancing the voice call experience.

[0097] Figure 5 FIG. 1 is a flow chart of another method for processing voice echo provided in an embodiment of the present application. Figure 5 As shown, the method for processing voice echo includes the following steps S501 to S509. Figure 5The method shown can be performed by a proximal device (such as Figure 1 the near-end device in the ). Alternatively, Figure 5 The execution subject of the method shown may be a chip in a proximal device, which is not limited here. Figure 5 The method is described below by taking the proximal device as the execution subject as an example, wherein steps S506 to S508 are a specific implementation of step S306.

[0098] S501: A local device plays a first sound signal of a first user sent by a remote device through a loudspeaker.

[0099] S502: The near-end device collects a sound signal through a microphone to obtain a second sound signal.

[0100] S503: The near-end device cancels the echo signal in the second sound signal based on the first sound signal to obtain a residual signal.

[0101] S504: The near-end device performs sub-band division processing on the residual signal to obtain first signals on one or more sub-bands.

[0102] S505: The near-end device determines the sound field state of the first signal on the one or more sub-bands.

[0103] The specific implementation of steps S501 to S505 may refer to the specific implementation of steps S301 to S305 above, and will not be described in detail here.

[0104] S506: The near-end device performs residual echo suppression on the first signal on one or more sub-bands based on the sound field state to obtain a second signal on one or more sub-bands.

[0105] In the embodiment of the present application, the near-end device processes the first signal on one or more sub-bands based on the sound field state in the frequency domain as an example.

[0106] In one possible implementation, when the near-end device performs residual echo suppression on the first signal on one or more sub-bands based on the sound field state to obtain the second signal on one or more sub-bands, the specific implementation may be one of the following two methods:

[0107] Method 1: specifically includes the following steps s21 to s24:

[0108] s21. The near-end device determines an initial residual echo estimate corresponding to the first signal on one or more subbands based on the leakage coefficient and the estimated echo signal, or determines an initial residual echo estimate corresponding to the first signal on one or more subbands based on the leakage coefficient and the first sound signal.

[0109] For example, the calculation may be performed using formula (8), that is, determining the initial residual echo estimate corresponding to the first signal on one or more sub-bands based on the leakage coefficient and the estimated echo signal:

[0110] RES ini (n,k)=η(n)*Y(n,k) (8)

[0111] Where n represents the signal frame number, k represents the frequency point number, η(n) represents the leakage coefficient, Y(n,k) represents the estimated echo signal, RES ini (n, k) represents the initial residual echo estimate corresponding to the first signal on the subband.

[0112] For another example, formula (9) may be used for calculation, that is, initial residual echo estimates corresponding to the first signal on one or more sub-bands may be determined based on the leakage coefficient and the first sound signal:

[0113] RES ini (n,k)=η(n)*X(n,k) (9)

[0114] Where n represents the signal frame number, k represents the frequency point number, η(n) represents the leakage coefficient, X(n,k) represents the first sound signal, RES ini (n, k) represents the initial residual echo estimate corresponding to the first signal on the subband.

[0115] Optionally, when the near-end device determines the initial residual echo estimate corresponding to the first signal on one or more subbands based on the leakage coefficient and the estimated echo signal, the specific implementation method may be: determining the fourth sound signal based on the estimated echo signal, the first sound signal, and the second sound signal; and determining the initial residual echo estimate corresponding to the first signal on one or more subbands based on the leakage coefficient and the fourth sound signal. The fourth sound signal may be determined by taking the maximum value of the estimated echo signal, the first sound signal, and the second sound signal; or by linear weighting; or by other methods, which are not limited here. Based on this method, it is more conducive to improving the accuracy of the residual echo estimate corresponding to the first signal on each subband.

[0116] For example, among the average value of the estimated echo signal, the average value of the first sound signal, and the average value of the second sound signal, the largest is the average value of the second sound signal, so the fourth sound signal is the second sound signal. Further, the initial residual echo estimate corresponding to the first signal on the subband can be calculated using formula (10):

[0117] RES ini (n,k)=η(n)*Z(n,k) (10)

[0118] Where n represents the signal frame number, k represents the frequency point number, η(n) represents the leakage coefficient, Z(n,k) represents the fourth sound signal, RES ini (n, k) represents the initial residual echo estimate corresponding to the first signal on the subband.

[0119] s22. The near-end device determines a residual echo gain corresponding to the first signal on one or more sub-bands based on the sound field state.

[0120] In the specific implementation, different sound field states correspond to different residual echo gains: No echo suppression is required in the silent state and the near-end single-talk state, so the residual echo gains for both the silent state and the near-end single-talk state are 0. In the abnormal echo state, residual echo estimation needs to be strengthened to ensure communication quality. In the far-end single-talk state, normal residual echo estimation is sufficient. In the double-talk state, residual echo estimation needs to be weakened to ensure effective communication between the two parties. Specifically, the residual echo gain for the abnormal echo state is greater than that for the far-end single-talk state, and the residual echo gain for the far-end single-talk state is greater than that for the double-talk state. Based on this approach, adjusting the residual echo gain according to the sound field state helps improve the accuracy of residual echo estimation.

[0121] For example, assume that the residual signal is divided into a first signal on subband 1, a first signal on subband 2, and a first signal on subband 3. The sound field state of the first signal on subband 1 is a double talk state, the sound field state of the first signal on subband 2 is a double talk state, and the sound field state of the first signal on subband 3 is an abnormal echo state. As shown in Table 1 below, the residual echo gain corresponding to the double talk state is 0.8, the residual echo gain corresponding to the far-end single talk state is 1, the residual echo gain corresponding to the abnormal echo state is 1.5, the residual echo gain corresponding to the silence state is 0, and the residual echo gain corresponding to the near-end single talk state is 0. Therefore, the residual echo gain corresponding to the first signal on subband 1 is 0.8, the residual echo gain corresponding to the first signal on subband 2 is 0.8, and the residual echo gain corresponding to the first signal on subband 3 is 1.5.

[0122] Table 1

[0123] Sound field status Residual echo gain Dual talk status 0.8 Remote single talk status 1 Abnormal echo state 1.5 Silent state 0 Near-end single talk status 0

[0124] s23. The near-end device determines a residual echo estimate corresponding to the first signal on one or more subbands based on the residual echo gain and the initial residual echo estimate.

[0125] For example, the residual echo estimate corresponding to the first signal on one or more sub-bands may be calculated using formula (11), which is as follows:

[0126] RES(n,k)=RESini (n,k)*G(l),k∈sub(l) (11)

[0127] Where n represents the signal frame number, k represents the frequency number, l represents the subband number, sub(l) represents the corresponding subband frequency range, G(l) represents the residual echo gain, RES ini (n, k) represents the initial residual echo estimate, and RES(n, k) represents the residual echo estimate corresponding to the first signal on the subband.

[0128] s24. The near-end device performs residual echo suppression on the first signal on one or more sub-bands based on the residual echo estimation to obtain a second signal on one or more sub-bands.

[0129] Optionally, when the near-end device performs residual echo suppression on the first signal on one or more sub-bands based on the residual echo estimation to obtain the second signal on one or more sub-bands, a specific implementation method may be one of the following two methods:

[0130] Method a: The near-end device determines a first suppression coefficient corresponding to a first signal in one or more subbands based on a residual echo estimate. The near-end device then performs residual echo suppression on the first signal in one or more subbands based on the first suppression coefficient to obtain a second signal in one or more subbands. This method directly determines the first suppression coefficient corresponding to the first signal in one or more subbands based on the magnitude of the residual echo estimate, which helps improve echo suppression efficiency.

[0131] For example, assume that the first signal on subband 1 corresponds to residual echo estimate 1, the first signal on subband 2 corresponds to residual echo estimate 2, and the first signal on subband 3 corresponds to residual echo estimate 3. The residual echo estimates can be divided into three ranges: the residual echo estimates in range 1 correspond to a first suppression coefficient of 0.8, the residual echo estimates in range 2 correspond to a first suppression coefficient of 0.5, and the residual echo estimates in range 3 correspond to a first suppression coefficient of 0.3. Residual echo estimate 1 is within range 1, residual echo estimate 2 is within range 1, and residual echo estimate 3 is within range 3. Therefore, the first suppression coefficient corresponding to the first signal on subband 1 is 0.8, the first suppression coefficient corresponding to the first signal on subband 2 is 0.8, and the first suppression coefficient corresponding to the first signal on subband 3 is 0.3.

[0132] Method b: The near-end device determines a noise estimate corresponding to a first signal on one or more subbands. Based on the noise estimate and the residual echo estimate, the near-end device determines a fused audio signal on the one or more subbands. Based on the fused audio signal, the near-end device performs residual echo suppression on the first signal on the one or more subbands to obtain a second signal on the one or more subbands. This method incorporates the residual echo into the noise estimate for fused noise reduction, which further improves voice communication quality.

[0133] For example, the near-end device may use minimum tracking to determine the noise estimate corresponding to the first signal on the sub-band, and then use formula (12) or formula (13) to calculate the fused sound signal on one or more sub-bands. Formula (12) and formula (13) are as follows:

[0134] S mix (n,k)=S min (n,k)+abs(RES(n,k)) (12)

[0135] Among them, n represents the signal frame number, k represents the frequency point number, S min (n, k) represents the noise estimate corresponding to the first signal on the subband, RES(n, k) represents the residual echo estimate corresponding to the first signal on the subband, S mix (n, k) represents the fused sound signal on the sub-band, and abs(·) represents the absolute value.

[0136] S mix (n,k)=max(S min (n,k),abs(RES(n,k))) (13)

[0137] Among them, n represents the signal frame number, k represents the frequency point number, S min (n, k) represents the noise estimate corresponding to the first signal on the subband, RES(n, k) represents the residual echo estimate corresponding to the first signal on the subband, S mix (n,k) represents the fused sound signal on the sub-band, abs(·) represents the absolute value, and max(·) represents the maximum value.

[0138] Then, the noise and residual echo are removed together by spectral subtraction, thereby achieving residual echo suppression. Specifically, formula (14) can be used for calculation, which is as follows:

[0139] OUT1(n,k)=(abs(E(n,k))-abs(S mix (n,k)))*phase(E(n,k)) (14)

[0140] Among them, n represents the signal frame number, k represents the frequency point number, S mix(n, k) represents the fused sound signal on the subband, E(n, k) represents the first signal on the subband, abs(·) represents taking the absolute value, phase(·) represents taking the phase, and OUT1(n, k) represents the second signal on one or more subbands.

[0141] Based on the first approach, the residual echo gain corresponding to the first signal on each sub-band is determined using the sound field state, thereby improving the accuracy of the residual echo estimation corresponding to the first signal on each sub-band and thereby improving the effectiveness of the residual echo suppression.

[0142] Method 2: The near-end device determines a second suppression coefficient corresponding to the first signal in one or more sub-bands based on the sound field state. The near-end device then performs residual echo suppression on the first signal in one or more sub-bands based on the second suppression coefficient to obtain the second signal in one or more sub-bands. In other words, the degree of residual echo suppression needs to be adjusted based on the sound field state to improve the accuracy of residual echo suppression.

[0143] For example, the sound field state of the first signal on subband 1 is double talk, the sound field state of the first signal on subband 2 is double talk, and the sound field state of the first signal on subband 3 is abnormal echo. As shown in Table 2 below, the second suppression coefficient corresponding to the double talk state is 0.8, the second suppression coefficient corresponding to the far-end single talk state is 0.5, and the second suppression coefficient corresponding to the abnormal echo state is 0.3. Therefore, the second suppression coefficients corresponding to the first signal on subband 1 are 0.8, the second suppression coefficients corresponding to the first signal on subband 2 are 0.8, and the second suppression coefficients corresponding to the first signal on subband 3 are 0.3.

[0144] Table 2

[0145] Sound field status The second suppression coefficient Dual talk status 0.8 Remote single talk status 0.5 Abnormal echo state 0.3

[0146] Then, by multiplying the first signal on subband 1 with the corresponding second suppression coefficient, the second signal on subband 1 can be obtained; by multiplying the first signal on subband 2 with the corresponding second suppression coefficient, the second signal on subband 2 can be obtained; and by multiplying the first signal on subband 3 with the corresponding second suppression coefficient, the second signal on subband 3 can be obtained.

[0147] Optionally, the second suppression coefficient corresponding to the dual-talk state is greater than the second suppression coefficient corresponding to the far-end single-talk state, and the second suppression coefficient corresponding to the far-end single-talk state is greater than the second suppression coefficient corresponding to the abnormal echo state. In other words, the degree of residual echo suppression in the dual-talk state is less than that in the far-end single-talk state, and the degree of residual echo suppression in the far-end single-talk state is less than that in the abnormal echo state. This approach ensures near-end speech continuity in the dual-talk state while ensuring clean echo suppression, and effectively identifies and suppresses abnormal echoes caused by sudden path changes.

[0148] S507: The near-end device performs nonlinear echo suppression on the second signal on one or more sub-bands based on the sound field state to obtain a third signal on one or more sub-bands.

[0149] In one possible implementation, when the near-end device performs nonlinear echo suppression on the second signal in one or more sub-bands based on the sound field state to obtain the third signal in one or more sub-bands, the specific implementation method may be: determining a third suppression coefficient corresponding to the second signal in one or more sub-bands based on the sound field state; and performing nonlinear echo suppression on the second signal in one or more sub-bands based on the third suppression coefficient to obtain the third signal in one or more sub-bands. In other words, the degree of nonlinear echo suppression needs to be adjusted according to the sound field state to improve the accuracy of nonlinear echo suppression.

[0150] For example, the sound field state of the first signal on subband 1 is double talk, the sound field state of the first signal on subband 2 is double talk, and the sound field state of the first signal on subband 3 is abnormal echo. As shown in Table 3 below, the third suppression coefficient corresponding to the double talk state is 0.8, the third suppression coefficient corresponding to the far-end single talk state is 0.5, and the third suppression coefficient corresponding to the abnormal echo state is 0.3. Therefore, the third suppression coefficient corresponding to the second signal on subband 1 is 0.8, the third suppression coefficient corresponding to the second signal on subband 2 is 0.8, and the third suppression coefficient corresponding to the second signal on subband 3 is 0.3.

[0151] Table 3

[0152] Sound field status The third suppression coefficient Dual talk status 0.8 Remote single talk status 0.5 Abnormal echo state 0.3

[0153] Then, by multiplying the second signal on subband 1 with the corresponding third suppression coefficient, the third signal on subband 1 can be obtained; by multiplying the second signal on subband 2 with the corresponding third suppression coefficient, the third signal on subband 2 can be obtained; and by multiplying the second signal on subband 3 with the corresponding third suppression coefficient, the third signal on subband 3 can be obtained.

[0154] Optionally, the third suppression coefficient corresponding to the dual-talk state is greater than the third suppression coefficient corresponding to the far-end single-talk state, and the third suppression coefficient corresponding to the far-end single-talk state is greater than the third suppression coefficient corresponding to the abnormal echo state. In other words, the degree of nonlinear echo suppression in the dual-talk state is less than that in the far-end single-talk state, and the degree of nonlinear echo suppression in the far-end single-talk state is less than that in the abnormal echo state. This approach ensures the continuity of near-end speech in the dual-talk state while ensuring clean echo suppression, and effectively identifies and suppresses abnormal echoes caused by sudden path changes.

[0155] S508: The near-end device combines the third signals on one or more sub-bands to obtain a third sound signal.

[0156] In a specific implementation, the near-end device performs an inverse fast Fourier transform (IFFT) and overlap-add operations on the third signal on one or more subbands to obtain a third sound signal, that is, the second sound signal completes echo suppression to obtain an effective third sound signal.

[0157] S509: The local device sends a third sound signal to the remote device. Correspondingly, the remote device may receive the third sound signal.

[0158] The specific implementation of step S509 may refer to the specific implementation of step S307 above, which will not be described in detail here.

[0159] It can be seen that based on Figure 5 The described method divides the residual signal after adaptive filtering into subbands and accurately judges the sound field state based on the first signal in the subband. At the same time, different residual echo suppression, nonlinear echo suppression and other processing methods are used for different sound field states to more effectively suppress echoes, which is conducive to improving voice communication quality and enhancing the voice call experience.

[0160] See Figure 6 , Figure 6 1 is a structural diagram of a voice echo processing device provided by an embodiment of the present invention. The voice echo processing device may be a near-end device or a device having near-end device functions (eg, a chip). Figure 6 As shown, the speech echo processing device 600 may include a processing unit 601 , a determining unit 602 , and a sending unit 603 .

[0161] in:

[0162] The processing unit 601 is configured to play a first voice signal of a first user sent by a remote device through a speaker;

[0163] The processing unit 601 is further configured to collect a sound signal through a microphone to obtain a second sound signal, where the second sound signal includes a third sound signal of the second user and an echo signal corresponding to the first sound signal;

[0164] The processing unit 601 is further configured to eliminate the echo signal in the second sound signal based on the first sound signal to obtain a residual signal;

[0165] The processing unit 601 is further configured to perform sub-band division processing on the residual signal to obtain a first signal on one or more sub-bands;

[0166] A determining unit 602 is configured to determine a sound field state of a first signal on one or more sub-bands;

[0167] The processing unit 601 is further configured to process the first signal on one or more sub-bands based on the sound field state to obtain a third sound signal;

[0168] The sending unit 603 is configured to send the third sound signal to a remote device.

[0169] In one possible implementation, when the processing unit 601 eliminates the echo signal in the second sound signal based on the first sound signal to obtain a residual signal, it can be specifically used to: filter the first sound signal based on an adaptive filter to obtain an estimated echo signal; and eliminate the echo signal in the second sound signal based on the estimated echo signal to obtain a residual signal.

[0170] In a possible implementation, the processing unit 601 is further configured to adjust an update rate of the adaptive filter based on the sound field state.

[0171] In one possible implementation, the sound field state includes a double-talk state, an abnormal echo state, and a far-end single-talk state. The determining unit 602, when determining the sound field state of the first signal on the one or more sub-bands, may be specifically configured to: if a speech signal exists in the second sound signal and the first sound signal, calculate a leakage coefficient based on the residual signal and the estimated echo signal, the leakage coefficient being used to indicate a degree of signal leakage echo; and if the leakage coefficient is greater than a first threshold, determine that the sound field state of all first signals in the one or more sub-bands is the abnormal echo state; If the leakage coefficient is less than or equal to the first threshold, it is determined whether the echo suppression ratio of the first signal on the first sub-band is greater than the second threshold, where the first sub-band is any one of the one or more sub-bands, and the echo suppression ratio is used to indicate the degree of suppression of the echo signal in the sound signal; if the echo suppression ratio of the first signal on the first sub-band is greater than the second threshold, it is determined that the sound field state of the first signal on the first sub-band is a far-end single-talk state; if the echo suppression ratio of the first signal on the first sub-band is less than or equal to the second threshold, it is determined that the sound field state of the first signal on the first sub-band is a double-talk state.

[0172] In a possible implementation, if a speech signal exists in the residual signal, the second threshold is the third threshold; if no speech signal exists in the residual signal, the second threshold is the fourth threshold, and the third threshold is greater than the fourth threshold.

[0173] In one possible implementation, the processing unit 601, when processing the first signal on one or more sub-bands based on the sound field state to obtain the third sound signal, can be specifically used to: perform residual echo suppression on the first signal on one or more sub-bands based on the sound field state to obtain the second signal on one or more sub-bands; perform nonlinear echo suppression on the second signal on one or more sub-bands based on the sound field state to obtain the third signal on one or more sub-bands; and perform merging processing on the third signal on one or more sub-bands to obtain the third sound signal.

[0174] In one possible implementation, the processing unit 601, when performing residual echo suppression on the first signal on one or more subbands based on the sound field state to obtain the second signal on one or more subbands, can be specifically used to: determine an initial residual echo estimate corresponding to the first signal on the one or more subbands based on the leakage coefficient and the estimated echo signal, or determine an initial residual echo estimate corresponding to the first signal on one or more subbands based on the leakage coefficient and the first sound signal; determine a residual echo gain corresponding to the first signal on one or more subbands based on the sound field state; determine a residual echo estimate corresponding to the first signal on one or more subbands based on the residual echo gain and the initial residual echo estimate; and perform residual echo suppression on the first signal on one or more subbands based on the residual echo estimate to obtain the second signal on one or more subbands.

[0175] In one possible implementation, the processing unit 601, when determining an initial residual echo estimate corresponding to a first signal on one or more subbands based on the leakage coefficient and the estimated echo signal, can be specifically used to: determine a fourth sound signal based on the estimated echo signal, the first sound signal, and the second sound signal; and determine an initial residual echo estimate corresponding to the first signal on one or more subbands based on the leakage coefficient and the fourth sound signal.

[0176] In one possible implementation, when performing residual echo suppression on a first signal on one or more subbands based on the residual echo estimate to obtain a second signal on one or more subbands, the processing unit 601 may be specifically configured to: determine a first suppression coefficient corresponding to the first signal on the one or more subbands based on the residual echo estimate; and perform residual echo suppression on the first signal on the one or more subbands based on the first suppression coefficient to obtain the second signal on the one or more subbands.

[0177] In one possible implementation, when performing residual echo suppression on the first signal on one or more subbands based on the residual echo estimate to obtain the second signal on the one or more subbands, the processing unit 601 may be specifically configured to: determine a noise estimate corresponding to the first signal on the one or more subbands; determine a fused sound signal on the one or more subbands based on the noise estimate and the residual echo estimate; and perform residual echo suppression on the first signal on the one or more subbands based on the fused sound signal to obtain the second signal on the one or more subbands.

[0178] In one possible implementation, the processing unit 601, when performing residual echo suppression on the first signal on one or more sub-bands based on the sound field state to obtain the second signal on one or more sub-bands, can be specifically used to: determine the second suppression coefficient corresponding to the first signal on one or more sub-bands based on the sound field state; and perform residual echo suppression on the first signal on one or more sub-bands based on the second suppression coefficient to obtain the second signal on one or more sub-bands.

[0179] In a possible implementation, the second suppression coefficient corresponding to the double-talk state is greater than the second suppression coefficient corresponding to the far-end single-talk state, and the second suppression coefficient corresponding to the far-end single-talk state is greater than the second suppression coefficient corresponding to the abnormal echo state.

[0180] In one possible implementation, the processing unit 601, when performing nonlinear echo suppression on the second signal on one or more sub-bands based on the sound field state to obtain the third signal on one or more sub-bands, can be specifically used to: determine the third suppression coefficient corresponding to the second signal on one or more sub-bands based on the sound field state; perform nonlinear echo suppression on the second signal on one or more sub-bands based on the third suppression coefficient to obtain the third signal on one or more sub-bands.

[0181] In a possible implementation, the third suppression coefficient corresponding to the double-talk state is greater than the third suppression coefficient corresponding to the far-end single-talk state, and the third suppression coefficient corresponding to the far-end single-talk state is greater than the third suppression coefficient corresponding to the abnormal echo state.

[0182] The present application also provides a chip that can execute the relevant steps of the terminal device in the aforementioned method embodiment. The chip includes a processor and a communication interface, and the processor is configured to enable the chip to perform the following operations: playing a first sound signal of a first user sent by a remote device through a speaker; collecting a sound signal through a microphone to obtain a second sound signal, the second sound signal including a third sound signal of the second user and an echo signal corresponding to the first sound signal; eliminating the echo signal in the second sound signal based on the first sound signal to obtain a residual signal; performing sub-band division processing on the residual signal to obtain a first signal on one or more sub-bands; determining the sound field state of the first signal on the one or more sub-bands; processing the first signal on the one or more sub-bands based on the sound field state to obtain a third sound signal; and sending the third sound signal to the remote device.

[0183] In one possible implementation, when the chip eliminates the echo signal in the second sound signal based on the first sound signal to obtain a residual signal, it can be specifically used to: filter the first sound signal based on an adaptive filter to obtain an estimated echo signal; and eliminate the echo signal in the second sound signal based on the estimated echo signal to obtain a residual signal.

[0184] In a possible implementation, the chip is further configured to adjust an update rate of the adaptive filter based on the sound field state.

[0185] In a possible implementation, the sound field state includes a double-talk state, an abnormal echo state, and a far-end single-talk state. The chip, when determining the sound field state of the first signal on the one or more sub-bands, can be specifically used to: if there is a speech signal in the second sound signal and the first sound signal, calculate a leakage coefficient based on the residual signal and the estimated echo signal, and the leakage coefficient is used to indicate the degree of signal leakage echo; if the leakage coefficient is greater than a first threshold, determine that the sound field state of all first signals in the first signals on the one or more sub-bands is the abnormal echo state; if the If the leakage coefficient is less than or equal to a first threshold, it is determined whether an echo suppression ratio of the first signal on a first sub-band is greater than a second threshold, where the first sub-band is any one of the one or more sub-bands, and the echo suppression ratio is used to indicate a degree of suppression of an echo signal in a sound signal; if the echo suppression ratio of the first signal on the first sub-band is greater than the second threshold, it is determined that the sound field state of the first signal on the first sub-band is a far-end single-talk state; if the echo suppression ratio of the first signal on the first sub-band is less than or equal to the second threshold, it is determined that the sound field state of the first signal on the first sub-band is a double-talk state.

[0186] In a possible implementation, if a speech signal exists in the residual signal, the second threshold is the third threshold; if no speech signal exists in the residual signal, the second threshold is the fourth threshold, and the third threshold is greater than the fourth threshold.

[0187] In one possible implementation, when the chip processes the first signal on one or more sub-bands based on the sound field state to obtain a third sound signal, it can be specifically used to: perform residual echo suppression on the first signal on one or more sub-bands based on the sound field state to obtain a second signal on one or more sub-bands; perform nonlinear echo suppression on the second signal on one or more sub-bands based on the sound field state to obtain a third signal on one or more sub-bands; and perform merging processing on the third signals on one or more sub-bands to obtain a third sound signal.

[0188] In one possible implementation, the chip, when performing residual echo suppression on a first signal on one or more sub-bands based on the sound field state to obtain a second signal on one or more sub-bands, can be specifically used to: determine an initial residual echo estimate corresponding to the first signal on the one or more sub-bands based on the leakage coefficient and the estimated echo signal, or determine an initial residual echo estimate corresponding to the first signal on one or more sub-bands based on the leakage coefficient and the first sound signal; determine a residual echo gain corresponding to the first signal on one or more sub-bands based on the sound field state; determine a residual echo estimate corresponding to the first signal on one or more sub-bands based on the residual echo gain and the initial residual echo estimate; and perform residual echo suppression on the first signal on one or more sub-bands based on the residual echo estimate to obtain a second signal on one or more sub-bands.

[0189] In one possible implementation, the chip, when determining an initial residual echo estimate corresponding to a first signal on one or more subbands based on the leakage coefficient and the estimated echo signal, can be specifically used to: determine a fourth sound signal based on the estimated echo signal, the first sound signal, and the second sound signal; and determine an initial residual echo estimate corresponding to the first signal on one or more subbands based on the leakage coefficient and the fourth sound signal.

[0190] In one possible implementation, when performing residual echo suppression on a first signal on one or more subbands based on the residual echo estimate to obtain a second signal on one or more subbands, the chip may be specifically configured to: determine a first suppression coefficient corresponding to the first signal on the one or more subbands based on the residual echo estimate; and perform residual echo suppression on the first signal on the one or more subbands based on the first suppression coefficient to obtain the second signal on the one or more subbands.

[0191] In one possible implementation, when performing residual echo suppression on a first signal on one or more subbands based on the residual echo estimate to obtain a second signal on one or more subbands, the chip may be specifically configured to: determine a noise estimate corresponding to the first signal on the one or more subbands; determine a fused sound signal on the one or more subbands based on the noise estimate and the residual echo estimate; and perform residual echo suppression on the first signal on the one or more subbands based on the fused sound signal to obtain the second signal on the one or more subbands.

[0192] In one possible implementation, when performing residual echo suppression on a first signal on one or more sub-bands based on the sound field state to obtain a second signal on one or more sub-bands, the chip can be specifically used to: determine a second suppression coefficient corresponding to the first signal on one or more sub-bands based on the sound field state; and perform residual echo suppression on the first signal on one or more sub-bands based on the second suppression coefficient to obtain the second signal on one or more sub-bands.

[0193] In a possible implementation, the second suppression coefficient corresponding to the double-talk state is greater than the second suppression coefficient corresponding to the far-end single-talk state, and the second suppression coefficient corresponding to the far-end single-talk state is greater than the second suppression coefficient corresponding to the abnormal echo state.

[0194] In one possible implementation, when the chip performs nonlinear echo suppression on the second signal on one or more sub-bands based on the sound field state to obtain a third signal on one or more sub-bands, the chip can be specifically used to: determine a third suppression coefficient corresponding to the second signal on one or more sub-bands based on the sound field state; and perform nonlinear echo suppression on the second signal on one or more sub-bands based on the third suppression coefficient to obtain the third signal on one or more sub-bands.

[0195] In a possible implementation, the third suppression coefficient corresponding to the double-talk state is greater than the third suppression coefficient corresponding to the far-end single-talk state, and the third suppression coefficient corresponding to the far-end single-talk state is greater than the third suppression coefficient corresponding to the abnormal echo state.

[0196] For each device or product applied to or integrated in the chip, each module contained therein can be implemented in the form of hardware such as circuits, or at least some of the modules can be implemented in the form of software programs, which run on the processor integrated inside the chip, and the remaining (if any) modules can be implemented in the form of hardware such as circuits.

[0197] See also Figure 7 , Figure 7 1 is a schematic diagram of the structure of a voice echo processing device provided in an embodiment of the present invention. The voice echo processing device 700 may include a memory 701 and a processor 702. Optionally, it may also include a communication interface 703. The memory 701, processor 702, and communication interface 703 are connected via one or more communication buses. Communication interface 703 is controlled by processor 702 to send and receive information.

[0198] The memory 701 may include a read-only memory and a random access memory, and provides instructions and data to the processor 702. A portion of the memory 701 may also include a nonvolatile random access memory.

[0199] The communication interface 703 is used to receive or send data.

[0200] The processor 702 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. A general-purpose processor may be a microprocessor, or alternatively, the processor 702 may be any conventional processor. Specifically:

[0201] The memory 701 is used to store program instructions.

[0202] The processor 702 is configured to call the program instructions stored in the memory 701 .

[0203] The processor 702 calls the program instructions stored in the memory 701 to enable the speech echo processing apparatus 700 to execute the method executed by the near-end device in the above method embodiment.

[0204] like Figure 8 As shown, Figure 8 800 is a schematic diagram of a module device provided in an embodiment of the present application. The module device 800 can execute the relevant steps of the proximal device in the aforementioned method embodiment, and the module device 800 includes: a communication module 801, a power module 802, a storage module 803 and a chip 804.

[0205] Among them, the power module 802 is used to provide power to the module device; the storage module 803 is used to store data and instructions; the communication module 801 is used for internal communication within the module device, or for the module device to communicate with external devices; the chip 804 is used to execute the method executed by the proximal device in the above method embodiment.

[0206] It should be noted that Figures 6 to 8 The details not mentioned in the corresponding embodiments and the specific implementation methods of each step can be found in Figure 3 and Figure 5 The illustrated embodiments and the aforementioned contents will not be described in detail here.

[0207] An embodiment of the present application also provides a computer-readable storage medium, which stores instructions. When the computer-readable storage medium is executed on a processor, the method flow of the above method embodiment is implemented.

[0208] An embodiment of the present application further provides a computer program product. When the computer program product is run on a processor, the method flow of the above method embodiment is implemented.

[0209] Regarding the various modules / units contained in the various devices and products described in the above embodiments, they can be software modules / units, hardware modules / units, or partly software modules / units and partly hardware modules / units. For example, for various devices and products applied to or integrated into a chip, the various modules / units contained therein can all be implemented in the form of hardware such as circuits, or at least some of the modules / units can be implemented in the form of software programs, which run on the integrated processor inside the chip, and the remaining (if any) modules / units can be implemented in the form of hardware such as circuits; for various devices and products applied to or integrated into a chip module, the various modules / units contained therein can all be implemented in the form of hardware such as circuits, and different modules / units can be located in the same part of the chip module (such as a chip, circuit module, etc.) or in different components, or at least some of the modules / units can be implemented in the form of software programs. It can be implemented in the form of a software program, which runs on the processor integrated inside the chip module, and the remaining (if any) modules / units can be implemented in the form of hardware such as circuits; for various devices and products applied to or integrated in the terminal, the modules / units contained therein can all be implemented in the form of hardware such as circuits, and different modules / units can be located in the same component (for example, chip, circuit module, etc.) or different components in the terminal, or, at least some modules / units can be implemented in the form of a software program, which runs on the processor integrated inside the terminal, and the remaining (if any) modules / units can be implemented in the form of hardware such as circuits.

[0210] It should be noted that for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, certain operations can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in this specification are all preferred embodiments, and the actions and modules involved are not necessarily required by this application.

[0211] The descriptions of the various embodiments provided in this application can refer to each other. The descriptions of each embodiment have their own focus. For parts not described in detail in a particular embodiment, please refer to the relevant descriptions of other embodiments. For the convenience and brevity of description, for example, the functions and operations performed by the various devices and equipment provided in the embodiments of this application can refer to the relevant descriptions of the method embodiments of this application. The various method embodiments and the various device embodiments can also refer to, be combined with, or quote each other.

[0212] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some or all of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for processing speech echo, characterized in that: The method comprises: Playing a first sound signal of a first user sent by a remote device through a loudspeaker; Collecting a sound signal through a microphone to obtain a second sound signal, where the second sound signal includes a third sound signal of a second user and an echo signal corresponding to the first sound signal; Performing filtering processing on the first sound signal based on an adaptive filter to obtain an estimated echo signal; Eliminate the echo signal in the second sound signal based on the estimated echo signal to obtain a residual signal; Performing sub-band division processing on the residual signal to obtain a first signal on one or more sub-bands; If there is a speech signal in the second sound signal and the first sound signal, a leakage coefficient is calculated based on the residual signal and the estimated echo signal, wherein the leakage coefficient is used to indicate the degree of signal leakage echo; If the leakage coefficient is greater than a first threshold, determining that the sound field state of all first signals in the one or more sub-bands is an abnormal echo state; If the leakage coefficient is less than or equal to the first threshold, and the echo suppression ratio of the first signal on the first sub-band is greater than the second threshold, it is determined that the sound field state of the first signal on the first sub-band is a far-end single talk state; the first sub-band is any one of the one or more sub-bands, and the echo suppression ratio is used to indicate the degree of suppressing the echo signal in the sound signal; If the leakage coefficient is less than or equal to the first threshold, and the echo suppression ratio of the first signal on the first sub-band is less than or equal to the second threshold, determining that the sound field state of the first signal on the first sub-band is a double-talk state; Processing the first signal on the one or more sub-bands based on the sound field state to obtain the third sound signal; The third sound signal is sent to the remote device.

2. The method according to claim 1, characterized in that The method further comprises: An update rate of the adaptive filter is adjusted based on the sound field state.

3. The method according to claim 1, characterized in that If there is a speech signal in the residual signal, the second threshold is the third threshold; if there is no speech signal in the residual signal, the second threshold is the fourth threshold, and the third threshold is greater than the fourth threshold.

4. The method according to claim 1, characterized in that The processing of the first signal on the one or more sub-bands based on the sound field state to obtain the third sound signal includes: Performing residual echo suppression on the first signal on the one or more sub-bands based on the sound field state to obtain a second signal on the one or more sub-bands; Performing nonlinear echo suppression on the second signal on the one or more sub-bands based on the sound field state to obtain a third signal on the one or more sub-bands; The third signal on the one or more sub-bands is combined to obtain the third sound signal.

5. The method according to claim 4, characterized in that The performing residual echo suppression on the first signal on the one or more sub-bands based on the sound field state to obtain a second signal on the one or more sub-bands includes: Determine an initial residual echo estimate corresponding to the first signal on the one or more sub-bands based on the leakage coefficient and the estimated echo signal, or determine an initial residual echo estimate corresponding to the first signal on the one or more sub-bands based on the leakage coefficient and the first sound signal; Determine a residual echo gain corresponding to the first signal on the one or more sub-bands based on the sound field state; Determine a residual echo estimate corresponding to the first signal on the one or more subbands based on the residual echo gain and the initial residual echo estimate; Residual echo suppression is performed on the first signal on the one or more sub-bands based on the residual echo estimation to obtain a second signal on the one or more sub-bands.

6. The method according to claim 5, characterized in that The determining, based on the leakage coefficient and the estimated echo signal, an initial residual echo estimate corresponding to the first signal on the one or more sub-bands comprises: determining a fourth sound signal based on the estimated echo signal, the first sound signal, and the second sound signal; An initial residual echo estimate corresponding to the first signal on the one or more subbands is determined based on the leakage coefficient and the fourth sound signal.

7. The method according to claim 5, characterized in that The performing residual echo suppression on the first signal on the one or more sub-bands based on the residual echo estimation to obtain a second signal on the one or more sub-bands includes: Determine a first suppression coefficient corresponding to a first signal on the one or more subbands based on the residual echo estimation; Residual echo suppression is performed on the first signal on the one or more sub-bands based on the first suppression coefficient to obtain a second signal on the one or more sub-bands.

8. The method according to claim 5, characterized in that The performing residual echo suppression on the first signal on the one or more sub-bands based on the residual echo estimation to obtain a second signal on the one or more sub-bands includes: Determining a noise estimate corresponding to the first signal on the one or more subbands; determining a fused sound signal on the one or more subbands based on the noise estimate and the residual echo estimate; Residual echo suppression is performed on the first signal on the one or more sub-bands based on the fused sound signal to obtain a second signal on the one or more sub-bands.

9. The method according to claim 4, characterized in that The performing residual echo suppression on the first signal on the one or more sub-bands based on the sound field state to obtain a second signal on the one or more sub-bands includes: Determine a second suppression coefficient corresponding to the first signal on the one or more sub-bands based on the sound field state; Residual echo suppression is performed on the first signal on the one or more sub-bands based on the second suppression coefficient to obtain second signals on the one or more sub-bands.

10. The method according to claim 9, characterized in that The second suppression coefficient corresponding to the double-talk state is greater than the second suppression coefficient corresponding to the far-end single-talk state, and the second suppression coefficient corresponding to the far-end single-talk state is greater than the second suppression coefficient corresponding to the abnormal echo state.

11. The method according to claim 4, characterized in that The performing nonlinear echo suppression on the second signal on the one or more sub-bands based on the sound field state to obtain a third signal on the one or more sub-bands includes: Determine a third suppression coefficient corresponding to the second signal on the one or more sub-bands based on the sound field state; Nonlinear echo suppression is performed on the second signal on the one or more sub-bands based on the third suppression coefficient to obtain a third signal on the one or more sub-bands.

12. The method according to claim 11, characterized in that The third suppression coefficient corresponding to the double-talk state is greater than the third suppression coefficient corresponding to the far-end single-talk state, and the third suppression coefficient corresponding to the far-end single-talk state is greater than the third suppression coefficient corresponding to the abnormal echo state.

13. A speech echo processing device, characterized in that: The processing device comprises: A processing unit, configured to play a first sound signal of a first user sent by a remote device through a loudspeaker; The processing unit is further configured to collect a sound signal through a microphone to obtain a second sound signal, where the second sound signal includes a third sound signal of a second user and an echo signal corresponding to the first sound signal; The processing unit is further configured to filter the first sound signal based on an adaptive filter to obtain an estimated echo signal; The processing unit is further configured to perform a process of eliminating the echo signal in the second sound signal based on the estimated echo signal to obtain a residual signal; The processing unit is further configured to perform sub-band division processing on the residual signal to obtain a first signal on one or more sub-bands; a determining unit, configured to calculate a leakage coefficient based on the residual signal and the estimated echo signal if a speech signal exists in the second sound signal and the first sound signal, wherein the leakage coefficient is used to indicate a degree of signal leakage echo; The determining unit is further configured to determine that the sound field state of all first signals in the one or more sub-bands is an abnormal echo state if the leakage coefficient is greater than a first threshold; The determining unit is further configured to determine that the sound field state of the first signal on the first sub-band is a far-end single talk state if the leakage coefficient is less than or equal to the first threshold and the echo suppression ratio of the first signal on the first sub-band is greater than a second threshold; the first sub-band is any one of the one or more sub-bands, and the echo suppression ratio is used to indicate a degree of suppressing an echo signal in a sound signal; The determining unit is further configured to determine that the sound field state of the first signal on the first sub-band is a double-talk state if the leakage coefficient is less than or equal to the first threshold and the echo suppression ratio of the first signal on the first sub-band is less than or equal to the second threshold; The processing unit is further configured to process the first signal on the one or more sub-bands based on the sound field state to obtain the third sound signal; A sending unit is used to send the third sound signal to the remote device.

14. A chip, characterized in that: The chip comprises a processor and a communication interface, wherein the processor is configured to cause the chip to execute the method according to any one of claims 1 to 12.

15. A module device, characterized in that: The module device includes a communication module, a power module, a storage module and a chip, wherein: The power module is used to provide electrical energy to the module device; The storage module is used to store data and instructions; The communication module is used for internal communication of the module device, or for communication between the module device and an external device; The chip is used to execute the method according to any one of claims 1 to 12.

16. A speech echo processing device, characterized in that: The device comprises a memory and a processor, wherein the memory is used to store a computer program, the computer program comprises program instructions, and the processor is configured to call the program instructions to enable the speech echo processing device to execute the method according to any one of claims 1 to 12.

17. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed on a speech echo processing device, the speech echo processing device executes the method according to any one of claims 1 to 12.

Images