Low latency bypass audio channel method and apparatus for critical voice interaction
By introducing a low-latency bypass audio channel into the operating system and utilizing TEE authentication and direct hardware-level audio data transmission, the problems of high latency, large fluctuations, and insufficient security in key voice interactions in existing technologies are solved, achieving a low-latency, efficient, and secure voice interaction solution.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-13
- Publication Date
- 2026-07-31
AI Technical Summary
Existing technologies suffer from high end-to-end latency, large latency fluctuations, susceptibility to interference from other applications or processes within the system, insufficient security, and low processing efficiency in key voice interaction scenarios.
By introducing a low-latency bypass audio channel into the operating system and using a Trusted Execution Environment (TEE) for security authentication and authorization verification, the standard audio stack is bypassed, and audio data is transmitted directly between the ADC/DAC and the audio processing unit. Priority preemption and resource exclusivity strategies are used to ensure the stability and security of critical audio streams.
It achieves end-to-end latency at the millisecond or even sub-millisecond level, improving system robustness and stability, reducing power consumption, ensuring the continuity and security of critical voice interactions, and possessing good compatibility and practicality.
Smart Images
Figure CN122496231A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of audio processing technology, and more specifically to a method and apparatus for low-latency bypass audio channels for critical voice interaction. Background Technology
[0002] Against the backdrop of rapid iteration in mobile communications, IoT, and AI technologies, voice interaction has become the core human-computer interaction method for smart terminals (such as mobile phones, in-vehicle systems, and smart homes). Especially in scenarios with stringent requirements for latency and stability, such as in-vehicle emergency calls and real-time voice control, audio processing performance directly impacts user experience and safety.
[0003] Mainstream smart device operating systems (such as Android and Linux) adopt a layered, software-centric audio architecture. Taking Android as an example, audio data needs to pass through multiple software stacks, including application buffers, framework layer mixing / sound effects processing, and Hardware Abstraction Layer (HAL), before finally reaching the ADC / DAC device to support concurrent applications and format conversion.
[0004] However, this general-purpose architecture has four core limitations when dealing with key voice interaction scenarios:
[0005] First, the latency is high and unstable. The multi-layered software processing chain results in end-to-end latency of tens to hundreds of milliseconds, and the latency fluctuates greatly due to factors such as system scheduling and memory contention, making it impossible to meet the requirements for millisecond-level low latency.
[0006] Second, there is the conflict of system resources. When multiple applications compete for audio resources, non-critical tasks can easily block high-priority voice interaction, leading to stuttering, loss of commands, and reduced system real-time performance and robustness.
[0007] Third, there is a lack of security mechanisms. The standard audio stack lacks strong authentication for audio streams, allowing malicious applications to illegally access audio devices and posing a risk of eavesdropping or interfering with critical services.
[0008] Fourth, it suffers from low processing efficiency. Complex algorithms such as noise suppression and speech recognition rely on CPU processing, increasing the load and power consumption, and further prolonging latency.
[0009] Although the industry has tried to make local adjustments such as increasing process priority and optimizing buffers, it has been unable to eradicate the inherent defects of multi-layered software stacks; standalone MCU solutions increase hardware costs and system complexity, and are difficult to coordinate.
[0010] Therefore, there is an urgent need for a system-level innovative solution to build a secure, low-latency, and efficient key voice interaction path on the basis of existing hardware, in order to break through the bottlenecks of traditional architecture. This invention is a solution proposed to address this pain point. Summary of the Invention
[0011] To address the shortcomings of existing technologies, this invention discloses a low-latency bypass audio channel method and apparatus for critical voice interaction, which solves the problems of high end-to-end latency, large latency fluctuations, susceptibility to interference from other applications or processes within the system, insufficient security, and low processing efficiency due to lengthy software processing paths in the prior art when processing critical voice interaction tasks.
[0012] This invention is achieved through the following technical solution:
[0013] According to a first aspect of the present invention, a low-latency bypass audio channel method for critical voice interaction is provided, the method being as follows:
[0014] The method includes: an application requiring critical voice interaction initiating an access request to a bypass audio channel; the system performing security authentication and permission verification on the application; after the application passes verification, marking its associated audio stream as a critical audio stream; activating a bypass channel controller, which allocates audio hardware resources to the critical audio stream based on a preset arbitration strategy, the audio hardware resources including at least an analog-to-digital converter (ADC) and / or a digital-to-analog converter (DAC); configuring the ADC / DAC to establish a bypass audio channel that can bypass the standard audio stack of the operating system; and finally enabling the audio data of the critical audio stream to be directly transmitted between the ADC / DAC and the audio processing unit through the bypass audio channel.
[0015] Furthermore, the security authentication and permission verification are preferably performed through a Trusted Execution Environment (TEE) to verify the identity and / or signature of the application, and the permission manager authorizes the application based on the verification results.
[0016] Furthermore, the preset arbitration strategy includes a priority preemption strategy and / or a resource exclusivity strategy. The priority preemption strategy is used to ensure that when the critical audio stream starts, lower-priority standard audio streams that are accessing the audio hardware resources can be paused or suspended. The resource exclusivity strategy is used to grant exclusive access to the ADC / DAC during the activity of the critical audio stream.
[0017] Furthermore, establishing the bypass audio channel includes configuring a direct data path between the ADC / DAC and a voice acceleration processing unit. The voice acceleration processing unit can be a digital signal processor (DSP) or a neural network processor (NPU) for real-time acceleration processing of audio data transmitted through the bypass audio channel.
[0018] Furthermore, the method also includes a safety fallback step: when the bypass audio channel cannot be successfully established due to insufficient resources or hardware failure, the critical audio stream is forwarded to the standard audio stack of the operating system for processing.
[0019] Furthermore, the method also includes a resource release and recovery step: when the critical audio stream finishes using the bypass audio channel, the audio hardware resources it occupies are released, and the audio service of the operating system is notified to restore the standard control mode of the ADC / DAC.
[0020] According to a second aspect of the present invention, a low-latency bypass audio channel device for critical voice interaction is provided, the device being as follows:
[0021] The device includes: a permission management module for receiving access requests from applications to a bypass audio channel and performing security authentication and permission verification on the applications; a channel control module connected to the permission management module for marking the associated audio stream as a critical audio stream after the application passes verification, and allocating audio hardware resources to the critical audio stream based on a preset arbitration strategy, wherein the audio hardware resources include at least an analog-to-digital converter (ADC) and / or a digital-to-analog converter (DAC); and a hardware configuration module connected to the channel control module for configuring the ADC / DAC to establish a bypass audio channel that bypasses the standard audio stack of the operating system; wherein the established bypass audio channel is used to enable the direct transmission of audio data of the critical audio stream between the ADC / DAC and the audio processing unit.
[0022] Furthermore, the permission management module can be integrated into or communicate with a Trusted Execution Environment (TEE) to utilize the security capabilities of the TEE to perform security authentication of the application.
[0023] Furthermore, the device also includes a voice acceleration processing unit, and the hardware configuration module is further used to configure a direct data path between the ADC / DAC and the voice acceleration processing unit. The voice acceleration processing unit may be a digital signal processor (DSP) or a neural network processor (NPU).
[0024] Furthermore, the channel control module also includes a rollback and recovery unit, used to: trigger a safe rollback mechanism when the bypass audio channel fails to be established, forwarding the critical audio stream to the standard audio stack of the operating system; and, when the critical audio stream ends, control the release of the audio hardware resources and notify the audio service of the operating system to resume standard control.
[0025] The beneficial effects of this invention are as follows:
[0026] This invention achieves extremely low latency by bypassing the complex software processing flow of the standard audio stack of the operating system through hardware-level bypass design, thereby significantly reducing the end-to-end latency of critical audio streams to the millisecond level or even sub-millisecond level, meeting extremely high real-time requirements.
[0027] This invention significantly improves the robustness and stability of the system. Through an independent bypass channel and priority arbitration mechanism, it ensures that the critical voice stream is not interfered with by other ordinary applications or processes in the system, thus guaranteeing the continuity and determinism of voice interaction. By integrating TEE security authentication and permission management, it strictly controls access to the bypass channel, preventing unauthorized applications from abusing high-priority audio resources or eavesdropping, and providing system-level security.
[0028] This invention improves processing efficiency and reduces power consumption, supports direct audio data flow from ADC / DAC to dedicated voice acceleration processing units (such as DSP / NPU), realizes hardware acceleration processing, reduces the burden on the main CPU, helps reduce the overall power consumption of the system, and has good compatibility and practicality. The proposed solution provides a high-performance bypass channel while retaining the complete standard audio path, and ensures compatibility with existing systems and applications through safe fallback and smooth switching mechanisms, making it easy to deploy applications in various platforms and scenarios. Attached Figure Description
[0029] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0030] Figure 1 This is a schematic diagram of the system architecture of a low-latency bypass audio channel device for critical voice interaction provided in an embodiment of the present invention;
[0031] Figure 2 This is a flowchart of a low-latency bypass audio channel method for critical voice interaction provided in an embodiment of the present invention. Detailed Implementation
[0032] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0033] Reference Figure 2 As shown, the low-latency bypass audio channel method for critical voice interaction provided by this invention mainly includes the following key steps and implementation details:
[0034] The initiation and security authentication of the key audio stream in this embodiment are as follows:
[0035] In implementation, when a user application (such as a pre-installed emergency call app or a system-level voice assistant service) needs to initiate a latency-sensitive voice interaction, it will send a request through a dedicated set of user-space APIs provided by the operating system. This API is designed to be streamlined and efficient, and may contain a function call such as 'requestCriticalAudioPath()'. The request includes the application identifier and the desired audio mode (capture, playback, or full-duplex).
[0036] Once the corresponding driver in the system kernel intercepts this request, it submits it to a separate security authentication module. The implementation of this module heavily relies on the device's Trusted Execution Environment (TEE). Within the secure isolation zone of the TEE, a Trusted Application (TA) runs, responsible for verifying the identity of the requesting application. Verification methods can be based on: a) whitelist verification of the application signature; b) specific runtime permissions granted by the system permission manager (such as 'android.permission.CRITICAL_AUDIO_ACCESS'); c) or a combination of both. Only applications that pass TEE verification are granted a token by the permission manager and marked as legitimate users. This step strictly prevents unauthorized applications from abusing high-privilege bypass channels and is the cornerstone of system security.
[0037] The establishment and resource configuration of the hardware bypass channel in this embodiment are as follows:
[0038] Once authorized, the system core is taken over by a bypass channel controller for subsequent processes. This controller can be a firmware-controlled hardware state machine within the SoC audio subsystem, or a driver module running at kernel privilege levels.
[0039] The key to its implementation lies in the direct management of hardware resources:
[0040] Resource query: The controller first checks the status of the target hardware resources, including the current working mode of the ADC / DAC, whether it is occupied by the standard audio stack, and the availability of dedicated voice acceleration processing units (such as DSP / NPU).
[0041] Arbitration and Preemption: The controller executes policies based on a preset audio arbitration mechanism. For example, if the ADC is being used by a regular music application, the controller will notify Android's AudioFlinger service via kernel events or IPC mechanisms, requesting it to pause or lower the priority of that regular stream. Similar coordination may be used for DAC resources. After gaining control of the resources, the controller can configure exclusive access for critical audio streams, ensuring their data paths are not disturbed.
[0042] Hardware Direct Attachment Configuration: This is the core of achieving low latency. The controller directly configures the audio codec or the audio routing matrix within the SoC by writing to registers. The goal is to establish a direct physical data path. For example, the digital audio stream from the microphone input, after ADC conversion, is not sent to the standard AudioHAL buffer, but instead directly to the input interface of a speech recognition accelerator (such as a DSP) via a dedicated DMA channel or hardware FIFO. Similarly, the processed audio can be directly connected from the DSP output interface to the DAC for playback. This configuration process bypasses the complex data copying and conversion process of Android AudioHAL.
[0043] The data flow management and security rollback in this embodiment are as follows:
[0044] Once the bypass channel is successfully established, the application interacts with the kernel driver via a dedicated API to send and receive audio data packets with extremely low overhead. The data stream is transmitted directly between the ADC / DAC and the accelerator at the hardware level.
[0045] As part of robust design, a safety fallback mechanism must be implemented. For example, if a hardware failure (such as a DSP overheating and crashing) or an uninterruptible system core function (such as a boot ringtone) is detected when the controller attempts to allocate resources, a bypass cannot be established. In this case, the controller returns a specific error code to the application via the API and automatically redirects all subsequent operations of this audio session to the standard Android audio stack (via AudioFlinger and AudioHAL). This ensures that even if the high-end optimized path fails, critical voice functions (such as emergency calls) can still work in the standard way, albeit with higher latency.
[0046] The session termination and resource release in this embodiment are as follows:
[0047] When the key voice interaction ends, the application calls the release API. The controller performs the reverse operation:
[0048] a) Stop the direct hardware data flow.
[0049] b) Remove the exclusive configuration of resources such as ADC / DAC.
[0050] c) Notify the AudioFlinger service to resume playback or capture of previously paused normal audio streams, and achieve a smooth transition through brief volume changes to avoid popping sounds.
[0051] d) Clear the "critical" markers of the audio stream, and the controller returns to an idle state.
[0052] In one embodiment, reference is made to Figure 1 As shown, the apparatus for implementing the above method is typically physically integrated into the system-on-chip (SoC) of a mobile terminal or vehicle host, and its logical structure and interaction relationship are as follows:
[0053] The implementation of the core functional modules in this embodiment is as follows:
[0054] Security Module / Trusted Execution Environment (TEE): This is a hardware area within the SoC with independent secure storage and execution capabilities. Trusted applications (TAs) running within it are responsible for executing sensitive authentication logic.
[0055] Authentication / Permission Manager: This module can run as a system service in a Rich Execution Environment (REE), but it has a secure channel with the TEE. It receives the authentication results from the TEE and maintains a dynamic list of permission policies, determining which application ID to grant bypass channel access rights to. It is also the decision point for issuing the start command to the bypass channel controller.
[0056] Bypass channel controller: The core logic unit of this invention. Its implementation can be as follows:
[0057] Hardware implementation: As part of the SoC audio subsystem or power management unit (PMU), it is a finite state machine defined by a hardware description language (HDL) that triggers a series of hardwired operations in response to specific register writes.
[0058] Firmware implementation: It runs on a low-power microcontroller (MCU) that is separate from the main CPU and is dedicated to managing low-latency peripherals such as audio and sensors.
[0059] Software implementation: As a high-priority Linux kernel driver module, it implements its functions by manipulating hardware registers and coordinating kernel scheduling.
[0060] Regardless of the implementation, the controller has built-in arbitration strategy logic (such as a priority comparator) and is responsible for interacting directly with the hardware configuration logic.
[0061] Hardware configuration module: This is the execution end of the controller. In essence, it is a set of low-level drivers inside the SoC that program and configure hardware such as audio codecs, crossbars, DMA controllers, and DSP / NPU interfaces.
[0062] Speech recognition accelerator: refers to a dedicated computing unit integrated in a SoC, such as a DSP or NPU. The key to its implementation lies in the fact that its audio data interface (such as McASP, I2S, SLIMbus) can be directly mapped to the digital stream interface of the ADC / DAC by the bypass channel controller through a hardware configuration module, forming a controllable hardware data link.
[0063] The implementation of this embodiment in conjunction with the Android system is as follows:
[0064] The entire device needs to work seamlessly with the existing Android audio architecture:
[0065] User-space API: A new set of Native APIs (C / C++ interfaces) is implemented in the Android framework layer (such as 'frameworks / av'), possibly encapsulated in the 'libaudiobypass.so' library, for legitimate system applications to call.
[0066] Kernel driver: Implement a character device driver (such as ' / dev / audiobypass'), handle ioctl calls from user-space APIs, communicate with the bypass channel controller (whether it is hardware or firmware), and perform necessary state synchronization with the existing AudioHAL.
[0067] Coordination with AudioFlinger / AudioHAL: This is the implementation challenge. The bypass channel controller (or the kernel driver representing it) needs to establish a communication mechanism with AudioFlinger. For example, an "audio resource priority notification" service can be defined through Android's HIDL or AIDL interface. When the bypass channel needs to preempt resources, it notifies AudioFlinger through this service, and AudioFlinger then pauses the corresponding Track or Stream. This coordination ensures the overall consistency of system audio behavior, avoiding logical conflicts or hardware access contention.
[0068] In one embodiment, an in-vehicle emergency call system (eCall) is provided: when a serious accident occurs and the airbags deploy, automatically triggering the eCall, the application immediately requests a bypass audio channel. The system verifies its legitimacy via a TEE (ensuring that the accident is not fabricated by malicious software), and then the bypass channel is established. Voice communication between the injured person and the rescue center is conducted directly through the in-vehicle microphone and speakers with minimal latency, avoiding call delays or interruptions caused by the in-vehicle infotainment system playing music or navigation voices, thus gaining valuable time for rescue.
[0069] In one embodiment, a real-time wake-up and response system for a smart voice assistant is provided: During driving, after the user utters a "wake-up word," the voice assistant's hot word detection module (running on the DSP) captures audio in real time through a bypass channel. Once detection is successful, the assistant enters an interactive state, and subsequent user voice commands continue to be collected through this low-latency channel and sent to the cloud or local NPU for recognition. This ensures an ultra-fast experience in the "wake-up-response" chain, giving the user the feeling that the device is "instantly responsive" without any perceptible lag.
[0070] In one embodiment, a highly reliable voice intercom application is provided: In security, industrial inspection, or specific team communication scenarios, a dedicated intercom app can be authorized to use a bypass channel. When the user presses the PTT (Push-to-Talk) button, voice data is directly transmitted and received through the bypass channel, with extremely low latency and stability, ensuring the immediacy and clarity of command issuance and situation reporting, unaffected by interference from other application notification tones or media playback on the device.
[0071] In one embodiment, an assistive hearing device compatibility mode is provided: for hearing aids or assistive hearing devices connected to a mobile phone, the accompanying app can request a bypass channel when extremely low-latency audio pass-through processing is required (such as real-time feedback after ambient sound enhancement). This can significantly reduce the latency between sound acquisition and processing by the mobile phone microphone and transmission to the hearing aid, avoiding unnatural auditory echoes for the user and improving the user experience.
[0072] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A low-latency bypass audio channel method for critical voice interaction, characterized in that, include: The application initiates an access request to the bypass audio channel; Perform security authentication and permission verification on the application; Once the application passes verification, its associated audio stream is marked as a critical audio stream; Activate the bypass channel controller and allocate audio hardware resources to the key audio stream based on a preset arbitration strategy; The audio hardware resources include at least analog-to-digital converters (ADCs) and / or digital-to-analog converters (DACs); The ADC / DAC is configured to establish a bypass audio channel that bypasses the standard audio stack of the operating system, allowing the audio data of the critical audio stream to be directly transmitted between the ADC / DAC and the audio processing unit through the bypass audio channel.
2. The low-latency bypass audio channel method for critical voice interaction according to claim 1, characterized in that, The security authentication and permission verification of the application includes: verifying the identity and / or signature of the application through a Trusted Execution Environment (TEE); and granting or denying access to the bypass audio channel by the permission manager based on the verification result.
3. The low-latency bypass audio channel method for critical voice interaction according to claim 1, characterized in that, The preset arbitration strategy includes a priority preemption strategy and / or a resource exclusive strategy; wherein, the priority preemption strategy is used to pause or suspend access to the audio hardware resources by a lower priority standard audio stream when the critical audio stream starts; the resource exclusive strategy is used to enable the critical audio stream to exclusively access the ADC / DAC during the activity of the critical audio stream.
4. The low-latency bypass audio channel method for critical voice interaction according to claim 1, characterized in that, The establishment of a bypass audio channel that bypasses the standard audio stack of the operating system includes: configuring a direct data path between the ADC / DAC and the voice acceleration processing unit; the voice acceleration processing unit includes a digital signal processor (DSP) or a neural network processor (NPU) for real-time processing of audio data transmitted through the bypass audio channel.
5. The low-latency bypass audio channel method for critical voice interaction according to any one of claims 1 to 4, characterized in that, Also includes: When the bypass audio channel fails to be established due to insufficient resources or hardware failure, a safe rollback operation is performed to forward the critical audio stream to the standard audio stack of the operating system for processing. When the critical audio stream finishes using the bypass audio channel, the occupied audio hardware resources are released, and the audio service of the operating system is notified to resume standard control of the ADC / DAC.
6. A low-latency bypass audio channel method for critical voice interaction, characterized in that, include: The application initiates an access request to the bypass audio channel; The application is subjected to security authentication and permission verification, including: verifying the identity and / or signature of the application through a Trusted Execution Environment (TEE), and the permission manager granting or denying access to the bypass audio channel based on the verification result; Once the application passes verification, its associated audio stream is marked as a critical audio stream; the bypass channel controller is activated, and audio hardware resources are allocated to the critical audio stream based on a preset arbitration strategy; the preset arbitration strategy includes a priority preemption strategy and / or a resource exclusive strategy; wherein, the priority preemption strategy is used to pause or suspend access to the audio hardware resources by lower-priority standard audio streams when the critical audio stream starts; the resource exclusive strategy is used to allow the critical audio stream to exclusively access the ADC / DAC during the activity of the critical audio stream; The audio hardware resources include at least analog-to-digital converters (ADCs) and / or digital-to-analog converters (DACs); The ADC / DAC is configured to establish a bypass audio channel that bypasses the standard audio stack of the operating system, allowing audio data of the critical audio stream to be directly transmitted between the ADC / DAC and the audio processing unit through the bypass audio channel. Establishing the bypass audio channel includes configuring a direct data path between the ADC / DAC and the voice acceleration processing unit. The voice acceleration processing unit includes a digital signal processor (DSP) or a neural network processor (NPU) for real-time processing of the audio data transmitted through the bypass audio channel. When the bypass audio channel fails to be established due to insufficient resources or hardware failure, a safe rollback operation is performed to forward the critical audio stream to the standard audio stack of the operating system for processing. When the critical audio stream finishes using the bypass audio channel, the occupied audio hardware resources are released, and the audio service of the operating system is notified to resume standard control of the ADC / DAC.
7. A low-latency bypass audio channel device for critical voice interaction, used to implement the low-latency bypass audio channel method for critical voice interaction as described in any one of claims 1-6, characterized in that, include: The permission management module is used to receive access requests from applications to the bypass audio channel and to perform security authentication and permission verification on the applications. The channel control module, connected to the permission management module, is used to mark the associated audio stream as a key audio stream after the application passes verification, and to allocate audio hardware resources to the key audio stream based on a preset arbitration strategy. The audio hardware resources include at least analog-to-digital converters (ADCs) and / or digital-to-analog converters (DACs). A hardware configuration module, connected to the channel control module, is used to configure the ADC / DAC to establish a bypass audio channel that bypasses the standard audio stack of the operating system. The bypass audio channel is used to enable the direct transmission of audio data of the key audio stream between the ADC / DAC and the audio processing unit.
8. The low-latency bypass audio channel device for critical voice interaction according to claim 7, characterized in that, The permission management module is integrated into or communicates with a Trusted Execution Environment (TEE) for performing security authentication of the application through the TEE.
9. The low-latency bypass audio channel device for critical voice interaction according to claim 7, characterized in that, It also includes a voice acceleration processing unit; the hardware configuration module is further used to configure a direct data path between the ADC / DAC and the voice acceleration processing unit; the voice acceleration processing unit includes a digital signal processor (DSP) or a neural network processor (NPU) for real-time processing of audio data transmitted through the bypass audio channel.
10. The low-latency bypass audio channel device for critical voice interaction according to claim 7, characterized in that, The channel control module is also used to: trigger a safety fallback mechanism when the bypass audio channel fails to be established, forwarding the critical audio stream to the standard audio stack of the operating system; and, when the critical audio stream ends, control the release of the audio hardware resources and notify the audio service of the operating system to resume standard control.