CN115223582B - Audio noise processing method, system, electronic device and medium - Google Patents

Abstract

The present invention provides an audio noise processing method and system. The method includes: obtaining the audio signal at the microphone end and determining the source of the audio signal; if the source of the audio signal is a call audio signal, the call audio signal is processed through a DSP noise reduction chip. Perform noise reduction processing, and input the obtained first noise reduction signal to the call module through the audio hardware abstraction layer; if the source of the audio signal is a local voice audio signal, then input it to the voice assistant application software through the audio hardware abstraction layer, and pass The voice noise reduction algorithm integrated in the voice assistant application software performs noise reduction processing and echo cancellation processing on the local voice audio signal; if the source of the audio signal is a mobile phone interconnection voice audio signal, the noise reduction unit of the hardware abstraction layer will perform noise reduction processing on the mobile phone interconnection. The speech audio signal undergoes noise reduction processing. The invention reduces the hardware cost and development cost of noise processing, and avoids quality risks caused by substantial modification of the audio hardware abstraction layer.

Description

Audio noise processing method, system, electronic device and medium

Technical field

The present invention relates to the technical field of noise processing, and in particular to an audio noise processing method, system, electronic device and medium.

Background technique

With the development of the Internet of Vehicles, voice, Bluetooth calls, etc. have become the mainstream and important interaction methods in the car. In the vehicle environment, the car contains environmental noise, wind noise, tire noise, played audio and other noises, resulting in vehicle voice The input signal of Bluetooth and other devices is poor, resulting in low quality of calls and recognition. In-car noise reduction application scenarios are divided into three categories: one is for front-end audio processing for speech recognition, and the other is for Bluetooth calls and CarPlay (a vehicle-mounted system released by Apple in the United States that integrates the user's iOS device with the dashboard system ), front-end audio processing for call functions such as telematics BOX, TBOX, etc. The third is used for mobile phone interconnection voice audio processing such as CarPlay and Carlife (a mobile phone and car intelligent interconnection product launched by Baidu). Currently, the industry mainly uses one of two processing solutions: hard noise reduction or soft noise reduction.

Existing vehicle noise reduction methods mainly use one of two processing solutions: hard noise reduction processing or soft noise reduction processing. Hard noise reduction refers to noise reduction and echo elimination through independent hardware noise reduction modules. However, the disadvantages of this method are: 1) high hardware cost and difficulty in hardware design for recording channels; 2) printed circuit board (Printed circuit board). CircuitBoard (PCB) takes up a lot of space; 3) Refreshing needs to be combined with firmware updates, and the noise reduction parameters/algorithms cannot be flexibly adjusted. It cannot be updated through Over-the-Air Technology (OTA). It is difficult to refresh during after-sales updates. big. Soft noise reduction refers to noise reduction through software algorithms. However, the disadvantages of this method are: 1) The cost is high. Since all algorithms are integrated in the system on chip (SOC), the performance of the CPU is greatly consumed. , therefore it is necessary to use a high-performance SOC chip, which increases the cost; if there is a delay when the CPU performance is insufficient, the noise reduction effect is not good; 2) Since all soft noise reduction algorithms are deployed in the audio hardware abstraction layer of the Android system ( Audio Hardware Abstraction Layer (Audio-HAL), leading to an increase in quality risks caused by modifying the HAL layer; 3) Due to the complex verification process of the soft noise reduction algorithm, production complexity is increased and production efficiency is reduced.

Contents of the invention

In order to solve the above existing technical problems, the present invention provides an audio noise processing method, system, electronic device and medium, which reduces vehicle audio noise interference by providing noise reduction processing methods suitable for audio signals from different sound sources. Improve the quality of in-vehicle audio and reduce the hardware cost and development cost of noise processing.

A first aspect of the present invention provides an audio noise processing method, including:

Obtain the audio signal from the microphone end and determine the source of the audio signal;

If the source of the audio signal is a call audio signal, then the DSP noise reduction chip performs noise reduction processing on the call audio signal, and the first noise reduction signal obtained by the noise reduction processing is input to the call module through the audio hardware abstraction layer. , to complete two-way call interaction and response;

If the source of the audio signal is a local voice audio signal, it is input to the voice assistant application software through the audio hardware abstraction layer, and the local voice audio signal is processed through the voice noise reduction algorithm integrated in the voice assistant application software. Noise reduction processing and echo cancellation processing;

If the source of the audio signal is a mobile phone Internet voice audio signal, the noise reduction process is performed on the mobile phone Internet voice audio signal through the noise reduction unit of the audio hardware abstraction layer.

Further, the noise reduction processing of the call audio signal through the DSP noise reduction chip includes:

Gain the call audio signal through a DSP noise reduction chip to obtain a first gain signal;

Perform noise separation and filtering on the first gain signal to obtain a monophonic audio signal;

The mono audio signal is resampled so that the sampling rate of the mono audio signal meets the sampling rate of the input signal of the call module, and the resampled mono audio signal is used as the first downscaled signal. Noisy signal.

Further, the mono audio signal is resampled, specifically:

The monophonic audio signal is resampled by hardware through the built-in ADSP chip in the SOC chip.

Further, before the voice audio signal is input to the voice assistant application software through the audio hardware abstraction layer, and the voice audio signal is subjected to noise reduction processing and echo cancellation processing through the voice noise reduction algorithm integrated in the voice assistant application software, include:

Obtain the audio signal at the speaker end as a reference signal;

Mix the reference signal and the speech audio signal to obtain a mixed audio signal;

The mixed audio signal is input into the CPU, and the voice noise reduction algorithm integrated in the voice assistant application software is called.

Further, before performing noise reduction processing on the call audio signal through the DSP noise reduction chip, it also includes:

Set the audio source program to call mode, and switch the DSP noise reduction chip to call channel mode through the SPI serial communication protocol.

Further, before performing noise reduction processing and echo cancellation processing on the local voice audio signal through the voice noise reduction algorithm integrated in the voice assistant application software, it also includes:

Set the audio source program to intelligent voice mode, and switch the DSP noise reduction chip to intelligent voice channel mode.

Further, the sound source of the audio signal is determined, specifically:

The sound source of the audio signal is determined through the sound source circulation channel of the audio signal.

A second aspect of the present invention provides an audio noise processing system, including:

Audio signal acquisition and sound source judgment module, the sound source obtains the audio signal from the microphone end, and determines the sound source of the audio signal;

A call audio signal processing module, used to perform noise reduction processing on the call audio signal through a DSP noise reduction chip when the source of the audio signal is a call audio signal, and pass the first noise reduction signal obtained by the noise reduction processing through The audio hardware abstraction layer is input to the call module to complete two-way call interaction and response;

A local voice audio signal processing module, used when the source of the audio signal is a local voice audio signal, input to the voice assistant application software through the audio hardware abstraction layer, and use the voice noise reduction integrated in the voice assistant application software The algorithm performs noise reduction processing and echo cancellation processing on the local voice audio signal;

The mobile phone Internet voice audio signal processing module is used to perform noise reduction processing on the mobile phone Internet voice audio signal through the noise reduction unit of the audio hardware abstraction layer when the source of the audio signal is the mobile phone Internet voice audio signal.

A third aspect of the present invention provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the above-mentioned steps are implemented. The audio noise processing method according to any one of the first aspects.

A fourth aspect of the present invention provides a computer-readable storage medium. The computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the above-mentioned first step. The audio noise processing method described in any one of the aspects.

Compared with the prior art, the beneficial effects of the embodiments of the present invention are:

The present invention provides an audio noise processing method and system, wherein the method includes: obtaining the audio signal at the microphone end, and determining the source of the audio signal; if the source of the audio signal is a call audio signal, then through the DSP noise reduction chip Perform noise reduction processing on the call audio signal, and input the first noise reduction signal obtained by the noise reduction processing to the call module through the audio hardware abstraction layer to complete two-way call interaction and response; if the source of the audio signal is local The voice audio signal is input to the voice assistant application software through the audio hardware abstraction layer, and the local voice audio signal is subjected to noise reduction processing and echo cancellation processing through the voice noise reduction algorithm integrated in the voice assistant application software; if If the source of the audio signal is a mobile phone Internet voice audio signal, the noise reduction process is performed on the mobile phone Internet voice audio signal through the noise reduction unit of the audio hardware abstraction layer. The present invention separates the audio signal according to its source, adopts different noise reduction methods for audio signals from different sources, reduces the hardware cost and development cost of noise processing, and avoids quality risks caused by substantial modifications to the audio hardware abstraction layer. .

Description of the drawings

In order to explain the technical solution of the present invention more clearly, the drawings needed to be used in the implementation will be briefly introduced below. Obviously, the drawings in the following description are only some implementations of the present invention. For ordinary people in the art, For technical personnel, other drawings can also be obtained based on these drawings without exerting creative work.

Figure 1 is a flow chart of an audio noise processing method provided by Embodiment 1 of the present invention;

Figure 2 is a flow chart of an audio noise processing method provided by Embodiment 2 of the present invention;

Figure 3 is a flow chart of an audio noise processing method provided by Embodiment 3 of the present invention;

Figure 4 is a device diagram of an audio noise processing system provided by Embodiment 5 of the present invention;

Figure 5 is a structural diagram of an electronic device provided by an embodiment of the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts fall within the scope of protection of the present invention.

It should be understood that the step numbers used in the text are only for convenience of description and are not intended to limit the execution order of the steps.

It should be understood that the terminology used in the description of the present invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in this specification and the appended claims, the singular forms "a", "an" and "the" are intended to include the plural forms unless the context clearly dictates otherwise.

The terms "comprises" and "comprising" indicate the presence of described features, integers, steps, operations, elements and/or components but do not exclude the presence of one or more other features, integers, steps, operations, elements, components and/or The existence or addition to its collection.

The term "and/or" refers to any and all possible combinations of one or more of the associated listed items and includes such combinations.

Currently, the industry mainly adopts one of two processing solutions: hard noise reduction processing or soft noise reduction: Hard noise reduction is used alone. Due to high requirements for hardware noise reduction modules such as voice calls, an independent noise reduction module needs to be specially used. The noise reduction module reduces noise echo, but this solution seriously increases the hardware cost, takes up a lot of space on the printed circuit board (PCB), and needs to update the firmware every time, making it difficult to flexibly adjust the noise reduction parameters/algorithm. Especially after-sales updates are difficult to refresh and cannot be updated through OTA.

Soft noise reduction is used alone. Since the algorithms required for soft noise reduction are integrated in the SOC, if soft noise reduction is used for call voice, it consumes a lot of CPU performance compared to hard noise reduction. If the performance is insufficient, there will be a delay problem. , reducing the call effect; and the call soft noise reduction algorithm needs to be deployed in the Audio-HAL of the Android system, resulting in an increase in quality risks caused by modifying the HAL layer, and the multi-channel soft noise reduction algorithm verification steps can easily affect the complexity of the production line, reducing Slow production pace.

Compared with the industry's method of using either hard noise reduction or soft noise reduction, the present invention proposes an audio noise processing method that combines hard noise reduction and soft noise reduction. The audio signal is separated according to the source of the audio signal, and the audio signal is separated. Audio signals from different sources use different noise reduction methods to reduce vehicle audio noise interference, improve vehicle audio quality, and reduce noise processing hardware costs and development costs.

first.

Please refer to Figure 1. Embodiment 1 of the present invention provides an audio noise processing method, including steps S100, S200, S300 and S400; it should be noted that after the execution of step S100 is completed, according to the result of S100, select S200 or S300 or One step in S400 is executed; and the relationship between steps S200, S300 and S400 is parallel rather than sequential. specifically:

S100. Obtain the audio signal from the microphone end and determine the source of the audio signal.

It should be noted that the audio signal at the microphone end is the user's input signal to the microphone end. After the DSP chip collects the audio signal at the microphone end, the vehicle system determines the source of the audio signal.

It can be understood that since different audio signals have different audio channels, determining the sound source of the audio signal may be: judging the sound source of the audio signal through the sound source circulation channel of the audio signal. Since different audio signals have different frequency domains, determining the sound source of the audio signal may also include determining the sound source of the audio signal based on the frequency domain of the audio signal.

Preferably, the sound source of the audio signal is determined, specifically:

The sound source of the audio signal is determined through the sound source circulation channel of the audio signal.

S200. If the source of the audio signal is a call audio signal, perform noise reduction processing on the call audio signal through the DSP noise reduction chip, and input the first noise reduction signal obtained by the noise reduction processing to the audio hardware abstraction layer. Call module to complete two-way call interaction and response.

It should be noted that the call audio signals include: Bluetooth calls, CarPlay calls, ECALL (car emergency call system), BCAll (one-button phone rescue, mainly road rescue) and other call audio signals. The noise reduction processing of the call audio signal through the DSP noise reduction chip is the hard noise reduction processing of the call audio signal.

It can be understood that after the DSP chip collects the audio signal from the microphone and the vehicle system determines that it is a call audio signal, the DSP chip directly performs noise reduction processing inside the chip to obtain the first noise reduction signal, and the first noise reduction signal is passed through the audio signal The hardware abstraction layer is transmitted to the call module to complete two-way call interaction and response.

S300. If the source of the audio signal is a local voice audio signal, it is input to the voice assistant application software through the audio hardware abstraction layer, and the local voice audio is processed through the voice noise reduction algorithm integrated in the voice assistant application software. The signal is processed for noise reduction and echo cancellation.

It should be noted that local voice audio signals include: vehicle-mounted artificial intelligence voice, vehicle-mounted voice recognition assistant and other local voice audio signals. Performing noise reduction processing and echo cancellation processing on the local voice audio signal through the voice noise reduction algorithm integrated in the voice assistant application software is to perform soft noise reduction on the local voice audio signal.

It can be understood that after the DSP chip collects the audio signal at the microphone end and the vehicle system determines that it is a local voice audio signal, the local voice audio signal is transmitted to the CPU and input to the voice assistant application software through the audio hardware abstraction layer, and through The voice noise reduction algorithm integrated in the voice assistant application software performs noise reduction processing and echo cancellation processing on the local voice audio signal.

S400. If the source of the audio signal is a mobile phone Internet voice audio signal, perform noise reduction processing on the mobile phone Internet voice audio signal through the noise reduction unit of the audio hardware abstraction layer.

It should be noted that the mobile phone Internet voice audio is a mix of voice and music. The mobile phone Internet voice audio signals include: mobile phone Internet voice audio signals from third-party applications such as Carplay and Carlife, and the third-party applications do not provide noise reduction functions. . The noise reduction processing of the mobile phone Internet voice audio signal through the noise reduction unit of the audio hardware abstraction layer is to perform soft noise reduction on the mobile phone Internet voice audio signal.

It can be understood that after the DSP chip collects the audio signal from the microphone and the vehicle system determines that it is the mobile phone Internet voice audio signal, it needs to perform voice monitoring when playing music, transmit the local voice audio signal to the CPU, and use it in the audio hardware The abstraction layer performs noise reduction processing on the mobile phone Internet voice audio signal through the vehicle-mounted SOC system.

The audio noise processing method provided in Embodiment 1 of the present invention separates the audio signal according to the source of the audio signal, uses a DSP noise reduction chip for call audio signal to reduce noise, and uses a voice noise reduction algorithm for local voice audio signal to reduce noise. , for mobile phone Internet voice audio signals, the noise reduction unit of the hardware abstraction layer is used for noise reduction, that is, different noise reduction methods are used for audio signals from different sources to reduce vehicle audio noise interference, improve vehicle audio quality, and reduce the hardware cost of noise processing. and development costs.

In addition, the noise reduction process of the call audio signal does not occupy CPU performance, does not need to modify the HAL layer code, and does not require authorization or activation to affect the production line. It decouples the voice resource package and Bluetooth phone, etc., reducing the probability of call failure; for calls After-sales updates for audio signal noise reduction do not require replacement of the DSP chip. After-sales updates are only performed through software upgrades and refreshes. The voice noise reduction algorithm can be upgraded OTA at any time to reduce hardware refresh costs.

For the noise reduction process of local voice audio signals, since each car model has different microphone layout positions, angles, spacing, interior decorations, etc., as well as different noise environments such as engines, air conditioners, tire noise, and road noise in the car, in order to ensure good speech recognition Effect, voice parameters such as voice wake-up threshold and in-car reverberation signal model need to be calibrated and adjusted as the vehicle model changes. The voice noise reduction algorithm proposed in this solution is integrated in the APP layer. Only the APP layer parameters are modified, and there is no need to modify the HAL layer. It can be quickly modified and adapted to each car model. The voice tuning parameters can be more flexibly adjusted according to different car models and different application scenarios. It can also reduce the need for separate hardware noise reduction modules, reduce PCB occupation, reduce dependence on hardware, and greatly reduce costs. , after-sales updates for noise reduction of local voice audio signals and flexible adjustment of tuning parameters without the need to update firmware, reducing refresh costs.

Regarding the noise reduction process of mobile phone Internet voice audio signals, since the DSP cannot support simultaneous noise reduction for voice and music mixing, it cannot support the mobile phone Internet voice audio for third-party applications such as Carplay and Carlife that do not provide noise reduction functions. The signal is denoised. Therefore, the noise reduction process of the mobile phone Internet voice audio signal is processed by the vehicle SOC system at the HAL layer, which reduces PCB occupation, reduces dependence on hardware, and greatly reduces costs.

Please refer to Figure 2. In the audio noise processing method provided in Embodiment 2 of the present invention, steps S100-S400 are the same as steps S100-S400 in Embodiment 1. The difference lies in that: in step S200, DSP is used to reduce noise. The chip performs noise reduction processing on the call audio signal, including:

S210. Use the DSP noise reduction chip to gain the call audio signal to obtain a first gain signal.

It should be noted that the gain is the signal amplification factor.

S220. Perform noise separation and filtering on the first gain signal to obtain a monophonic audio signal.

It should be noted that the separator filtering is used to denoise the signal.

It is understandable that the DSP noise reduction chip gains the call audio signal, separates and filters the external environmental noise, and eliminates the noise, thereby improving the call quality.

S230 resamples the mono audio signal so that the sampling rate of the mono audio signal meets the sampling rate of the input signal of the call module, and uses the resampled mono audio signal as the first Noise reduction signal.

It should be noted that resampling means resampling the signal to obtain a signal that satisfies the sampling rate of the input signal of the call module. Since the sampling rate of the signal output by the DSP noise reduction chip is different from the sampling rate of the input signal of the call module, the signal needs to be resampled. For example, the sampling rate of the signal output by the DSP noise reduction chip is 48k, and the sampling rate of the signal is changed to one of the sampling rates of 8K, 24K or 16K of the input signal of the call module through resampling.

Preferably, the mono audio signal is resampled in step S230, specifically:

The monophonic audio signal is resampled by hardware through the built-in ADSP chip in the SOC chip.

It should be noted that the built-in ADSP chip in the SOC chip is used for hardware resampling and then sent to the core algorithm processing engine instead of software resampling implementation at the Android framework layer, which greatly reduces system overhead and increases system efficiency.

The audio noise processing method provided in Embodiment 2 of the present invention uses a DSP noise reduction chip to gain, separate, filter and resample the call audio signal, so that the call audio signal output by the DSP can be completely input to the call module. To complete two-way call interaction and response. The noise reduction process for call audio signals does not occupy CPU performance, does not require changes to the HAL layer code, and does not require authorization or activation to affect the production line. It decouples the voice resource package from Bluetooth phones, etc., reducing the probability of call failures; for call audio signals After-sales updates for noise reduction do not require replacement of the DSP chip. After-sales updates are only performed through software upgrades and refreshes. The voice noise reduction algorithm can be upgraded OTA at any time to reduce hardware refresh costs.

Please refer to Figure 3. In the audio noise processing method provided in Embodiment 3 of the present invention, steps S100-S400 are the same as steps S100-S400 in Embodiment 1. The difference is that in step S300, audio hardware abstraction is used. layer is input to the voice assistant application software, and before the voice audio signal is subjected to noise reduction processing and echo cancellation processing through the voice noise reduction algorithm integrated in the voice assistant application software, it also includes:

S310. Obtain the audio signal at the speaker end as a reference signal.

It should be noted that the DSP collects the audio signal from the speaker in real time as a reference signal.

S320. Mix the reference signal and the speech audio signal to obtain a mixed audio signal.

It should be noted that the DSP collects and converts the reference signal and the voice audio signal and other multi-channel signals through an ADC (analog-to-digital converter), and then mixes them through time-division multiplexing (TDM) mode. After receiving one signal, it transmits it to the CPU to realize simultaneous transmission of multiple signals.

S330: Input the mixed audio signal into the CPU, and call the voice noise reduction algorithm integrated in the voice assistant application software.

It should be noted that the voice noise reduction algorithm integrated in the voice assistant application software in the SOC is called to achieve noise reduction and echo elimination of the voice audio signal to ensure good recognition effect. The interface functions of the speech noise reduction algorithm are as shown in Table 1:

Table 1

It can be understood that the voice noise reduction algorithm for voice audio signal noise reduction and echo cancellation is integrated in the voice assistant application software, which does not require HAL layer modification, and can reduce the need for separate hardware noise reduction modules, reduce PCB occupation, and reduce Reliance on hardware greatly reduces costs, and tuning parameters can be flexibly adjusted without the need to update firmware, reducing refresh costs.

The audio noise processing method provided in Embodiment 3 of the present invention, for the noise reduction process of local voice audio signals, due to the differences in the microphone arrangement position, angle, spacing, interior decoration, etc. of each vehicle model, as well as the in-car engine, air conditioner, tire noise, Noise environments such as road noise are different. In order to ensure good speech recognition effects, speech parameters such as voice wake-up threshold and in-car reverberation signal model need to be calibrated and adjusted as the car model changes. By mixing multi-channel audio, it is integrated into The voice noise reduction algorithm in the voice assistant application software de-noises the voice audio signal. Since the voice noise reduction algorithm is integrated in the APP layer, only the APP layer parameters are modified. It can be quickly modified and adapted to each car model without the need for HAL layer modification. According to different car models and different application scenarios, the voice tuning parameters can be more flexibly adjusted, and separate hardware noise reduction modules can be reduced, reducing PCB occupation, reducing dependence on hardware, and greatly reducing costs. For the noise reduction of local voice audio signals After-sales updates allow for flexible adjustment of tuning parameters without the need to update firmware, reducing refresh costs.

In the audio noise processing method provided in Embodiment 4 of the present invention, steps S100-S400 are the same as steps S100-S400 in Embodiment 1, except that:

Before performing noise reduction processing on the call audio signal through the DSP noise reduction chip in step S200, it also includes:

Set the audio source program to call mode, and switch the DSP noise reduction chip to call channel mode through the SPI serial communication protocol.

It should be noted that the SPI serial communication protocol is a synchronous serial interface technology. It is a high-speed, full-duplex, synchronous communication bus. It has the advantages of supporting full-duplex communication, simple communication, and fast data transmission rate.

It can be understood that after the DSP chip collects the audio signal from the microphone and the vehicle system determines that it is a call audio signal, it sets the audio source program to call mode and notifies the DSP to switch to call channel mode through an SPI serial port message.

In step S300, before performing noise reduction processing and echo cancellation processing on the local voice audio signal through the voice noise reduction algorithm integrated in the voice assistant application software, it also includes:

Set the audio source program to intelligent voice mode, and switch the DSP noise reduction chip to intelligent voice channel mode.

It is understandable that after the DSP chip collects the audio signal from the microphone and the vehicle system determines that it is a local voice audio signal, it sets the audio source program to the intelligent voice mode and notifies the DSP to switch to the intelligent voice channel mode through the SPI serial port message.

If the same soft noise reduction mode as the local voice audio signal is used to reduce noise on the call audio signal, there will be many difficulties and problems. The first problem is that the complexity and space occupied of the soft noise reduction algorithm will increase, which will consume CPU running performance and lead to possible delays. The second problem is that call audio signal noise reduction is different from voice audio signal noise reduction. Since the voice audio signal noise reduction algorithm can be integrated in the application layer voice assistant APP, the Bluetooth call algorithm BTNR in the call audio signal soft noise reduction solution involves the underlying layer. Deeper, it is necessary to integrate the soft noise reduction library in the HAL layer, call the interface library and modify the function to configure and create a recording thread, and finally reduce the recording and send it to the voice assistant to achieve soft noise reduction. This solution will require deep changes to the Android HAL layer interface, which is risky and prone to problems. The third problem is that general soft noise reduction algorithms require authorization activation. If the call audio signal also uses the soft noise reduction algorithm, it will be the same as the local voice audio signal soft noise reduction. The production line needs to perform at least two call and voice algorithm calibrations. It increases the complexity of the production line, slows down the production cycle, and is prone to unauthorized issues in certain scenarios. The fourth problem is that the call audio signal soft noise reduction algorithm and the local voice audio signal soft noise reduction algorithm need to be stored in the voice resource package. When the voice resource package is missing, it will also directly affect the phone recording function. The coupling is large, and the phone The function itself requires a higher level than voice, and the voice resource package is large and prone to defects due to copy upgrades, etc., resulting in an increased probability of phone unavailability.

Therefore, Embodiment 4 of the present invention proposes an audio noise processing method. The call audio signal noise reduction is based on the DSP chip for hard noise reduction: the user inputs audio to the microphone, and the DSP samples the call audio such as Bluetooth/CarPlay phone/ECALL/BCAll and then directly The underlying DSP chip performs noise reduction processing internally, and the noise-reduced recording is finally transmitted to the Bluetooth APP application to complete the call recognition function. Using this solution, call audio signal noise reduction does not require CPU performance, does not need to modify the HAL layer code, and does not require authorization or activation to affect the production line. It decouples the voice resource package from Bluetooth phones and reduces the probability of call failures.

The second aspect.

Please refer to Figure 4. Embodiment 5 of the present invention provides an audio noise processing system, including:

The audio signal acquisition and audio source determination module 100 acquires the audio signal from the microphone and determines the audio source of the audio signal.

It should be noted that the audio signal at the microphone end is the user's input signal to the microphone end. After the DSP chip collects the audio signal at the microphone end, the vehicle system determines the source of the audio signal.

It can be understood that since different audio signals have different audio channels, determining the sound source of the audio signal may be: judging the sound source of the audio signal through the sound source circulation channel of the audio signal. Since different audio signals have different frequency domains, determining the sound source of the audio signal may also include determining the sound source of the audio signal based on the frequency domain of the audio signal.

Preferably, the sound source of the audio signal is determined, specifically:

The sound source of the audio signal is determined through the sound source circulation channel of the audio signal.

The call audio signal processing module 200 is used to perform noise reduction processing on the call audio signal through a DSP noise reduction chip when the source of the audio signal is a call audio signal, and use the first noise reduction signal obtained by the noise reduction processing to It is input to the call module through the audio hardware abstraction layer to complete two-way call interaction and response.

It should be noted that the call audio signals include: Bluetooth calls, CarPlay calls, ECALL, BCAll and other call audio signals. The noise reduction processing of the call audio signal through the DSP noise reduction chip is the hard noise reduction processing of the call audio signal.

It can be understood that after the DSP chip collects the audio signal from the microphone and the vehicle system determines that it is a call audio signal, the DSP chip directly performs noise reduction processing inside the chip to obtain the first noise reduction signal, and the first noise reduction signal is passed through the audio signal The hardware abstraction layer is transmitted to the call module to complete two-way call interaction and response.

The local voice audio signal processing module 300 is used to input the audio signal to the voice assistant application software through the audio hardware abstraction layer when the source of the audio signal is a local voice audio signal, and use the voice reduction function integrated in the voice assistant application software to The noise algorithm performs noise reduction processing and echo cancellation processing on the local voice audio signal.

It should be noted that local voice audio signals include: vehicle-mounted artificial intelligence voice, vehicle-mounted voice recognition assistant and other local voice audio signals. Performing noise reduction processing and echo cancellation processing on the local voice audio signal through the voice noise reduction algorithm integrated in the voice assistant application software is to perform soft noise reduction on the local voice audio signal.

It can be understood that after the DSP chip collects the audio signal at the microphone end and the vehicle system determines that it is a local voice audio signal, the local voice audio signal is transmitted to the CPU and input to the voice assistant application software through the audio hardware abstraction layer, and through The voice noise reduction algorithm integrated in the voice assistant application software performs noise reduction processing and echo cancellation processing on the local voice audio signal.

The mobile phone Internet voice audio signal processing module 400 is used to perform noise reduction processing on the mobile phone Internet voice audio signal through the noise reduction unit of the audio hardware abstraction layer when the source of the audio signal is a mobile phone Internet voice audio signal.

It should be noted that the mobile phone Internet voice audio is a mix of voice and music. The mobile phone Internet voice audio signal includes: the mobile phone Internet voice audio signal of third-party applications such as carplay and Carlife, and the third-party application does not provide noise reduction function. . The noise reduction processing of the mobile phone Internet voice audio signal through the noise reduction unit of the audio hardware abstraction layer is to perform soft noise reduction on the mobile phone Internet voice audio signal.

It can be understood that after the DSP chip collects the audio signal from the microphone and the vehicle system determines that it is a mobile phone interconnected voice audio signal, it needs to perform voice monitoring when playing music, transmit the local voice audio signal to the CPU, and perform it at the HAL layer The mobile phone Internet voice audio signal is subjected to noise reduction processing through the vehicle-mounted SOC system.

The audio noise processing system provided in Embodiment 5 of the present invention separates the audio signal according to the source of the audio signal, uses a DSP noise reduction chip for call audio signal to reduce noise, and uses a speech noise reduction algorithm for local voice audio signal to reduce noise. , for mobile phone Internet voice audio signals, the noise reduction unit of the hardware abstraction layer is used for noise reduction, that is, different noise reduction methods are used for audio signals from different sources to reduce vehicle audio noise interference, improve vehicle audio quality, and reduce the hardware cost of noise processing. and development costs.

In addition, the noise reduction process of the call audio signal does not occupy CPU performance, does not need to modify the HAL layer code, and does not require authorization or activation to affect the production line. It decouples the voice resource package and Bluetooth phone, etc., reducing the probability of call failure; for calls After-sales updates for audio signal noise reduction do not require replacement of the DSP chip. After-sales updates are only performed through software upgrades and refreshes. The voice noise reduction algorithm can be upgraded OTA at any time to reduce hardware refresh costs.

For the noise reduction process of local voice audio signals, since each car model has different microphone layout positions, angles, spacing, interior decorations, etc., as well as different noise environments such as engines, air conditioners, tire noise, and road noise in the car, in order to ensure good speech recognition Effect, voice parameters such as voice wake-up threshold and in-vehicle reverberation signal model need to be calibrated and adjusted as the vehicle model changes. The voice noise reduction algorithm proposed in this solution is integrated in the APP layer, and only the APP layer parameters are modified, without the need for HAL layer modification, that is It can be quickly modified and adapted to each car model. The voice tuning parameters can be more flexibly adjusted according to different car models and different application scenarios. It can also reduce the need for separate hardware noise reduction modules, reduce PCB occupation, reduce dependence on hardware, and greatly reduce costs. After-sales updates for noise reduction of local voice audio signals and flexible adjustment of tuning parameters without the need to update firmware, reducing refresh costs.

Regarding the noise reduction process of mobile phone Internet voice audio signals, since the DSP cannot support simultaneous noise reduction for voice and music mixing, it cannot support the mobile phone Internet voice audio for third-party applications such as Carplay and Carlife that do not provide noise reduction functions. The signal is denoised. Therefore, the noise reduction process of the mobile phone Internet voice audio signal is processed by the vehicle SOC system at the HAL layer, which reduces PCB occupation, reduces dependence on hardware, and greatly reduces costs.

The modules 100-400 of the audio noise processing system provided in Embodiment 6 of the present invention are all the same as the modules 100-400 in Embodiment 5. The difference is that the call audio signal processing module 200 is also used to perform steps S210-S230:

S210. Use the DSP noise reduction chip to gain the call audio signal to obtain a first gain signal.

It should be noted that the gain is the signal amplification factor.

S220. Perform noise separation and filtering on the first gain signal to obtain a monophonic audio signal.

It should be noted that the separator filtering is used to denoise the signal.

It is understandable that the DSP noise reduction chip gains the call audio signal, separates and filters the external environmental noise, and eliminates the noise, thereby improving the call quality.

S230 resamples the mono audio signal so that the sampling rate of the mono audio signal meets the sampling rate of the input signal of the call module, and uses the resampled mono audio signal as the first Noise reduction signal.

It should be noted that resampling means resampling the signal to obtain a signal that satisfies the sampling rate of the input signal of the call module. Since the sampling rate of the signal output by the DSP noise reduction chip is different from the sampling rate of the input signal of the call module, the signal needs to be resampled. For example, the sampling rate of the signal output by the DSP noise reduction chip is 48k, and the sampling rate of the signal is changed to one of the sampling rates of 8K, 24K or 16K of the input signal of the call module through resampling.

Preferably, the mono audio signal is resampled in step S230, specifically:

The monophonic audio signal is resampled by hardware through the built-in ADSP chip in the SOC chip.

It should be noted that the built-in ADSP chip in the SOC chip is used for hardware resampling and then sent to the core algorithm processing engine instead of software resampling implementation at the Android framework layer, which greatly reduces system overhead and increases system efficiency.

The audio noise processing system provided in Embodiment 6 of the present invention uses a DSP noise reduction chip to gain, separate, filter and resample the call audio signal, so that the call audio signal output by the DSP can be completely input to the call module. To complete two-way call interaction and response. The noise reduction process for call audio signals does not occupy CPU performance, does not require changes to the HAL layer code, and does not require authorization or activation to affect the production line. It decouples the voice resource package from Bluetooth phones, etc., reducing the probability of call failures; for call audio signals After-sales updates for noise reduction do not require replacement of the DSP chip. After-sales updates are only performed through software upgrades and refreshes. The voice noise reduction algorithm can be upgraded OTA at any time to reduce hardware refresh costs.

The modules 100-400 of the audio noise processing system provided in Embodiment 7 of the present invention are all the same as the modules 100-400 in Embodiment 5. The difference is that the local voice audio signal processing module 300 is also used to execute Steps S310-S3330:

S310. Obtain the audio signal at the speaker end as a reference signal.

It should be noted that the DSP collects the audio signal from the speaker in real time as a reference signal.

S320. Mix the reference signal and the speech audio signal to obtain a mixed audio signal.

It should be noted that the DSP collects and converts the reference signal and the voice audio signal and other multi-channel signals through the ADC, and mixes them through the TDM time division multiplexing mode. After obtaining one channel of signal, it is transmitted to the CPU to realize multiple signals. signals are transmitted simultaneously.

S330: Input the mixed audio signal into the CPU, and call the voice noise reduction algorithm integrated in the voice assistant application software.

It should be noted that the voice noise reduction algorithm integrated in the voice assistant application software in the SOC is called to achieve noise reduction and echo elimination of the voice audio signal to ensure good recognition effect. The interface functions of the speech noise reduction algorithm are as shown in Table 1:

Table 1

It can be understood that the voice noise reduction algorithm for voice audio signal noise reduction and echo cancellation is integrated in the voice assistant application software, which does not require HAL layer modification, and can reduce the need for separate hardware noise reduction modules, reduce PCB occupation, and reduce Reliance on hardware greatly reduces costs, and tuning parameters can be flexibly adjusted without the need to update firmware, reducing refresh costs.

The audio noise processing system provided in Embodiment 7 of the present invention, for the noise reduction process of local voice audio signals, due to the differences in the microphone arrangement position, angle, spacing, interior decoration, etc. of each vehicle model, as well as the in-car engine, air conditioner, tire noise, Noise environments such as road noise are different. In order to ensure good speech recognition effects, speech parameters such as voice wake-up threshold and in-car reverberation signal model need to be calibrated and adjusted as the car model changes. By mixing multi-channel audio, it is integrated into The voice noise reduction algorithm in the voice assistant application software de-noises the voice audio signal. Since the voice noise reduction algorithm is integrated in the APP layer, only the APP layer parameters are modified. It can be quickly modified and adapted to each car model without the need for HAL layer modification. According to different car models and different application scenarios, the voice tuning parameters can be more flexibly adjusted, and separate hardware noise reduction modules can be reduced, reducing PCB occupation, reducing dependence on hardware, and greatly reducing costs. For the noise reduction of local voice audio signals After-sales updates allow for flexible adjustment of tuning parameters without the need to update firmware, reducing refresh costs.

The modules 100-400 of the audio noise processing system provided in Embodiment 8 of the present invention are all the same as the modules 100-400 in Embodiment 5. The difference is that the call audio signal processing module 200 is also used to:

Set the audio source program to call mode, and switch the DSP noise reduction chip to call channel mode through the SPI serial communication protocol.

It should be noted that the SPI serial communication protocol is a synchronous serial interface technology. It is a high-speed, full-duplex, synchronous communication bus. It has the advantages of supporting full-duplex communication, simple communication, and fast data transmission rate.

It can be understood that after the DSP chip collects the audio signal from the microphone and the vehicle system determines that it is a call audio signal, it sets the audio source program to call mode and notifies the DSP to switch to call channel mode through an SPI serial port message.

The local voice audio signal processing module 300 is also used to:

Set the audio source program to intelligent voice mode, and switch the DSP noise reduction chip to intelligent voice channel mode.

It is understandable that after the DSP chip collects the audio signal from the microphone and the vehicle system determines that it is a local voice audio signal, it sets the audio source program to the intelligent voice mode and notifies the DSP to switch to the intelligent voice channel mode through the SPI serial port message.

If the same soft noise reduction mode as the local voice audio signal is used to reduce noise on the call audio signal, there will be many difficulties and problems. The first problem is that the complexity and space occupied of the soft noise reduction algorithm will increase, which will consume CPU running performance and lead to possible delays. The second problem is that call audio signal noise reduction is different from voice audio signal noise reduction. Since the voice audio signal noise reduction algorithm can be integrated in the application layer voice assistant APP, the Bluetooth call algorithm BTNR in the call audio signal soft noise reduction solution involves the underlying layer. Deeper, it is necessary to integrate the soft noise reduction library in the HAL layer, call the interface library and modify the function to configure and create a recording thread, and finally reduce the recording and send it to the voice assistant to achieve soft noise reduction. This solution will require deep changes to the Android HAL layer interface, which is risky and prone to problems. The third problem is that general soft noise reduction algorithms require authorization activation. If the call audio signal also uses the soft noise reduction algorithm, it will be the same as the local voice audio signal soft noise reduction. The production line needs to perform at least two call and voice algorithm calibrations. It increases the complexity of the production line, slows down the production pace, and is prone to unauthorized problems in certain scenarios. The fourth problem is that the call audio signal soft noise reduction algorithm and the local voice audio signal soft noise reduction algorithm need to be stored in the voice resource package. When the voice resource package is missing, it will also directly affect the phone recording function. The coupling is large, and the phone The function itself requires a higher level than voice, and the voice resource package is large and prone to defects due to copy upgrades, etc., resulting in an increased probability of phone unavailability.

Therefore, Embodiment 8 of the present invention proposes an audio noise processing system. The call audio signal noise reduction is based on the DSP chip for hard noise reduction: the user inputs audio to the microphone, and the DSP samples the call audio such as Bluetooth/CarPlay phone/ECALL/BCAll and then directly The underlying DSP chip performs noise reduction processing internally, and the noise-reduced recording is finally transmitted to the Bluetooth APP application to complete the call recognition function. Using this solution, call audio signal noise reduction does not require CPU performance, does not need to modify the HAL layer code, and does not require authorization or activation to affect the production line. It decouples the voice resource package from Bluetooth phones and reduces the probability of call failures.

The third aspect.

The invention provides an electronic device, which includes:

Processors, memories, and buses;

The bus is used to connect the processor and the memory;

The memory is used to store operating instructions;

The processor is configured to, by invoking the operation instructions, have executable instructions that cause the processor to perform operations corresponding to the audio noise processing method shown in the first aspect of this application.

In an optional embodiment, an electronic device is provided, as shown in Figure 5. The electronic device 5000 shown in Figure 5 includes: a processor 5001 and a memory 5003. Among them, the processor 5001 and the memory 5003 are connected, such as through a bus 5002. Optionally, electronic device 5000 may also include a transceiver 5004. It should be noted that in practical applications, the number of transceivers 5004 is not limited to one, and the structure of the electronic device 5000 does not limit the embodiments of the present application.

The processor 5001 may be a CPU, a general-purpose processor, a DSP, an ASIC, an FPGA, or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute the various illustrative logical blocks, modules, and circuits described in connection with this disclosure. The processor 5001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

Bus 5002 may include a path that carries information between the above-mentioned components. The bus 5002 may be a PCI bus or an EISA bus, etc. The bus 5002 can be divided into an address bus, a data bus, a control bus, etc. For ease of presentation, only one thick line is used in Figure 5, but it does not mean that there is only one bus or one type of bus.

The memory 5003 can be ROM or other types of static storage devices that can store static information and instructions, RAM or other types of dynamic storage devices that can store information and instructions, or it can be EEPROM, CD-ROM or other optical disk storage, optical disk storage (including compressed optical discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), disk storage media or other magnetic storage devices, or can be used to carry or store desired program code in the form of instructions or data structures and can be used by a computer Any other medium for access, but not limited to this.

The memory 5003 is used to store application program code for executing the solution of the present application, and is controlled by the processor 5001 for execution. The processor 5001 is configured to execute the application program code stored in the memory 5003 to implement the content shown in any of the foregoing method embodiments.

Among them, electronic devices include but are not limited to: mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PAD (tablet computers), PMP (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc. mobile terminals such as digital TVs, desktop computers, etc.

The fourth aspect.

The present invention provides a computer-readable storage medium. A computer program is stored on the computer-readable storage medium. When the program is executed by a processor, the audio noise processing method shown in the first aspect of the application is implemented.

Another embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When run on a computer, the computer can execute the corresponding content in the foregoing method embodiment.

Claims (9)

1. The audio noise processing method is characterized by being applied to a vehicle-mounted terminal and comprising the following steps of:

Acquiring an audio signal of a microphone end and determining a sound source of the audio signal;

if the sound source of the audio signal is a call audio signal, the call audio signal is subjected to noise reduction through a DSP noise reduction chip, and a first noise reduction signal obtained through the noise reduction is input to a call module through an audio hardware abstraction layer so as to complete two-way call interaction and response;

if the audio source of the audio signal is a local voice audio signal, the local voice audio signal is input into voice assistant application software through an audio hardware abstraction layer, and noise reduction processing and echo cancellation processing are carried out on the local voice audio signal through a voice noise reduction algorithm integrated in the voice assistant application software;

if the sound source of the audio signal is a mobile phone interconnection voice audio signal, the mobile phone interconnection voice audio signal is subjected to noise reduction processing through a noise reduction unit of an audio hardware abstraction layer, and the mobile phone interconnection voice audio signal is subjected to noise reduction processing through a vehicle-mounted SOC system of a HAL layer;

the noise reduction processing of the call audio signal by the DSP noise reduction chip comprises the following steps:

the communication audio signal is gained through the DSP noise reduction chip, and a first gain signal is obtained;

Noise separation and filtering are carried out on the first gain signal, and a mono audio signal is obtained;

resampling the mono audio signal to enable the sampling rate of the mono audio signal to meet the sampling rate of the input signal of the call module, and taking the resampled mono audio signal as a first noise reduction signal.

2. The noise processing method of audio according to claim 1, wherein the resampling of the mono audio signal is performed, in particular:

and carrying out hardware resampling on the mono audio signal through an ADSP chip built in the SOC chip.

3. The method for processing audio noise according to claim 1, wherein the audio signal is input to a voice assistant application software via an audio hardware abstraction layer and subjected to noise reduction processing and echo cancellation processing by a voice noise reduction algorithm integrated in the voice assistant application software, further comprising:

acquiring an audio signal of a loudspeaker end as a reference signal;

mixing the reference signal with the voice audio signal to obtain a mixed audio signal;

and inputting the mixed audio signal into a CPU, and calling the voice noise reduction algorithm integrated in the voice assistant application software.

4. The method for processing audio noise according to claim 1, wherein before the noise reduction processing is performed on the call audio signal by the DSP noise reduction chip, the method further comprises:

and setting the audio source program as a call mode, and switching the DSP noise reduction chip into a call channel mode through an SPI serial communication protocol.

5. The method for noise processing of audio according to claim 1, wherein said noise reduction processing and echo cancellation processing of said local speech audio signal by a speech noise reduction algorithm integrated in a speech assistant application software is preceded by:

and setting the audio source program to an intelligent voice mode, and switching the DSP noise reduction chip to an intelligent voice channel mode.

6. The method for noise processing of audio according to claim 1, wherein said determining the audio source of said audio signal, in particular:

and judging the sound source of the audio signal through the sound source circulation channel of the audio signal.

7. An audio noise processing system, which is applied to a vehicle-mounted terminal, comprising:

the audio signal acquisition and sound source judgment module is used for acquiring an audio signal of a microphone end by a sound source and determining the sound source of the audio signal;

The communication audio signal processing module is used for carrying out noise reduction processing on the communication audio signal through the DSP noise reduction chip when the audio source of the audio signal is the communication audio signal, and inputting a first noise reduction signal obtained by the noise reduction processing to the communication module through the audio hardware abstraction layer so as to complete two-way communication interaction and response;

the local voice audio signal processing module is used for inputting the local voice audio signal into voice assistant application software through an audio hardware abstraction layer when the audio source of the audio signal is the local voice audio signal, and carrying out noise reduction processing and echo cancellation processing on the local voice audio signal through a voice noise reduction algorithm integrated in the voice assistant application software;

the mobile phone interconnection voice audio signal processing module is used for carrying out noise reduction processing on the mobile phone interconnection voice audio signal through a noise reduction unit of an audio hardware abstraction layer when the sound source of the audio signal is the mobile phone interconnection voice audio signal;

carrying out noise reduction processing on the mobile phone interconnection voice audio signals by a vehicle-mounted SOC system of the HAL layer;

the noise reduction processing of the call audio signal by the DSP noise reduction chip comprises the following steps:

the communication audio signal is gained through the DSP noise reduction chip, and a first gain signal is obtained;

Noise separation and filtering are carried out on the first gain signal, and a mono audio signal is obtained;

resampling the mono audio signal to enable the sampling rate of the mono audio signal to meet the sampling rate of the input signal of the call module, and taking the resampled mono audio signal as a first noise reduction signal.

8. An electronic device comprising a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, the processor implementing the method of noise processing of audio according to any one of claims 1 to 6 when the computer program is executed.

9. A computer readable storage medium, characterized in that the computer readable storage medium comprises a stored computer program, wherein the computer program, when run, controls a device in which the computer readable storage medium is located to perform the noise processing method of audio according to any one of claims 1 to 6.