Voice communication detection method for virtual social space and related device

Abstract

The application discloses a voice communication detection method of a virtual social space and related equipment, which can obtain parameters of a target terminal, the target terminal including a terminal logged in by a target account joining the virtual social space; based on the parameters of the target terminal, a target simulation node is called, and the configuration parameters of the target simulation node are matched with the parameters of the target terminal; the target simulation node is accessed into the virtual social space, and voice communication data in the virtual social space is collected through the target simulation node; voice communication quality detection is performed based on the voice communication data collected by the target simulation node, and a voice communication quality detection result of the virtual social space is obtained, so that the voice communication data can be acquired by configuring the virtual node according to the parameters of the terminal, and the voice communication quality is detected by using the acquired voice communication data, so that the active detection of the voice communication quality can be realized, and the timeliness of the voice communication detection is improved.

Description

Technical Field

[0001] This application relates to the field of voice technology, and more specifically, to a voice communication detection method, electronic device, storage medium, and program product in a virtual social space. Background Technology

[0002] In related technologies, users can enter virtual social spaces for voice communication, such as virtual rooms. Issues during voice communication, such as network packet loss, latency jitter, link interruption, mute stuttering, audio distortion, and signaling anomalies, directly impact the user's call experience. Therefore, real-time and comprehensive anomaly detection of the RTC voice communication link is crucial to ensuring stable service operation.

[0003] However, the mainstream voice communication detection solutions in the industry are all passive detection solutions, which mainly rely on server log collection, user active feedback, and background abnormal error capture to achieve fault monitoring, resulting in a lag in voice communication detection. Summary of the Invention

[0004] In view of the above problems, this application proposes a voice communication detection method, electronic device, storage medium and program product in virtual social spaces to improve the timeliness of voice communication detection in virtual social spaces.

[0005] In a first aspect, embodiments of this application propose a method for voice communication detection in a virtual social space, comprising: acquiring parameters of a target terminal, the target terminal including the terminal logged in by a target account that has joined the virtual social space; based on the parameters of the target terminal, calling a target simulation node, the configuration parameters of the target simulation node matching the parameters of the target terminal; connecting the target simulation node to the virtual social space, and collecting voice communication data in the virtual social space through the target simulation node; and performing voice communication quality detection based on the voice communication data collected by the target simulation node to obtain the voice communication quality detection result of the virtual social space.

[0006] In one possible implementation, the method further includes: obtaining the voice communication type of the social virtual space; and invoking the target simulation node based on the parameters of the target terminal, including: invoking the target simulation node based on the parameters of the target terminal and the voice communication type.

[0007] In one possible implementation, the voice communication type includes private voice communication, multi-person voice communication, or cross-regional voice communication. Based on the parameters of the target terminal and the voice communication type, a target simulation node is invoked, including: invoking multiple sets of candidate simulation nodes based on the parameters of the target terminal, where the configuration parameters of each set of candidate simulation nodes match the parameters of the target terminal, and each set of candidate simulation nodes includes at least one node; if the voice communication type is private voice communication, then the first set of candidate simulation nodes in the multiple sets of candidate simulation nodes is used as the target simulation node, and the first set of candidate simulation nodes is configured to perform private voice communication detection; if the voice communication type is... For multi-person voice communication, the second group of candidate simulated nodes from multiple candidate simulated nodes is used as the target simulated node, and the second group of candidate simulated nodes is configured to detect multi-person voice communication. If the voice communication type is cross-regional voice communication, the third group of candidate simulated nodes from multiple candidate simulated nodes is used as the target simulated node, and the third group of candidate simulated nodes is configured to detect cross-regional voice communication. Among them, each candidate simulated node in the first group is configured for single-channel audio decoding, each candidate simulated node in the second group is configured for multi-channel concurrent audio decoding, and the third group of candidate simulated nodes is configured for packet loss-resistant decoding.

[0008] In one possible implementation, voice communication quality detection is performed based on voice communication data collected from the target simulated node to obtain the voice communication quality detection result of the virtual social space, including: performing voice communication quality detection based on voice communication data collected from the target simulated node to obtain a target voice communication quality level, wherein the target voice communication quality level is one of multiple preset voice communication quality levels; and using the target voice communication quality level as the voice communication quality detection result.

[0009] In one possible implementation, the voice communication data includes voice audio data. Voice communication quality detection is performed based on the voice communication data collected by the target simulated node to obtain the voice communication quality detection result of the virtual social space. This includes: performing voice blurring and desensitization processing on the voice audio data collected by the target simulated node to obtain desensitized voice audio data; and performing voice communication quality detection based on the desensitized voice audio data to obtain the voice communication quality detection result of the virtual social space.

[0010] In one possible implementation, the voice audio data collected by the target simulated node is subjected to voice blurring and desensitization processing, including: assessing the need for audio fault localization in the virtual social space; if audio fault localization is not required, then the voice audio data collected by the target simulated node is subjected to voice blurring and desensitization processing.

[0011] In one possible implementation, the voice communication data includes voice audio data and network indicator data. Voice communication quality detection is performed based on the voice communication data collected by the target simulated node to obtain the voice communication quality detection result of the virtual social space. This includes: performing voice communication quality detection based on the voice audio data and network indicator data collected by the target simulated node to obtain the voice communication quality detection result of the virtual social space.

[0012] Secondly, embodiments of this application propose a voice communication detection module for a virtual social space, comprising: an acquisition module for acquiring parameters of a target terminal, the target terminal including the terminal logged in by a target account that has joined the virtual social space; a detection module for calling a target simulation node based on the parameters of the target terminal, the configuration parameters of the target simulation node matching the parameters of the target terminal; connecting the target simulation node to the virtual social space and collecting voice communication data in the virtual social space through the target simulation node; and performing voice communication quality detection based on the voice communication data collected by the target simulation node to obtain the voice communication quality detection result of the virtual social space.

[0013] Thirdly, embodiments of this application provide an electronic device, including: a processor; a memory, on which computer-readable instructions are stored, and when the computer-readable instructions are executed by the processor, the above-mentioned voice communication detection method for virtual social spaces is implemented.

[0014] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-readable instructions, which, when executed by a processor, implement the above-described method for detecting voice communication in a virtual social space.

[0015] Fifthly, embodiments of this application provide a computer program product, including computer instructions, which, when executed by a processor, implement the above-described voice communication detection method for a virtual social space.

[0016] In this embodiment, parameters of the target terminal can be obtained, including the terminal logged in by the target account that has joined the virtual social space. Based on the parameters of the target terminal, a target simulation node is invoked, and the configuration parameters of the target simulation node match the parameters of the target terminal. The target simulation node is connected to the virtual social space, and voice communication data in the virtual social space is collected through the target simulation node. Voice communication quality detection is performed based on the voice communication data collected by the target simulation node to obtain the voice communication quality detection result of the virtual social space. In this way, the virtual node can be configured to acquire voice communication data according to the parameters of the terminal, and the acquired voice communication data can be used to detect the voice communication quality, thereby realizing active detection of voice communication quality and improving the timeliness of voice communication detection. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0018] Figure 1 This is a flowchart illustrating a voice communication detection method in a virtual social space according to an embodiment of this application.

[0019] Figure 2 This is a schematic diagram of a system framework according to an embodiment of this application.

[0020] Figure 3 This is a flowchart of a simulated user access and data collection process provided according to an embodiment of this application.

[0021] Figure 4 This is a logic block diagram for multi-dimensional anomaly detection and privacy protection provided according to an embodiment of this application.

[0022] Figure 5 This is a schematic diagram of the structure of a voice communication detection device for a virtual social space according to an embodiment of this application.

[0023] Figure 6 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown. Detailed Implementation

[0024] The embodiments of this application are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0025] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present application without creative effort are within the scope of protection of the present application.

[0026] In the following description, the terms "first," "second," etc., are used merely to distinguish similar objects and do not represent a specific ordering of the objects. It is understood that "first" and "second" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein. "A plurality of" means not less than two.

[0027] In this document, "multiple" refers to two or more. "And / or" describes the relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following associated objects are in an "or" relationship. In the following description, references to "some embodiments or some embodiment methods" describe a subset of all possible embodiments. However, it is understood that "some embodiments" can be the same subset or different subsets of all possible embodiments and can be combined with each other without conflict.

[0028] In practice, the collection and processing of data in this application should strictly comply with the requirements of relevant national laws and regulations, obtain the informed consent or separate consent of the data subject, and carry out subsequent data use and processing within the scope of laws and regulations and the authorization of the data subject.

[0029] Before providing a detailed description of the embodiments of this application, the nouns and terms involved in the embodiments of this application will be explained, and the nouns and terms involved in the embodiments of this application shall be interpreted as follows.

[0030] Real-time communication (RTC) technology is widely used in internet scenarios such as online voice calls, cloud conferencing, and real-time voice interaction. The voice communication in the virtual social space of this embodiment can be RTC communication.

[0031] Saving (Dumping) Audio Data: In computer terminology, "dump" means "to transfer, dump, or export." When a program crashes or encounters an exception, the system "dumps" the current memory state; this is called a memory dump. Similarly, in RTC calls, dumping audio means directly "dumping" (saving) the raw audio data packets received by the network transport layer or audio decoder to the local or server hard drive. In this embodiment, the audio data dumped can be raw data containing a large amount of low-level technical information: Raw waveform data: PCM audio stream before final rendering. Timestamp: The precise arrival time of each audio frame, used to analyze latency and stuttering. Network transmission characteristics: Traces of packet loss, out-of-order delivery, and jitter. Low-level processing traces: For example, intermediate state data after audio processing modules such as echo cancellation (AEC) and noise reduction (ANS).

[0032] Simulated User Nodes: These are not real terminal devices, but rather "probes" or "bots" dynamically built by the system backend to detect voice communications of terminals used by users in virtual social spaces. Simulated user nodes are independent and controllable detection units designed to replicate the network and device environment of real users. Terminal Attribute Simulation: Nodes can be configured to simulate real physical device parameters (such as specific mobile device models, PC models, operating system versions, etc.). Network Environment Simulation: Supports configuration of multiple network types (4G / 5G / WiFi) and can artificially inject specific weak network environments (such as high packet loss, high latency, network jitter, network disconnection and reconnection, etc.) to match complex online real-world business scenarios. Call Parameter Configuration: Each node can independently configure underlying parameters such as call duration and audio encoding format. To perform detection without affecting the real user experience, simulated user nodes adopt a "silent listening" mode: Downlink Reception: Downlink audio stream reception and decoding are enabled throughout, replicating the listening behavior of real users. Uplink Silence: Uplink voice acquisition and transmission functions are forcibly disabled, generating no uplink data stream. Zero Interference: This one-way monitoring mechanism ensures zero interference between the simulated nodes and the real room's call quality, bandwidth usage, and interaction status. During operation, the simulated user nodes capture and report three core categories of metrics in real time as the basis for anomaly detection: Network Transmission Metrics: Real-time monitoring and recording of underlying QoS data such as latency, packet loss rate, and network jitter. Communication Link Metrics: Recording signaling interaction status, reconnection count, and connection establishment time. Audio Quality Metrics: Analyzing audio distortion, stuttering frequency, and popping sounds, while simultaneously capturing the underlying raw dump audio data. To accurately cover different business scenarios, the simulated user nodes are grouped and connected to different test scenarios with differentiated configurations: Group 1 (One-on-One Private Calls): Configured for single-channel audio decoding. Focuses on verifying point-to-point communication connectivity, instant connection latency, and end-to-end encryption privacy security. Group 2 (Multi-Party Voice Conferencing): Configured for concurrent multi-channel audio decoding. The focus is on verifying the bandwidth load, mixed-stream computing power, and synchronization stability of room status signaling on the server-side SFU / MCU under high-concurrency scenarios. The third group (cross-regional node calls): configured for long-link weak network transmission and packet loss-resistant decoding. The focus is on verifying the intelligent routing and scheduling capabilities of the globally distributed network in cross-border / cross-regional long-link environments, the compensation and recovery effect of the packet loss-resistant algorithm, and the disaster recovery switching capabilities in the event of node failure.

[0033] The relevant technologies will be explained below.

[0034] In related technologies, users can enter virtual social spaces for voice communication, such as virtual rooms. Issues during voice communication, such as network packet loss, latency jitter, link interruption, mute stuttering, audio distortion, and signaling anomalies, directly impact the user's call experience. Therefore, real-time and comprehensive anomaly detection of the RTC voice communication link is crucial to ensuring stable service operation.

[0035] However, current mainstream voice communication detection solutions in the industry are all passive detection solutions, mainly relying on server-side log collection, user feedback, and backend error capture to achieve fault monitoring. Voice communication detection is lagging behind, and the specific problems are analyzed as follows: 1) Delayed detection and inability to provide early warning: Existing solutions can only detect problems after an anomaly occurs, a user complaint is received, or a system error occurs. They cannot achieve 24 / 7 proactive inspection. A large number of hidden and transient minor communication anomalies cannot be detected in advance, which can easily lead to a large number of user experience problems.

[0036] 2) The detection dimensions are one-sided and there are blind spots: Traditional detection relies only on the server-side backend data, which cannot cover key links such as the user terminal side, end-to-end transmission links, and local network environment. It cannot restore the real user call scenario. Problems such as occasional lag on the terminal side, abnormal local weak network adaptation, and abnormal audio decoding on the terminal side have long been in the detection blind spot.

[0037] 3) Difficult to reproduce and locate anomalies: RTC voice anomalies are often related to personalized factors such as user network environment, terminal equipment, call duration, and cross-regional transmission. Existing backend log data cannot simulate real user operation behavior and usage scenarios, making it difficult to reproduce faults. Maintenance personnel cannot accurately locate the root cause of anomalies, resulting in extremely low troubleshooting efficiency.

[0038] 4) High manual operation and maintenance costs and low detection efficiency: Existing solutions rely heavily on manual spot checks and user feedback reviews, which cannot achieve automated and routine full-call scenario inspections. The anomaly coverage in complex scenarios such as high concurrency, cross-regional, and long-term calls is extremely low.

[0039] 5) Risk of user privacy leakage: Existing RTC call quality testing solutions generally lack a robust privacy protection mechanism when collecting, storing, and exporting call dump audio data. The raw dump audio contains a large amount of users' real-time call privacy information. If it is stored in plaintext without access control or encryption protection, it is highly susceptible to security risks such as data leakage, unauthorized access, unauthorized export, and secondary misuse, failing to meet the compliance requirements for network data security and personal information protection.

[0040] In summary, existing RTC voice communication detection technologies suffer from shortcomings such as passive lag, incomplete coverage, difficulty in reproducing faults, low positioning accuracy, and lack of privacy data protection. There is an urgent need for a new solution that combines active, interference-free detection with audio data privacy and security protection.

[0041] In view of this, embodiments of this application propose a voice communication detection method, electronic device, storage medium, and program product for virtual social spaces. This method can acquire parameters of a target terminal, including the terminal logged into by a target account that has joined the virtual social space. Based on the parameters of the target terminal, a target simulation node is invoked, the configuration parameters of which match the parameters of the target simulation node. The target simulation node is connected to the virtual social space, and voice communication data in the virtual social space is collected through the target simulation node. Voice communication quality is detected based on the voice communication data collected by the target simulation node, resulting in a voice communication quality detection result for the virtual social space. This allows for the configuration of virtual nodes to acquire voice communication data based on the terminal parameters, and the use of the acquired voice communication data for voice communication quality detection. This enables proactive detection of voice communication quality, improving the timeliness of voice communication detection.

[0042] Please see Figure 1 , Figure 1 This is a flowchart illustrating a voice communication detection method in a virtual social space according to an embodiment of this application. The method can be executed by an electronic device, such as a server or terminal. The terminal can be a smartphone, tablet, laptop, desktop computer, smart TV, smart home device, in-vehicle terminal, etc., and is not specifically limited thereto. The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0043] like Figure 1 The methods shown may include: S110. Obtain parameters of the target terminal, including the terminal logged in by the target account that joined the virtual social space.

[0044] The virtual social space can refer to a digital environment built on internet technology that supports real-time interaction, content sharing, and social activities among multiple users with virtual identities. In this embodiment, the virtual social space may include, but is not limited to, a metaverse social platform, online voice chat rooms, voice channels within multiplayer online games, virtual meeting rooms, and online education live streaming rooms. The target terminal can refer to the physical hardware device or virtual operating environment used by the user to access the virtual social space and conduct real-time audio and video communication. In this embodiment, the virtual social space may include at least two target accounts, which can communicate via voice to generate audio data. The target account may be the account logged into by the user.

[0045] It should be understood that the virtual social space in this embodiment can conduct voice communication, but this does not mean that the virtual social space can only conduct voice communication. In some cases, the virtual social space can also conduct video communication, which is not limited here.

[0046] The parameters of the target terminal can refer to a set of quantified data or attributes that characterize the hardware status, software environment, network conditions, and audio processing capabilities at a specific moment. In this embodiment, the parameters of the target terminal may include, but are not limited to, at least one of the following: device static parameters, network parameters, or audio link parameters. Among them, device static parameters include, but are not limited to, device model, operating system version number, central processing unit (CPU) architecture, sound card driver version, and microphone / speaker hardware specifications. Network parameters may include, but are not limited to, the type of network currently accessed, such as 5G, WiFi, or 4G, Internet Protocol (IP) address, network signal strength, round-trip time (RTT), packet loss rate, and jitter value. Audio link parameters may include, but are not limited to, audio sampling rate (e.g., 44.1kHz / 48kHz), bit depth, number of channels, type of audio codec currently in use (Opus Audio Codec (Opus) / Advanced Audio Coding (AAC) / G.711 Audio Coding (G.711)), the on / off status and gain value of Acoustic Echo Cancellation (AEC) and Automatic Gain Control (AGC).

[0047] S120. Based on the parameters of the target terminal, call the target simulation node. The configuration parameters of the target simulation node match the parameters of the target terminal.

[0048] In this embodiment, the configuration parameters of the target simulation node match the parameters of the target terminal, which means that the configuration parameters of the target simulation node are identical to those of the target terminal. The configuration parameters of the target simulation node can be found in the description of the target terminal's parameters, and will not be repeated here.

[0049] S130. Connect the target simulation node to the virtual social space and collect voice communication data in the virtual social space through the target simulation node.

[0050] In this embodiment, when target accounts in the virtual social space make a call, voice communication data is generated, and the target virtual node that has accessed the virtual social space can collect the voice communication data in the virtual social space.

[0051] In this embodiment, a target simulation node can be invoked based on the parameters of the target terminal. The configuration parameters of the target simulation node match the parameters of the target terminal. In other words, when collecting voice communication data in the virtual social space through the target simulation node, it can be understood as acquiring voice communication data in an approximate state of the target terminal. This can help improve the reference value of the acquired voice communication data.

[0052] It should be noted that in this embodiment, the target simulated node is only responsible for downlink reception: downlink audio stream reception and decoding are enabled throughout the entire process, replicating the listening behavior of a real user. Uplink silence: uplink voice acquisition and transmission functions are forcibly disabled, and no uplink data stream is generated.

[0053] S140. Based on the voice communication data collected by the target simulation node, perform voice communication quality detection to obtain the voice communication quality detection results of the virtual social space.

[0054] The voice communication quality detection result is used to represent the quality of voice communication within the virtual social space. Optionally, the voice communication quality detection result can indicate whether the voice communication is good or a quality level; there is no limitation on this.

[0055] It should be noted that, in this embodiment, the voice communication quality of different accounts in the virtual social space may vary depending on the terminal and the environment. Therefore, in this embodiment, a target simulation node may be configured for at least one target account, that is, a target simulation node may be configured for some or all target accounts, without any limitation.

[0056] In one possible implementation, target emulation nodes can be configured for a subset of target accounts. In this case, target emulation nodes can be configured for target accounts that have enabled voice communication quality features, or for target accounts with the worst corresponding target terminal performance; there are no restrictions on this.

[0057] In another possible implementation, target simulation nodes can be configured for all target accounts. In this implementation, each target account joining the virtual social space can be configured with a target simulation node. For example, the target simulation node can be configured after the target account enters the virtual social space, or when voice data is detected. There are no restrictions on this.

[0058] Optionally, the target simulated node corresponding to the target account can be destroyed after the target account exits the virtual social space, or it can be retained for a certain period of time so that the target account can call the target simulated node to obtain voice communication data after re-entering the virtual social space. There are no restrictions on this.

[0059] It should be understood that by configuring a simulated node for each account, the voice communication quality for each account can be tested.

[0060] In another possible implementation, it could be achieved by detecting the mixed audio of the entire channel (i.e., a virtual social space). This mixed audio could be a combination of all audio within the virtual social space. In this implementation, because it detects the mixed audio of the entire channel, it allows for real-time monitoring of the voice communication of each target account joining the virtual social space, thus improving the comprehensiveness of audio detection.

[0061] In this embodiment, by acquiring the parameters of the target terminal, which includes the terminal logged in by the target account that has joined the virtual social space; based on the parameters of the target terminal, a target simulation node is invoked, the configuration parameters of which match the parameters of the target simulation node; the target simulation node is connected to the virtual social space, and voice communication data in the virtual social space is collected through the target simulation node; voice communication quality is detected based on the voice communication data collected by the target simulation node, and the voice communication quality detection result of the virtual social space is obtained. In this way, the virtual node can be configured to acquire voice communication data according to the parameters of the terminal, and the acquired voice communication data can be used to detect the voice communication quality, thereby realizing active detection of voice communication quality and improving the timeliness of voice communication detection.

[0062] In one possible implementation, the method further includes: obtaining the voice communication type of the social virtual space; and invoking the target simulation node based on the parameters of the target terminal, including: invoking the target simulation node based on the parameters of the target terminal and the voice communication type.

[0063] In this embodiment, the voice communication type may include, but is not limited to, private voice communication, multi-person voice communication, or cross-regional voice communication. Among these, Private voice communication typically refers to one-on-one real-time voice calls. It usually employs peer-to-peer (P2P) transmission or lightweight relays: for the shortest data link, direct P2P connections are typically prioritized to reduce latency; if network limitations exist, fast forwarding is achieved through a single-node server (Selective Forwarding Unit, SFU). End-to-End Encryption (E2EE) is used. Voice data is encrypted at the sending end and decrypted only at the receiving end. Even with server relay, the server cannot access the plaintext voice, preventing privacy leaks at the fundamental level. Extremely low latency is required: to ensure smooth and natural conversation, end-to-end latency must typically be controlled within 200ms-300ms, and the ability to connect within seconds is also required.

[0064] Multi-person voice communication includes multi-person voice conferencing, voice chat rooms, and game chat channels. Multi-person voice communication can employ server-side mixing and distribution. However, due to the large number of participants, P2P mesh connections can cause terminal computing power and bandwidth degradation. Therefore, a server-side architecture is necessary, where the server centrally receives the audio stream, performs multipoint control (MCU) or selective forwarding (SFU), and then distributes it to each terminal.

[0065] Cross-regional voice communication can refer to real-time voice calls across different cities or even continents (such as international conferences and overseas social matching). For example, using a globally distributed network (SD-RTN / MSDN): cross-border public network transmission is extremely unstable. This technology constructs a virtual "global high-speed network" by deploying massive edge access nodes and backbone leased lines globally, allowing users' voice data to bypass congested public networks and instead be transmitted across borders via internal leased lines.

[0066] In this embodiment, by obtaining the voice communication type of the social virtual space, and based on the parameters of the target terminal and the voice communication type, the target simulation node is invoked. This allows the target simulation node to be invoked by combining the voice communication type and the parameters of the target terminal, which can improve the accuracy of simulation node invocation and thus help improve the accuracy of voice communication quality detection.

[0067] In another possible implementation, it is also possible to eliminate the need to use voice communication to call simulated nodes, which can reduce the required computing resources.

[0068] In one possible implementation, the voice communication type includes private voice communication, multi-person voice communication, or cross-regional voice communication. Based on the parameters of the target terminal and the voice communication type, a target simulation node is invoked, including: invoking multiple sets of candidate simulation nodes based on the parameters of the target terminal, where the configuration parameters of each set of candidate simulation nodes match the parameters of the target terminal, and each set of candidate simulation nodes includes at least one node; if the voice communication type is private voice communication, then the first set of candidate simulation nodes in the multiple sets of candidate simulation nodes is used as the target simulation node, and the first set of candidate simulation nodes is configured to perform private voice communication detection; if the voice communication type is... For multi-person voice communication, the second group of candidate simulated nodes from multiple candidate simulated nodes is used as the target simulated node, and the second group of candidate simulated nodes is configured to detect multi-person voice communication. If the voice communication type is cross-regional voice communication, the third group of candidate simulated nodes from multiple candidate simulated nodes is used as the target simulated node, and the third group of candidate simulated nodes is configured to detect cross-regional voice communication. Among them, each candidate simulated node in the first group is configured for single-channel audio decoding, each candidate simulated node in the second group is configured for multi-channel concurrent audio decoding, and the third group of candidate simulated nodes is configured for packet loss-resistant decoding.

[0069] In private voice communication, due to the one-to-one nature of the calls and the short link, the requirements for real-time performance and privacy are extremely high. Therefore, the simulated nodes are configured with single-channel audio decoding. This means that during testing, the group of nodes establishes only one uplink receiving channel and one downlink transmitting channel, handling only the encoding and decoding of a single audio stream. Optionally, the success rate of point-to-point (P2P) direct connections, the handshake status of end-to-end encryption (E2EE), and the clarity and latency of the single audio stream can be verified. Since private calls do not involve mixing, the simulated nodes must simulate a realistic single-person listening state to avoid introducing unnecessary CPU load due to multi-channel decoding, thereby masking real device performance issues.

[0070] In multi-person voice communication, due to the many-to-many nature of the calls (such as voice chat rooms and conferences), server-side mixing or forwarding is involved, resulting in high concurrency pressure. Therefore, the simulated nodes are configured for multi-channel concurrent audio decoding. This means that the group of nodes is configured to simultaneously subscribe to and decode the audio streams of multiple speakers in the room (e.g., 4, 9, or even more streams). Optionally, this can be used to verify the server's (SFU / MCU) concurrent distribution capabilities, simulate the CPU / memory load of the terminal under multi-stream concurrency, and the synchronization and stuttering issues after multi-channel audio mixing. Since failures in multi-person scenarios often stem from insufficient terminal performance to support multi-channel concurrent decoding or insufficient server bandwidth, only by configuring "multi-channel concurrent decoding" can the failure scenarios of "stuttering when there are many people" or "intermittent sound" be realistically reproduced.

[0071] In cross-regional voice communication, due to the long-distance transmission across countries / provinces, physical latency is high, public network jitter is significant, and the risk of packet loss is high. Therefore, the simulated nodes are configured with packet loss-resistant decoding. That is, this group of nodes is typically deployed in different geographical locations (e.g., one in Beijing and one in New York), and enhanced weak network adversarial algorithms (such as FEC forward error correction and ARQ automatic retransmission) are enabled. Optionally, the effectiveness of global routing scheduling (SD-RTN), the recovery capability of packet loss-resistant algorithms at a specific packet loss rate (e.g., 20%), and jitter buffering strategies under long links can be verified. Since the core pain point of cross-regional communication is network transmission quality, this group of nodes focuses on testing whether "the sound can still be heard clearly under poor network conditions," emphasizing the detection of audio packet loss resistance and transmission link stability.

[0072] In this embodiment, different groups of simulated nodes are configured for different communication scenario types, and the different groups of simulated nodes have different focuses, which can improve the performance utilization of simulated nodes and reduce resource waste.

[0073] In one possible implementation, voice communication quality detection is performed based on voice communication data collected from the target simulated node to obtain the voice communication quality detection result of the virtual social space, including: performing voice communication quality detection based on voice communication data collected from the target simulated node to obtain a target voice communication quality level, wherein the target voice communication quality level is one of multiple preset voice communication quality levels; and using the target voice communication quality level as the voice communication quality detection result.

[0074] Multiple preset voice communication quality levels can be set as needed, such as minor abnormality, general abnormality, severe abnormality, etc., without any restrictions.

[0075] Optionally, a tiered anomaly detection threshold model can be pre-established, combining industry RTC voice communication quality standards to distinguish between three levels: minor anomalies, general anomalies, and severe anomalies. Real-time comparative analysis of the collected data is performed; when any indicator exceeds the preset threshold or exhibits abnormal behavior, an anomaly is determined in the corresponding communication link or audio scenario. The model also distinguishes between anomaly types, including network anomalies, link signaling anomalies, audio quality anomalies, and terminal compatibility anomalies.

[0076] In other words, the voice communication data collected in this embodiment includes, but is not limited to, the following data: Network transmission data: used to evaluate objective physical metrics such as end-to-end latency, network jitter, and packet loss.

[0077] Voice audio data: used to evaluate the clarity, distortion, stuttering frequency, and popping sounds of audio.

[0078] Communication link data: used to evaluate connection establishment time, number of reconnections, signaling interaction success rate, etc.

[0079] In this embodiment, by classifying the voice communication quality, it is possible to determine whether the voice communication quality is abnormal and the degree of abnormality, which helps to understand the voice communication quality more accurately.

[0080] In addition, for various anomalies identified, the system accurately attributes the causes by combining the terminal environment, network status, and silent listening call scenarios of simulated users, locating whether the anomaly occurs on the terminal side, transmission link side, server side, or node scheduling side; it automatically generates anomaly detection logs, recording the anomaly time, scenario, indicator data, anomaly type, and root cause; and it pushes early warning information in real time for serious anomalies, enabling early detection and early handling of faults.

[0081] Another possible implementation is to eliminate the need for differentiating voice communication quality levels, which would reduce the required computing resources.

[0082] In one possible implementation, the voice communication data includes voice audio data. Voice communication quality detection is performed based on the voice communication data collected by the target simulated node to obtain the voice communication quality detection result of the virtual social space. This includes: performing voice blurring and desensitization processing on the voice audio data collected by the target simulated node to obtain desensitized voice audio data; and performing voice communication quality detection based on the desensitized voice audio data to obtain the voice communication quality detection result of the virtual social space.

[0083] In this embodiment, to avoid the leakage of users' voiceprints and other privacy information, the voice audio data can be processed to blur and desensitize human voices.

[0084] Optionally, in this embodiment, after the simulated node collects the original dump audio data, it is not directly used for analysis or storage, but instead, human voice fuzzing and desensitization is triggered. During the human voice fuzzing and desensitization process, semantic destruction and physical feature preservation can be performed. Semantic destruction involves using frequency band attenuation, masking, or voice-changing algorithms to destructively process the human language features (voiceprint, intonation, vocabulary) in the audio. Physical feature preservation involves intentionally preserving the underlying "physical transmission characteristics" of the audio, such as: silence segments (representing disconnection), popping / mechanical sounds (representing decoding anomalies), and volume fluctuations (representing network jitter).

[0085] Then, during voice communication quality testing, continuity testing, purity testing, and synchronization testing can be performed. Continuity testing analyzes the duration of silence in the anonymized audio to determine if there is severe network lag or packet loss. Purity testing identifies abnormal mechanical sounds, echoes, or background noise in the anonymized audio to determine if the audio processing algorithm (such as AEC / ANS) is malfunctioning. Synchronization testing verifies whether the audio waveform is aligned with the underlying timestamp to determine if there is audio-visual asynchrony or jitter buffer overflow. In this embodiment, the evaluation of RTC communication quality can be performed without relying on semantic information.

[0086] In this embodiment, the voice audio data collected by the target simulated node is subjected to human voice blurring and desensitization processing to obtain desensitized voice audio data; based on the desensitized voice audio data, voice communication quality detection is performed to obtain the voice communication quality detection result of the virtual social space. In this way, while detecting the voice communication quality, the user's privacy and security can also be reduced.

[0087] In another possible implementation, it is not necessary to perform human voice blurring and desensitization processing on the voice audio data, which can improve the comprehensiveness of voice communication quality detection.

[0088] In one possible implementation, the voice audio data collected by the target simulated node is subjected to voice blurring and desensitization processing, including: assessing the need for audio fault localization in the virtual social space; if audio fault localization is not required, then the voice audio data collected by the target simulated node is subjected to voice blurring and desensitization processing.

[0089] In this embodiment, the audio fault localization requirement in virtual social spaces can be evaluated. Specifically, the audio fault localization requirement refers to whether, when the system detects an anomaly in a voice call, it is necessary to rely on specific voice content to determine the root cause of the fault. The following explains the scenarios where an audio fault localization requirement exists.

[0090] In cases where the fault manifests as a degradation in audio quality that cannot be directly explained by simple network metrics or signaling logs, the system requires audio fault localization. In such cases, the original dump audio must be preserved for analysis. In this embodiment, scenarios requiring audio fault localization include, but are not limited to, algorithm false positives / processing anomalies, acoustic echoes and feedback, device incompatibility and noise, and audio-visual asynchrony. Algorithm false positives / processing anomalies: For example, users report "my voice fluctuates in volume" or "the other person can't hear me." This may be due to a malfunction in Automatic Gain Control (AGC) or Noise Reduction (ANS) algorithms, which are cutting off normal human voices as noise. Listening to the original audio is necessary for confirmation. Acoustic echoes and feedback: Users report "hearing an echo in their own voice" or "a piercing feedback sound." This is usually due to a failure in the Echo Cancellation (AEC) algorithm, requiring analysis of the phase and delay in the audio waveform for localization. Device incompatibility and noise: Certain mobile phone models may have "electrical microphone" or "mechanical noise" during recording. This is a low-level hardware acquisition issue, requiring listening to the raw audio to confirm the frequency and characteristics of the noise. Audio and video are out of sync: the video and audio don't match, requiring analysis of the audio timestamps and waveform characteristics.

[0091] In cases where the fault manifests as a pure physical transmission interruption or system resource bottleneck, and objective network layer data (QoS) is sufficient to prove the cause of the fault, there is no need for audio fault localization. In this case, there is no need to listen to the specific content; simply perform human voice desensitization on the audio. Scenarios where audio fault localization is unnecessary include, but are not limited to, pure network packet loss / disconnection, server concurrency overload, terminal device hibernation / permission denial, and physical delays across long-distance cross-border links. Pure network packet loss / disconnection: The system detects a network packet loss rate as high as 30%, or the network is completely disconnected. The cause of the stuttering is 100% due to poor network conditions; there is no need to listen to audio for verification. Server concurrency overload: In a multi-person voice chat room, the server CPU is fully loaded or the bandwidth is saturated, causing everyone in the room to stutter simultaneously. This is a system resource issue, unrelated to the audio content. Terminal device hibernation / permission denial: The user's phone screen is locked, causing the RTC process to be suspended, or a system pop-up window denies microphone permissions. This is an operating system-level blockage; there is no need to analyze the audio. Cross-border long-link physical delay: Transoceanic calls will inevitably have a physical delay of more than 200ms, which is determined by the physical distance on Earth. There is no need to listen to the audio to confirm the delay.

[0092] In this embodiment, the audio fault location requirement of the virtual social space is evaluated. If audio fault location is not required, the voice audio data collected by the target simulated node is subjected to human voice blurring and desensitization processing. This avoids the situation where human voice blurring and desensitization processing is still performed when audio fault location is required, which would prevent normal voice communication quality detection and improve the accuracy of voice communication quality detection.

[0093] Optionally, in one possible implementation, if audio fault localization is required, speech detection processing can be performed based on the original speech audio data. Optionally, in this embodiment, the original speech audio data also needs to be retained for use when needed.

[0094] In one possible implementation, the voice communication data includes voice audio data and network indicator data. Voice communication quality detection is performed based on the voice communication data collected by the target simulated node to obtain the voice communication quality detection result of the virtual social space. This includes: performing voice communication quality detection based on the voice audio data and network indicator data collected by the target simulated node to obtain the voice communication quality detection result of the virtual social space.

[0095] In this embodiment, the speech audio data reflects the sound actually received by the user. When performing detection based on speech audio data, the audio waveform itself can be analyzed in depth, and digital signal processing technology can be used to evaluate the purity and integrity of the sound. Specifically, this embodiment can perform detection based on the original speech audio data (i.e., the actual acquired audio data) or based on speech audio data after human voice blurring processing; no limitation is made here.

[0096] In this embodiment, when performing detection based on network metric data, it can be based on objective quality assessment (QoS) of network metrics. This detection mainly focuses on the quality of the transport layer, determining whether the data arrives at its destination completely and on time, including but not limited to packet loss analysis, jitter analysis, and latency analysis.

[0097] Packet loss analysis includes random packet loss and burst packet loss. Random packet loss detects network congestion levels. High packet loss rates typically cause intermittent voice transmission. Burst packet loss detects link stability. Continuous packet loss is more disruptive to the listening experience than scattered packet loss, potentially causing prolonged silence or mechanical sounds.

[0098] Jitter analysis: Calculates the variance of data packet arrival times. Excessive jitter can cause frequent adjustments to the receiver's jitter buffer, leading to increased playback delays or the dropping of data packets to catch up on arrival times (exacerbating the perception of packet loss).

[0099] Latency analysis: Measuring end-to-end latency. Excessive latency can cause noticeable interruptions or pauses in conversations, severely impacting the user experience.

[0100] For ease of understanding, the following embodiments are described in conjunction with the above embodiments. Please refer to... Figure 2 , Figure 2 This is a schematic diagram of a system framework according to an embodiment of this application.

[0101] like Figure 2The framework shown may include a simulated user node configuration module 210, a silent listening and interference-free access module 220, a full-link data and dump audio acquisition module 230, a dump audio privacy and security protection module 240, a multi-dimensional anomaly judgment module 250, an anomaly attribution and early warning module 260, and an automated polling and inspection module 270; each module is connected in series to form a closed-loop process of "detection and acquisition + privacy protection + anomaly analysis".

[0102] Please see Figure 3 , Figure 3 This is a flowchart illustrating a simulated user access and data collection process according to an embodiment of this application. Figure 3 The process shown may include: S301. Initialize the simulated user cluster.

[0103] S302, Configure terminal / network / call parameters.

[0104] S303, Actively join RTC voice room.

[0105] S304. Establish an end-to-end communication link.

[0106] S305, Silent Listening, Downlink Audio Receive Only, Uplink Audio Disabled.

[0107] S306, Real-time acquisition of network / link / audio metrics + dumping audio data.

[0108] S307, Audio privacy encryption protection and storage.

[0109] S308, Synchronously push data to the anomaly analysis module.

[0110] When building a simulated user detection cluster, 10 simulated user nodes can be set up, with mobile and PC terminal attributes configured respectively, matching four network environments: 4G, 5G, public WiFi, and weak network speed limit, covering service nodes in multiple regions across the country; standard RTC voice call parameters are uniformly configured: audio encoding format is OPUS, sampling rate is 48kHz, and the normal call duration is 5-30 minutes, matching the parameters of mainstream online services.

[0111] Then, the simulated user nodes are grouped and connected to different business scenarios: the first group is connected to a one-to-one private voice call room, the second group is connected to a multi-person voice conference room, and the third group simulates cross-regional node calls; each group of simulated users remains in a silent listening state throughout the process, only continuously receiving and decoding the room's voice stream downlink, without sending any uplink voice data, replicating the entire usage behavior of real users listening to the call, without affecting the real users' calls in the room.

[0112] Furthermore, comprehensive data and audio dumps are collected. Throughout the call, key metrics are collected in real-time: at the network level, latency, jitter, uplink and downlink packet loss rates, and bandwidth fluctuations are collected; at the link level, signaling interaction success rate, link reconnection counts, and heartbeat timeout counts are collected; at the audio level, the frequency of stuttering, duration of abnormal silence, audio distortion, and the number of pops are collected, with data sampling completed every 100ms. Simultaneously, raw audio dump files are recorded and generated for subsequent analysis of audio quality anomalies.

[0113] In addition, privacy and security protection can be implemented for the dumped audio data. Comprehensive privacy protection is applied to the collected dumped audio data: (1) The file is encrypted and stored immediately after it is generated using the AES-256 algorithm, and the server does not retain plaintext audio; (2) Configure role permissions: Ordinary operation and maintenance personnel can only see the detection index data and have no audio access permissions. Only administrators can apply for temporary decryption permissions. (3) For scenarios where precise location of audio faults is not required, the dumped audio is automatically desensitized by blurring human voices; (4) Set up a 7-day automatic expiration and destruction mechanism to automatically clean up redundant audio data after the review is completed; (5) All audio decryption, viewing and download operations are recorded in audit logs and can be traced and verified.

[0114] Then, a graded anomaly detection and handling process is performed. Preset industry-adaptive thresholds are used: packet loss rate >3% is considered a minor anomaly, and packet loss rate >8% is considered a severe anomaly; latency jitter >50ms is considered a minor fluctuation, and >100ms is considered a communication anomaly; more than 3 stutters per minute are considered audio quality anomalies. The system compares sampled data with these thresholds in real time, automatically identifying and classifying various anomalies.

[0115] In addition, anomaly attribution and early warning output are performed. If high-frequency packet loss or jitter is detected in a weak network environment, it is attributed to network transmission anomalies; if frequent audio decoding stuttering is detected on fixed terminals, it is attributed to terminal adaptation anomalies; if signaling interaction timeouts are detected, it is attributed to service link anomalies. For severe anomalies, real-time SMS and platform pop-up warnings are sent, and a complete anomaly report is generated.

[0116] In addition, routine and cyclical inspections are required. The simulated user nodes adopt a polling mechanism, continuously switching between call scenarios, network environments, and access rooms to achieve 24 / 7 automated inspections, continuously monitoring the overall stability of online RTC voice communication, and promptly identifying various hidden faults.

[0117] Please see Figure 4 , Figure 4 This is a logic block diagram for multi-dimensional anomaly detection and privacy protection provided according to an embodiment of this application.

[0118] In this embodiment, the input consists of real-time collected full-dimensional detection data and dumped audio data, which are processed in parallel on two paths: one path is the anomaly detection branch 410, which completes anomaly classification and root cause attribution through network, link, and audio indicator verification; the other path is the privacy protection branch 420, which performs audio encryption, access control, de-identification processing, log auditing, and expiration destruction; and finally outputs anomaly warning results and secure and compliant audio storage data.

[0119] In summary, this embodiment addresses the shortcomings of existing RTC real-time voice communication anomaly detection solutions, such as passive lag, limited detection dimensions, blind spots in edge detection, difficulty in reproducing anomalies, low fault location accuracy, and lack of privacy protection for dumped audio data, making them susceptible to leakage. This embodiment provides a method for RTC real-time voice communication anomaly detection based on simulated users. This embodiment achieves interference-free active detection by constructing a silent listening simulated user node, while implementing full lifecycle security protection for the collected call dumped audio data. It balances accurate RTC call quality detection with user privacy protection, improving fault detection capabilities while meeting data security compliance requirements.

[0120] 2 Core Technology Solution This embodiment discloses a method for detecting anomalies in real-time voice communication (RTC) based on simulated users. Specifically, it involves building simulated user nodes that can simulate real terminal attributes, network environment, and user operation behavior. Simulated users actively access the RTC voice room and participate in real-time voice calls in a silent, listen-only mode, without generating uplink voice data to interfere with real users. The method fully replicates the entire process of user-side voice reception and decoding, simultaneously capturing and storing the original audio data dumped from the call and end-to-end communication data. Furthermore, it designs a comprehensive privacy protection mechanism for the dumped audio data, including encrypted storage, tiered access permissions, de-identification processing, and operation auditing. This achieves accurate detection of RTC voice anomalies while preventing the leakage of user privacy information. The specific steps are as follows: 1) Simulating user node construction and environment configuration: A multi-terminal, multi-network compatible simulated user cluster is pre-built. These simulated users replicate the terminal device parameters, system version, network type (4G / 5G / WiFi / weak network), and geographical location attributes of real users. Custom call parameters are also configured, including call duration, audio encoding format, sampling rate, and bandwidth threshold, perfectly matching real online RTC voice call scenarios. Each simulated user is an independently controllable detection node, capable of accessing different voice rooms and matching different call scenarios.

[0121] 2) Uninterrupted silent listening to call scenarios: The system controls simulated user nodes to initiate join requests and access the target RTC voice call room, establishing a complete real-time voice communication link with real users and other communication nodes in the room. The simulated user operates in silent listening mode throughout the process, only enabling downlink audio reception and decoding functions while disabling uplink voice acquisition and transmission functions. It only replicates the listening behavior of real users without generating any uplink voice data, thus avoiding sound interference and business impact on real calls and real users in the room, and fully restoring the entire process of real users receiving voice.

[0122] 3) End-to-end data and audio dump acquisition: During the simulated silent listening process, real-time end-to-end multi-dimensional detection data is collected, covering three main categories: network transmission metrics, communication link metrics, and audio quality metrics. Simultaneously, the original audio file of the call is captured and generated, completely preserving the original audio data for subsequent anomaly review, fault tracing, and audio quality problem analysis. Network transmission metrics include round-trip latency, latency jitter, uplink / downlink packet loss rate, and bandwidth fluctuation; communication link metrics include signaling interaction status, link connection duration, number of disconnections and reconnections, and heartbeat response status; audio quality metrics include audio distortion, silence duration, number of stutters, popping noises, and number of decoding anomalies.

[0123] 4) Comprehensive privacy protection processing for dumped audio data: For the collected dumped audio data containing user privacy, full lifecycle security protection is implemented, including multiple mechanisms such as encrypted storage, hierarchical access control, data anonymization, and operation log auditing to prevent the leakage of personal privacy: Audio encryption storage: All original dump audio files are encrypted using the AES symmetric encryption algorithm before being stored in the database and file server. Plaintext storage of original call audio is prohibited to prevent leakage of local and server static data. Hierarchical access control: Establish a multi-role permission system to distinguish between ordinary operation and maintenance, advanced operation and maintenance, and administrator roles. Ordinary personnel can only view abnormal indicator data and do not have permission to play, download, or export dump audio. Only authorized administrators can temporarily decrypt and view audio for fault review, thus implementing the principle of least privilege. Audio desensitization and time-limited destruction: For non-essential review scenarios, dump audio is automatically muted and desensitized, and key segments are blurred; at the same time, an automatic data expiration policy is configured to automatically clean up expired dump audio data after the review is completed, avoiding the risk of leakage caused by long-term data retention. Full-process operation audit: The entire process of viewing, decrypting, downloading, deleting, and exporting dumped audio is logged, including the operator, operation time, operation content, and IP address, so as to make privacy data traceable and auditable.

[0124] 5) Multi-dimensional intelligent anomaly detection: A pre-established tiered anomaly detection threshold model, combined with industry RTC voice communication quality standards, distinguishes between three levels of anomalies: minor, general, and severe. Real-time comparative analysis is performed on the collected data; when any indicator exceeds the preset threshold or exhibits abnormal behavior, an anomaly is determined in the corresponding communication link or audio scenario. The model also differentiates anomaly types, including network anomalies, link signaling anomalies, audio quality anomalies, and terminal compatibility anomalies.

[0125] 6) Anomaly attribution, recording, and early warning push: For various anomalies identified, the system accurately attributes the causes by combining the simulated user's terminal environment, network status, and silent eavesdropping call scenarios, pinpointing whether the anomaly occurred on the terminal side, transmission link side, server side, or node scheduling side; it automatically generates anomaly detection logs, recording the anomaly time, scenario, indicator data, anomaly type, and root cause; and it pushes early warning information in real time for serious anomalies, enabling early detection and early handling of faults.

[0126] 7) Automated cyclic inspection and detection: By polling multiple simulated user nodes to access RTC voice calls in different regions, rooms, and network scenarios, it achieves 24 / 7 uninterrupted automated inspection, covering various complex business scenarios such as high-concurrency calls, weak network calls, cross-regional calls, and long-term continuous calls, and realizes full-coverage detection of anomalies in all scenarios.

[0127] The technical effects of this embodiment may include, but are not limited to, the following: 1) Achieve interference-free proactive pre-detection: This embodiment enters the room by simulating a silent listening mode for users, receiving only downlink audio and not transmitting uplink sound, thus preventing interference with real user calls. At the same time, it proactively detects hidden and transient communication anomalies, enabling early warning of faults and significantly reducing the failure rate of user experience.

[0128] 2) Filling in blind spots in end-side detection and achieving more comprehensive detection coverage: By simulating the silent listening operation of a real user on the terminal side, it covers scenarios such as terminal adaptation, local decoding, and end-side network fluctuations that cannot be monitored by traditional server logs, achieving end-to-end full-link detection without blind spots and significantly improving the anomaly detection coverage.

[0129] 3) Faults are reproducible and location accuracy is greatly improved: Simulated users can reproduce various call scenarios and network environments. Maintenance personnel can rely on simulated detection scenarios to restore the fault process, accurately distinguish the level of anomaly occurrence, and solve the pain points of traditional solutions where faults are difficult to reproduce and root causes are difficult to locate.

[0130] 4) New dump audio privacy and security protection mechanism: For user call dump audio data collected during detection, multiple protections are implemented, including encrypted storage, hierarchical permissions, de-identification processing, automatic destruction, and operation auditing. This completely solves the problem of user privacy leakage caused by plaintext storage and unauthorized access of audio data in traditional detection solutions, and fully meets the requirements for personal information protection and network data security compliance.

[0131] 5) Balancing security and detection for enhanced practicality: While maintaining the accuracy of RTC call anomaly detection and fault recovery capabilities, it maximizes the protection of user call privacy, balancing business operation and maintenance needs with user privacy and security, and is suitable for large-scale commercial RTC voice service deployment.

[0132] 6) Achieve fully automated inspection and reduce operation and maintenance costs: 24 / 7 uninterrupted automated inspection can be achieved without manual intervention, reducing the workload of manual spot checks and fault analysis, and significantly lowering the operation and maintenance costs of RTC voice services. The following describes an apparatus embodiment of this application, which can be used to perform the methods described in the above embodiments of this application. For details not disclosed in the apparatus embodiments of this application, please refer to the method embodiments described in the above embodiments of this application.

[0133] Figure 5 This is a block diagram of a voice communication detection device for a virtual social space according to an embodiment of this application, such as... Figure 5 As shown, the voice communication detection device for this virtual social space is applied to an electronic device. The device may include an acquisition module 510 and a detection module 520, wherein: The acquisition module 510 is used to acquire parameters of the target terminal, including the terminal logged in by the target account that has joined the virtual social space; the detection module 520 is used to call the target simulation node based on the parameters of the target terminal, the configuration parameters of the target simulation node matching the parameters of the target terminal; connect the target simulation node to the virtual social space, and collect voice communication data in the virtual social space through the target simulation node; perform voice communication quality detection based on the voice communication data collected by the target simulation node, and obtain the voice communication quality detection result of the virtual social space.

[0134] In one possible implementation, the detection module 520 is further configured to: obtain the voice communication type of the social virtual space; and invoke the target simulation node based on the parameters of the target terminal, including: invoking the target simulation node based on the parameters of the target terminal and the voice communication type.

[0135] In one possible implementation, the voice communication type includes private voice communication, multi-person voice communication, or cross-regional voice communication. When the detection module 520 calls the target simulation node based on the parameters of the target terminal and the voice communication type, it is used to: call multiple sets of candidate simulation nodes based on the parameters of the target terminal, where the configuration parameters of each set of candidate simulation nodes match the parameters of the target terminal, and each set of candidate simulation nodes includes at least one node; if the voice communication type is private voice communication, then the first set of candidate simulation nodes in the multiple sets of candidate simulation nodes is used as the target simulation node, and the first set of candidate simulation nodes is configured to detect private voice communication; if the voice communication type is private voice communication... If the communication type is multi-person voice communication, the second group of candidate simulation nodes from multiple candidate simulation nodes will be used as the target simulation node, and the second group of candidate simulation nodes will be configured to detect multi-person voice communication. If the voice communication type is cross-regional voice communication, the third group of candidate simulation nodes from multiple candidate simulation nodes will be used as the target simulation node, and the third group of candidate simulation nodes will be configured to detect cross-regional voice communication. Among them, each candidate simulation node in the first group of candidate simulation nodes is configured for single-channel audio decoding, each candidate simulation node in the second group of candidate simulation nodes is configured for multi-channel concurrent audio decoding, and the third group of candidate simulation nodes is configured for packet loss-resistant decoding.

[0136] In one possible implementation, when the detection module 520 performs voice communication quality detection based on the voice communication data collected by the target simulated node and obtains the voice communication quality detection result of the virtual social space, it is used to: perform voice communication quality detection based on the voice communication data collected by the target simulated node to obtain the target voice communication quality level, wherein the target voice communication quality level is one of multiple preset voice communication quality levels; and use the target voice communication quality level as the voice communication quality detection result.

[0137] In one possible implementation, the voice communication data includes voice audio data. When the detection module 520 performs voice communication quality detection based on the voice communication data collected by the target simulated node and obtains the voice communication quality detection result of the virtual social space, it is used to: perform human voice blurring and desensitization processing on the voice audio data collected by the target simulated node to obtain desensitized voice audio data; and perform voice communication quality detection based on the desensitized voice audio data to obtain the voice communication quality detection result of the virtual social space.

[0138] In one possible implementation, when the detection module 520 performs voice blurring and desensitization processing on the voice audio data collected by the target simulated node, it is used to: assess the audio fault location requirements of the virtual social space; if audio fault location is not required, then the voice audio data collected by the target simulated node is subjected to voice blurring and desensitization processing.

[0139] In one possible implementation, the voice communication data includes voice audio data and network indicator data. When the detection module 520 performs voice communication quality detection based on the voice communication data collected by the target simulated node and obtains the voice communication quality detection result of the virtual social space, it is used to: perform voice communication quality detection based on the voice audio data and network indicator data collected by the target simulated node and obtain the voice communication quality detection result of the virtual social space.

[0140] The apparatus in this embodiment can be referred to the description of the method embodiment above, and will not be repeated here.

[0141] Figure 6 A schematic diagram of a computer system suitable for implementing the embodiments of this application is shown. It should be noted that... Figure 6 The computer system 600 of the illustrated electronic device is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this application. This electronic device can be used to execute the voice communication detection method for the virtual social space provided in the application.

[0142] like Figure 6 As shown, the computer system 600 includes a Central Processing Unit (CPU) 601, which can perform various appropriate actions and processes, such as executing the methods described in the above embodiments, based on programs stored in Read-Only Memory (ROM) 602 or programs loaded from storage portion 608 into Random Access Memory (RAM) 603. The RAM 603 also stores various programs and data required for system operation. The CPU 601, ROM 602, and RAM 603 are interconnected via a bus 604. An Input / Output (I / O) interface 605 is also connected to the bus 604.

[0143] The following components are connected to I / O interface 605: an input section 606 including a keyboard, mouse, etc.; an output section 607 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to I / O interface 605 as needed. A removable medium 611, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 610 as needed so that computer programs read from it can be installed into storage section 608 as needed.

[0144] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 609, and / or installed from removable medium 611. When the computer program is executed by central processing unit (CPU) 601, it performs various functions defined in the system of this application.

[0145] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such transmitted data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0146] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0147] The units described in the embodiments of this application can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0148] In another aspect, this application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable storage medium carries computer-readable instructions that, when executed by a processor, implement the methods in any of the above embodiments.

[0149] In the embodiments of this application, the terms "module" or "unit" refer to a computer program or part of a computer program that has a predetermined function and works with other related parts to achieve a predetermined goal. It can be implemented wholly or partially using software, hardware (e.g., processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that functions as a whole.

[0150] According to one aspect of the embodiments of this application, a computer program product is provided, the computer program product including computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the electronic device to perform the methods of any of the above embodiments.

[0151] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0152] Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, electronic device, touch terminal, or network device, etc.) to execute the method according to the embodiments of this application.

[0153] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.

[0154] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for detecting voice communication in a virtual social space, characterized in that, include: Obtain parameters of the target terminal, wherein the target terminal includes the terminal logged into by the target account that has joined the virtual social space; Based on the parameters of the target terminal, a target simulation node is invoked, wherein the configuration parameters of the target simulation node are matched with the parameters of the target terminal; The target simulation node is connected to the virtual social space, and voice communication data in the virtual social space is collected through the target simulation node. Voice communication quality is detected based on the voice communication data collected by the target simulation node, and the voice communication quality detection result of the virtual social space is obtained.

2. The method according to claim 1, characterized in that, The method further includes: Obtain the voice communication type of the social virtual space; The step of calling the target simulation node based on the parameters of the target terminal includes: Based on the parameters of the target terminal and the voice communication type, the target simulation node is invoked.

3. The method according to claim 2, characterized in that, The voice communication types include private voice communication, multi-person voice communication, or cross-regional voice communication. The step of invoking the target simulated node based on the parameters of the target terminal and the voice communication type includes: Based on the parameters of the target terminal, multiple sets of candidate simulation nodes are invoked. The configuration parameters of each set of candidate simulation nodes are matched with the parameters of the target terminal. Each set of candidate simulation nodes includes at least one node. If the voice communication type is private voice communication, then the first group of candidate simulation nodes in the multiple groups of candidate simulation nodes is taken as the target simulation node, and the first group of candidate simulation nodes is configured to detect private voice communication. If the voice communication type is multi-person voice communication, then the second group of candidate simulation nodes in the multiple groups of candidate simulation nodes is taken as the target simulation node, and the second group of candidate simulation nodes is configured to perform multi-person voice communication detection. If the voice communication type is cross-regional voice communication, then the third group of candidate simulation nodes among the multiple groups of candidate simulation nodes is taken as the target simulation node, and the third group of candidate simulation nodes is configured to detect cross-regional voice communication. In this configuration, each candidate simulation node in the first group of candidate simulation nodes is configured for single-channel audio decoding, each candidate simulation node in the second group of candidate simulation nodes is configured for multi-channel concurrent audio decoding, and the third group of candidate simulation nodes is configured for packet loss resistant decoding.

4. The method according to claim 1, characterized in that, The process of performing voice communication quality detection based on the voice communication data collected from the target simulated node to obtain the voice communication quality detection result of the virtual social space includes: Based on the voice communication data collected by the target simulation node, voice communication quality is detected to obtain a target voice communication quality level, which is one of a plurality of preset voice communication quality levels; The target voice communication quality level is used as the voice communication quality detection result.

5. The method according to claim 1, characterized in that, The voice communication data includes voice audio data. The voice communication quality detection based on the voice communication data collected by the target simulated node, to obtain the voice communication quality detection result of the virtual social space, includes: The voice audio data collected by the target simulation node is subjected to human voice blurring and desensitization processing to obtain the desensitized voice audio data; Voice communication quality detection is performed on the anonymized audio data to obtain the voice communication quality detection result of the virtual social space.

6. The method according to claim 5, characterized in that, The process of desensitizing and blurring human voices in the voice audio data collected from the target simulation node includes: Assess the audio fault localization requirements of the virtual social space; If audio fault location is not required, the voice audio data collected by the target simulation node is subjected to human voice blurring and desensitization processing.

7. The method according to any one of claims 1-6, characterized in that, The voice communication data includes voice audio data and network indicator data. The voice communication quality detection based on the voice communication data collected by the target simulated node, to obtain the voice communication quality detection result of the virtual social space, includes: Based on the voice audio data and network indicator data collected by the target simulation node, voice communication quality detection is performed to obtain the voice communication quality detection result of the virtual social space.

8. An electronic device, characterized in that, include: processor; A memory storing computer-readable instructions that, when executed by the processor, implement the method as described in any one of claims 1-7.

9. A computer-readable storage medium storing computer-readable instructions thereon, characterized in that, When the computer-readable instructions are executed by a processor, the method as described in any one of claims 1-7 is implemented.

10. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the method of any one of claims 1-7.

Images