CN101155140A - Method, device and system for audio stream error concealment - Google Patents

Abstract

The invention discloses a method for audio stream error concealment, the method comprising: a. classifying the transmitted audio frame according to the content to obtain the type information of the audio frame; b. combining the type information of the audio frame with the audio frame The encoding result is packaged and sent out; c. When frame loss occurs, for the lost audio frame, determine the type information of the audio frame obtained when it is classified according to the content; d. According to the type information of the lost audio frame, use the corresponding The error recovery strategy for audio frame reconstruction. The error concealment method of the present invention makes the reconstruction of the lost frame more pertinent, and can adaptively reconstruct the audio frame to achieve a better compensation effect. The invention also discloses a sending and receiving method for audio stream error concealment. The invention also discloses a transmitter, a receiver and a system for audio stream error concealment.

Description

Method, device and system for audio stream error concealment

technical field

The invention relates to real-time audio communication technology, in particular to a method, device and system for audio stream error concealment.

Background technique

People have been engaged in audio classification research for a long time, but for different application scenarios, the methods and classification results of audio classification are not the same. For example, 1. In the noise suppression of advanced audio equipment, frequency modulation analysis or Bayesian classifiers are commonly used to classify audio signals. 2. In order to better index and retrieve audio resources on the Internet, researches on content-based audio classification and retrieval have been carried out. In the more representative content-based audio classification work, the distinguishing features of audio are analyzed in detail, including loudness, pitch, harmonicity, etc., and an audio classifier is designed. 3. Another application of audio classification is to serve audio—especially the voice activation detector (VAD) of the speech coder. The purpose is to detect whether there is voice in voice communication. Coding method to achieve the effect of saving voice channel resources without reducing the voice quality.

In a real-time audio transmission system (such as VoIP), the main reasons for the degradation of sound quality caused by network transmission are delay, static interference and packet loss, among which packet loss is the most important reason for the degradation of sound quality. There is always some interference in the actual communication network, so there is a certain probability of packet loss. Although the lower layer of the protocol has error correction codes, it can only solve the error code in the packet, but cannot solve the problem of packet loss. In addition, since real-time audio services have strict delay constraints, packets delayed beyond a limit will also be considered lost. In order to ensure a certain communication quality under a certain packet loss rate, many codec algorithms integrate error concealment technology to eliminate the impact of packet loss.

Figure 1 is a block diagram of audio stream error concealment. As shown in Figure 1, after the compressed audio signal is transmitted through the IP network or other unreliable networks, the received audio data packets are usually stored in the jitter buffer, which is used to solve the reordering of late packets and early packets, etc. function, and then perform packet loss and error packet detection. If there is a packet loss or error packet, the system will start error concealment to compensate for packet loss, otherwise it will receive the audio packet decoding output correctly.

The packet loss recovery technology in real-time audio transmission can be divided into two categories according to the processing stage: repair based on the sender and repair based on the receiver.

●Sender-based error concealment

Packet loss recovery based on the sender is initiated by the sender and requires the cooperation of the sender and the receiver. Common methods include increasing redundancy, forward error correction, priority setting, and classification processing.

1. Increased redundancy: Increased data redundancy can improve the fault tolerance of the system, but at the same time increase the bandwidth.

2. Forward Error Correction (FEC): This strategy also relies on adding information that can repair lost packets in the transport stream, using blocks or algebraic codes to generate additional transport packets that can assist error correction, which also requires increased bandwidth.

3. Priority setting method: This technology requires network support and transmits packets according to priority, otherwise it cannot be realized, and it can only improve the packet loss probability caused by network congestion.

4. Classification processing method: In speech coding, in order to better adopt waveform replacement technology at the receiving end, the sending end can classify and process the speech signal according to the characteristics of the speech signal. For example, 3GPP2VMR-WB and ITU-T G.729.1 further describe the speech frame It is voiced, unvoiced, voiced transition, unvoiced transition, onset and other types, and after the decoder receives it, it can use the voice frame type of the previous frame and the next frame to infer the type of frame loss, and the decoder gets the frame loss type After that, the information of the lost frame can be better recovered.

● Receiver-based error concealment

The error concealment technology at the receiving end that does not require the participation of the sending end is essentially to estimate the lost data through a series of methods for the received data, and optimize it according to human physiological characteristics. It is basically a passive repair, usually Relatively easy to implement and does not increase bandwidth requirements. Receiver-based error concealment methods can be divided into three categories:

1. Insertion-based strategies: These techniques include methods such as splicing, mute and noise replacement. The splicing technique messes up the timing of the media stream and doesn't work well. The scope of application of silence replacement (filling the frame loss position with silence frames) is very limited. When the loss frequency of data packets is very low (less than 2%) and the gap width is less than 4ms, this method is more effective; when the gap width reaches 40ms, The effect would be unacceptable. Compared with silence replacement, noise replacement (filling the lost frame position with noise frames) can bring people a better subjective auditory experience, and can improve the distinguishability of speech signals at the same time. When background noise is used instead of silence, the human brain can subconsciously fill in the missing parts of the speech signal with the correct sound. The insertion method has nothing to do with the speech coding, nor the coding of the packet, but only processes the lost speech after decoding.

2. Interpolation-based strategy: Compared with interpolation technology, interpolation technology makes the processed sound bring relatively better subjective experience to people.

3. Regeneration-based strategy: extract the decoding state from the information around the lost packet, and generate a replacement packet for the lost packet. The implementation process of this method is more complicated, but it will achieve better results.

Generally speaking, the error concealment based on the sender will increase the network bandwidth and computational complexity, and the effect is better than that based on the receiver, but if the error concealment of the sender is independent of the receiver, that is, it has nothing to do with the media content, then it will not be able to do so according to all The characteristics of the frame loss adopt the corresponding error concealment strategy (for example, the stable speech frame is very similar to the previous frame, and the frame copy strategy can achieve a good hiding effect, and the transition frame needs to consider the state of the previous frame to determine the hiding strategy) . The technology at the receiving end is simple and can achieve certain concealment effects. However, if the concealment strategy has nothing to do with audio coding, that is, the content characteristics of the current lost frame and surrounding audio frames are not analyzed, and a targeted error concealment strategy can be adopted. Strategies for error concealment will be very limited.

At present, there are more and more audio coding standards that consider error concealment requirements at both the encoding end and the receiving end. There are mainly two representative methods:

(1) The encoder analyzes the characteristics of the audio frame before formal encoding, and adopts different encoding methods for audio frames with different characteristics. For example, AMR-WB+ adopts ACELP and TCX encoding to the signal frame according to the content of the audio frame, forming 26 kinds of superframe (every four frames form a superframe) coding mode. Coding mode information is used for error concealment. In the case of a frame loss, the receiving end infers or estimates the coding mode of the superframe according to the coding types of the remaining 3 frames of the superframe, so as to achieve a certain error concealment function.

(2) According to the pitch, spectrum and other characteristics of the speech frame, the speech frame is divided into types such as voiced, unvoiced, voiced transition, unvoiced transition, and onset. For example, in the latest ITU-TG.729.1 protocol, the encoder divides the speech frame into voiced, unvoiced, voiced transition, unvoiced transition, onset according to the frame content and its characteristics (VMR-WB also divides the speech frame into these 5 class), and 2 bits are used to mark its type in layer 2. In order to further enhance the accuracy of error concealment, G.729.1 also calculates the phase and energy of the frame, and transmits them in layer 3 and layer 4 of the next frame respectively. The decoder will try to recover the mode identification of the lost frame from the known category identification (including the category identification of the previous frame), so as to reconstruct the audio waveform according to the category mode of the lost frame by combining its phase and energy information.

For the first type of method: Although AMR-WB+ uses ACELP, TCX256, TCX512 and TCX1024 four modes of encoding according to the characteristics of the audio frame, 2, 2, 4, and 8 bits are used to represent the encoding mode information, and the encoding mode is used for error concealment. Mode information, infer or estimate the encoding mode of the superframe (composed of 4 frames and 1024 sampling points), so as to achieve a certain error concealment function, but only the encoding mode of the audio encoding is marked, and it cannot be based on the content of the audio frame Strategies are employed for frame-drop reconstruction, so efficient error concealment cannot be achieved.

For the second category of methods: Currently this type of error concealment technique is designed for speech frames and does not work well for other types of audio frames. For the classification and detection of music and natural sounds, especially in the case of packet loss, how to reconstruct the packet loss information so that audio communication can tolerate a high packet loss rate, there is currently no effective method.

To sum up, the current audio stream error concealment technology cannot realize efficient error concealment of audio frames, so that the requirement for packet loss rate in the audio communication process cannot be lowered.

Contents of the invention

In view of this, the embodiments of the present invention provide a method for audio stream error concealment, which can realize efficient error concealment for audio streams.

The embodiment of the present invention also provides an audio stream error concealment device and system, and the application of the device and system can realize efficient error concealment of the audio stream.

In order to realize the above-mentioned purpose of the invention, adopt following technical scheme:

A method for sending audio stream error concealment, characterized in that the method comprises:

a. Classify the sent audio frame according to the content to obtain the type information of the audio frame;

b. The type information of the audio frame and the encoding result of the audio frame are packaged and sent out.

A receiving method for audio stream error concealment, the method comprising:

a. When frame loss occurs, for the lost audio frame, determine the type information of the audio frame obtained when it is classified according to the content;

b. According to the type information of the lost audio frame, a corresponding error recovery strategy is used to reconstruct the audio frame.

A method for audio stream error concealment, the method comprising:

a. Classify the sent audio frame according to the content to obtain the type information of the audio frame;

b. Encapsulate and package the type information of the audio frame with the encoding result of the audio frame and send it to the receiving end;

c. When a frame is lost, for the lost audio frame, the receiving end determines the type information of the audio frame obtained when it is classified according to the content;

d. According to the type information of the lost audio frame, a corresponding error recovery strategy is used to reconstruct the audio frame.

A transmitter for audio stream error concealment comprising an audio encoder module, a frame packing module and an audio frame classifier module;

The audio frame classifier module is used to classify the sent audio frame according to the content, obtain the type information of the audio frame, and send the type information to the frame encapsulation module;

The frame encapsulation module is configured to receive the type information of the audio frame sent by the audio frame classifier module and the encoding result of the audio frame sent by the audio encoder module, and encode the type information of the audio frame and the audio frame The results are packaged and sent out.

A receiver for audio stream error concealment, the receiver includes a frame type discrimination module and an error concealment module,

The frame type discrimination module is used for the type information of the audio frame obtained when the lost audio frame is classified according to the content, and sends the type information to the error concealment module;

The error concealment module is configured to use a corresponding error recovery strategy to reconstruct the audio frame according to the type information of the received lost audio frame.

An audio stream error concealment system, the system includes: a transmitter and a receiver;

The transmitter is configured to classify the transmitted audio frame according to the content, obtain the type information of the audio frame, and package the type information of the audio frame and the encoding result of the audio frame to send to the receiver;

The receiver is configured to determine the type information obtained when the lost audio frame is classified according to content when a frame loss occurs, and use a corresponding error recovery strategy to reconstruct the audio frame according to the type information.

It can be seen from the above technical solution that the present invention classifies audio frames according to the content of the audio frame at the sending end, and sends the type information of the audio frame and the encoding result of the audio frame together; The different types obtained when the frame is classified according to the content, and the corresponding error concealment strategy is used to reconstruct the audio signal. It can be seen that the error concealment method described in the present invention makes the reconstruction of the lost frame more pertinent, and can adaptively reconstruct the audio frame to achieve a better compensation effect and bring better benefits to the receiving end user. Subjective auditory experience, while improving the resolution of audio frame signals, enabling audio communication to tolerate higher packet loss rates.

Description of drawings

Figure 1 is a block diagram of audio stream error concealment.

Fig. 2 is an overall flow chart of the method for audio stream error concealment in the present invention.

Fig. 3 is a general structural diagram of the system for audio stream error concealment in the present invention.

Fig. 4 is an overall structural diagram of a transmitter for audio stream error concealment in the present invention.

Fig. 5 is a general structural diagram of a receiver for audio stream error concealment in the present invention.

FIG. 6 is a specific flow chart of the sending method for audio stream error concealment in the embodiment of the present invention.

Fig. 7 is a schematic diagram of classifying audio frames in an embodiment of the present invention.

FIG. 8 is a specific flow chart of a receiving method for audio stream error concealment in an embodiment of the present invention.

FIG. 9 is a specific structural diagram of a transmitter for audio stream error concealment in an embodiment of the present invention.

FIG. 10 is a specific structural diagram of a receiver for audio stream error concealment in an embodiment of the present invention.

Detailed ways

In order to make the purpose, technical means and advantages of the present invention clearer, the specific implementation manners of the present invention will be described below in conjunction with the accompanying drawings and examples.

The basic idea of the present invention is to classify according to the audio frame content, and adopt corresponding error concealment strategy to reconstruct the audio signal according to the different types of the lost audio frame in case of packet loss.

Fig. 2 is an overall flow chart of the method for audio stream error concealment in the present invention. As shown in Figure 2, the method includes:

Step 201, classify the transmitted audio frame according to the content, and obtain the type information of the audio frame.

In step 202, the type information of the audio frame and the encoding result of the audio frame are packaged and sent out.

Step 203, when a frame loss occurs, for the lost audio frame, determine the type information of the audio frame obtained when it is classified according to the content;

Step 204, according to the type information of the lost audio frame, use a corresponding error recovery strategy to reconstruct the audio frame.

Wherein, steps 201-202 constitute the overall flow of the sending method for audio stream error concealment; steps 203-204 constitute the overall flow of the receiving method for audio stream error concealment.

Fig. 3 is a general structural diagram of the system for audio stream error concealment in the present invention. As shown in FIG. 3 , the system includes a transmitter 301 and a receiver 302 . In this system, the transmitter 301 is used to classify the transmitted audio frame according to the content, obtain the type information of the audio frame, package the type information of the audio frame and the encoding result of the audio frame and send it to the receiver 302 ; The receiver 302 is used to determine the type information obtained when the lost audio frame is classified according to the content when a frame loss occurs, and according to the type information, use a corresponding error recovery strategy to reconstruct the audio frame. The transmitter and receiver in this system can respectively adopt the specific structures of the transmitter 400 and the receiver 500 shown in FIG. 4 and FIG. 5 below.

Fig. 4 is an overall structural diagram of a transmitter for audio stream error concealment in the present invention. As shown in FIG. 4 , the transmitter 400 includes an audio encoder module 410 , an audio frame classifier module 420 and a frame packing module 430 .

In the transmitter 400 , the audio encoder module 410 is configured to encode the transmitted audio frame, and send the encoding result to the frame packing module 430 . The audio frame classifier is configured to classify the sent audio frame according to the content, obtain the type information of the audio frame, and send the type information to the frame encapsulation module 430 . The frame encapsulation module 430 is used to receive the encoding result of the audio frame sent by the audio encoder module 410 and the type information of the audio frame sent by the audio frame classifier module 420, and package the type information of the audio frame and the encoding result of the audio frame send it out. When encoding an audio frame in the audio encoder 410, different encoding methods may be used for the audio frame according to the type information of the audio frame sent by the audio frame classifier, or the same encoding method may be directly used for all encoded frames.

Corresponding to the transmitter, FIG. 5 is a general structural diagram of a receiver for audio stream error concealment in the present invention. The receiver includes: a frame type discrimination module 510 and an error concealment module 520 .

In the receiver, the frame type discrimination module 510 uses the type information of the audio frame obtained when the lost audio frame is classified according to the content, and sends the type information to the error concealment module 520 . The error concealment module 520 is configured to use a corresponding error recovery strategy to reconstruct the audio frame according to the type information of the received lost audio frame.

As can be seen from the above, the present invention classifies audio frames by content at the sending end to obtain the type information of the audio frame and sends it to the receiving end. When a frame is lost, the receiving end adopts different error recovery strategies according to the type information of the lost audio frame Perform audio frame reconstruction for efficient error concealment.

The above is a general overview of the method, device and system of the present invention, and the present invention will be further described in detail through specific embodiments below.

FIG. 6 is a specific flowchart of a method for audio stream error concealment in an embodiment of the present invention. As shown in Figure 6, the method includes:

Step 601, divide the audio signal into equally spaced audio frames.

In this step, the frame length of the audio frame is determined according to the encoding protocol.

Step 602: Analyze the content and characteristics of the audio frame to obtain type information of the audio frame.

In this step, the audio frame is divided into voice signal frame, noise signal frame, silent signal frame, musical tone signal frame and other types, and then each type can be further subdivided, for example, the voice signal frame can be further divided into voiced , unvoiced, voiced transition, unvoiced transition, onset and other types, the tone signal frame can be simply divided into stable tone frame (steady), transition tone frame (transition) and other types according to the stability characteristics of the signal.

Step 603, encoding and compressing the sent audio frame.

In this step, the same coding method may be used for the entire audio signal, or different coding methods may be used according to different types of audio frames.

Step 604, package the type of the audio frame and the result of encoding and compression, and send them out.

In this step, when encapsulating and packaging, the type information of the audio frame can be identified in the frame header of the current frame or the next frame.

So far, the sending process of audio stream error concealment in this embodiment ends.

In the above process, when audio frames are classified in step 602, the method shown in FIG. 7 may be used. Referring to FIG. 7 , firstly, VAD is used to detect whether the audio frame is a noise signal frame. If it is a noise signal frame, perform spectrum energy analysis on the audio frame, and if it is a non-noise signal frame, perform spectrum stability analysis on the audio frame.

According to the result of spectral energy analysis, the audio frame is divided into a silent signal frame and a noise signal frame, and then the silent signal frame or the noise signal frame can be further classified to obtain the type information of the audio frame.

According to the results of spectrum stability analysis, the audio frame is divided into speech signal frame and musical tone signal frame, and then the classification of the speech signal frame or musical tone signal frame can be further refined, such as the speech signal frame can be refined into voiced , unvoiced, voiced transition, etc. For the tone signal frame, it can be refined into stable tone frame, transition tone frame, etc.

Similar to the classification of audio frames in the above sending method, at the receiving end, the method shown in FIG. 8 is used for receiving in this embodiment. As shown in Figure 8, the receiving method includes:

Step 801 , perform frame loss detection on the audio signal, if frame loss occurs, execute step 804 and its subsequent steps, otherwise execute step 802 and its subsequent steps.

In this step, it is determined whether an audio frame is lost according to the frame number carried in the audio frame.

Step 802, detect and record the type of the audio frame.

When a frame is lost, the type information of the lost audio frame can be determined by using the type of the audio frame recorded in this step.

Step 803, decode the audio frame, and output the decoding result, and end this process.

In this step, according to the encoding method of the audio frame by the sending end, the corresponding decoding method is used for decoding.

Step 804, determine the type information obtained when the lost audio frames are classified according to content.

In this step, if the type information of the audio frame is transmitted in the audio frame, the receiving end extracts historical data, and infers the type of the currently lost frame according to the type information of the correctly received frame; if the type information of the audio frame is carried in other If it is transmitted in the correctly received audio frame, the receiving end can directly extract the type information of the currently lost frame from the corresponding correctly received audio frame.

Step 805, according to the type of the lost audio frame, adaptively adopt the corresponding error recovery strategy to reconstruct the audio frame, and output the reconstructed result, and end this process.

In this step, according to the type of the lost audio frame, the most suitable error recovery strategy for the type can be selected to reconstruct the audio frame. For example, a stable speech frame is very similar to its previous frame, and a frame copy strategy can be used to achieve a good hiding effect, while a transition frame needs to consider the state of the preceding and following frames to determine the hiding strategy.

So far, the receiving method of audio stream error concealment ends.

In the sending method of this embodiment, the audio frame is classified in the manner shown in FIG. 7 , and of course other content-based audio frame classification methods can also be used, as long as the audio frame can be classified according to the content. purpose.

It can be seen from the above that the cooperation between the sending end and the receiving end can use the classification of audio frames to efficiently implement error concealment, and greatly improve the tolerance of real-time audio communication to packet loss rate.

The foregoing is a specific implementation manner of the audio stream error concealment sending and receiving method provided in this embodiment. The specific implementation of the method for audio stream error concealment in the present invention can be constituted by the mutual cooperation of the two implementations. In addition, this embodiment also provides specific implementation manners of a corresponding audio stream error concealment transmitter and receiver.

FIG. 9 is a specific structural diagram of a transmitter for audio stream error concealment in an embodiment of the present invention. As shown in FIG. 9 , the transmitter 900 includes: an audio encoder module 910 , an audio frame classifier module 920 , a frame packing module 930 and an audio frame dividing module 940 .

In the transmitter 900, the audio frame division module 940 is used to divide the audio signal into equally spaced audio frames according to different encoding protocols, and send the audio frames to the audio encoder module 910 and the audio frame classifier module 920 .

The audio encoder module 910 is configured to encode the audio frame and send the encoding result to the frame packing module 930 . The audio frame classifier module 920 is used to classify the audio frames according to the content, and the specific classification method can adopt the method shown in FIG. 7 , and send the type information of the audio frame to the frame encapsulation module 930 .

The frame encapsulation module 930 is configured to receive the audio frame encoding result sent by the audio encoder module 910 and the audio frame type information sent by the audio frame classifier module 920, encapsulate the type information and the audio frame encoding result, and send them out. When performing encapsulation and packaging, the type information of the audio frame can be encapsulated in the current audio frame or the next audio frame, specifically in the frame header.

When encoding an audio frame in the audio encoder 910, different encoding methods may be used for the audio frame according to the type information of the audio frame sent by the audio frame classifier, or the same encoding method may be directly used for all encoded frames.

FIG. 10 is a specific structural diagram of a receiver for audio stream error concealment in an embodiment of the present invention. As shown in FIG. 10 , the receiver 1000 includes a frame type discrimination module 1010 , an error concealment module 1020 , an error detection module 1030 and an audio decoder module 1040 . Among them, the frame type judgment module 1010 includes a judgment sub-module 1011 and a storage sub-module 1012 ; the error concealment module 1020 includes a policy decision sub-module 1021 and an error concealment sub-module 1022 .

In the receiver 1000, the error detection module 1030 is used to receive the audio frame from the channel, send the received audio frame to the discrimination sub-module 1011 in the frame type discrimination module 1010, and detect whether there is a frame loss, if there is If the frame is lost, the discrimination sub-module 1011 in the frame type discrimination module 1010 will be notified.

In the frame type discrimination module 1010, when determining the type obtained by classifying the audio frame according to the content, if the type information of the audio frame is carried in the correctly received audio frame, the type information is directly extracted and stored in the storage submodule In 1012: if the type information of the audio frame is carried in the lost audio frame, infer the type information obtained when the lost audio frame is classified according to the content according to the types of the preceding and following frames.

In the error concealment module 1020, the policy judgment submodule 1021 is used to receive the type information of the lost frame sent by the discrimination submodule 1011, and according to the type information, determine the error recovery strategy adopted, and send the result to the error concealment submodule 1022. The error concealment sub-module 1022 is configured to reconstruct the lost audio frame according to the error recovery policy decision result sent by the policy decision sub-module 1021 .

The audio frame decoder module 1040 is configured to decode the received audio frame and output the decoding result.

In the transmitter 900 of this embodiment, the audio frame classifier module 920 uses the method shown in FIG. 7 to classify the audio frames. Of course, other content-based audio frame classification methods may also be used, which will not be repeated here. In the receiver 1000, the frame type decision module 1010 is subdivided into a decision submodule 1011 and a storage submodule 1012, which respectively perform frame type judgment and storage; the error concealment module 1020 is subdivided into a policy decision submodule 1021 and an error concealment submodule Module 1022, respectively perform policy decision and error concealment. Wherein, the error concealment sub-module 1022 can be further divided into multiple different types of error concealment units, such as noise error concealment units, voice error concealment units, etc., for handling the error concealment of different types of audio frames.

The implementation manner of the audio stream error concealment system of the present invention may be: use the transmitter 900 and the receiver 1000 shown in FIG. 9 and FIG. 10 as the specific implementation manners of the transmitter and receiver in the audio stream error concealment system, and the The audio frame output by the frame encapsulation module 930 in the transmitter 900 is sent to the error detection module 1030 in the receiver 1000 . In this way, an embodiment of the audio stream error concealment system of the present invention can be constituted.

It can be seen from the above specific implementation of the method, device and system of the present invention that the technical solution of the present invention makes the reconstruction of lost frames more pertinent, and can adaptively reconstruct audio frames to achieve better The compensation effect brings better subjective auditory experience to the receiving end user, and at the same time improves the distinguishability of the audio frame signal, making the audio communication tolerant to a higher packet loss rate.

The above are only preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims (24)

1. A transmission method for audio stream error concealment, characterized in that the method comprises: a. Classify the sent audio frame according to the content to obtain the type information of the audio frame; b. The type information of the audio frame and the encoding result of the audio frame are packaged and sent out. 2. The method according to claim 1, wherein classifying the audio frames sent according to the content in step a comprises: a1, utilize the voice activation detector VAD to detect whether the audio frame is a noise signal frame, if so, then perform step a2, otherwise perform step a3; a2. Perform spectrum energy analysis on the noise signal frame, and determine that the audio frame is a silent signal frame or a noise signal frame; a3. Spectrum stability analysis is performed on the non-noise signal frame, and the audio frame is determined to be a speech signal frame or a musical tone signal frame. 3. The method according to claim 2, characterized in that the silent signal frame, the noise signal frame, the speech signal frame or the tone signal frame are further finely divided. 4. The method according to claim 1, wherein said encapsulating and packing the type information of the audio frame and the encoding result of the audio frame is: packing the type information of the audio frame into a frame header of the audio frame. 5. The method according to claim 4, wherein the frame header of the audio frame is: the frame header of the audio frame represented by the type information of the audio frame or the lower part of the audio frame represented by the type information of the audio frame The header of a frame. 6. The method according to claim 1, wherein the encoding result of the audio frame is: the encoding result obtained by using the same encoding method for all audio frames, or, according to the different types of audio frames, different encoding methods are used. The encoded result of the method. 7. A receiving method for audio stream error concealment, characterized in that the method comprises: a. When frame loss occurs, for the lost audio frame, determine the type information of the audio frame obtained when it is classified according to the content; b. According to the type information of the lost audio frame, a corresponding error recovery strategy is used to reconstruct the audio frame. 8. The receiving method according to claim 7, characterized in that, the type information of the audio frame obtained when determining that it is classified according to content as described in step a is: When the type information of the lost audio frame is located in the correctly received audio frame, directly extracting the type information of the lost audio frame from the correctly received audio frame; When the type information of the lost audio frame is located in the lost audio frame, the type information of the lost audio frame is deduced according to the type information of the correctly received frame. 9. The receiving method according to claim 7 or 8, further comprising: before step a, detecting whether there is frame loss in the received audio signal, if so, performing step a and subsequent steps thereof, otherwise, The correctly received audio frame is decoded, and the type information of the audio frame carried in the frame is extracted and stored. 10. A method for audio stream error concealment, characterized in that the method comprises: a. Classify the sent audio frame according to the content to obtain the type information of the audio frame; b. Encapsulate and package the type information of the audio frame with the encoding result of the audio frame and send it to the receiving end; c. When a frame is lost, for the lost audio frame, the receiving end determines the type information of the audio frame obtained when it is classified according to the content; d. According to the type information of the lost audio frame, a corresponding error recovery strategy is used to reconstruct the audio frame. 11. The method according to claim 10, wherein classifying the audio frames sent according to the content in step a comprises: a1, utilize the voice activation detector VAD to detect whether the audio frame is a noise signal frame, if so, then perform step a2, otherwise perform step a3; a2. Perform spectrum energy analysis on the noise signal frame, and determine that the audio frame is a silent signal frame or a noise signal frame; a3. Spectrum stability analysis is performed on the non-noise signal frame, and the audio frame is determined to be a speech signal frame or a musical tone signal frame. 12. The method according to claim 10, characterized in that, the type information of the audio frame and the encoding result of the audio frame are packaged and packed as described in step b: the type information of the audio frame is packed into the frame header of the audio frame middle. 13. The method according to claim 12, wherein the frame header of the audio frame is: the frame header of the audio frame represented by the type information of the audio frame or the lower part of the audio frame represented by the type information of the audio frame The header of a frame. 14. The method according to claim 10, wherein the type information of the audio frame obtained when determining that it is classified according to content as described in the step c is: When the type information of the lost audio frame is located in the correctly received audio frame, directly extracting the type information of the lost audio frame from the correctly received audio frame; When the type information of the lost audio frame is located in the lost audio frame, the type information of the lost audio frame is deduced according to the type information of the correctly received frame. 15. The method according to any one of claims 10 to 14, characterized in that, between steps b and c, further comprising: detecting whether a frame is lost in the received audio signal, if so, performing steps c and Its subsequent steps, otherwise, decode the correctly received audio frame, extract and store the type information of the audio frame carried in the frame. 16. A transmitter for audio stream error concealment, comprising an audio encoder module and a frame packing module, characterized in that the transmitter also includes an audio frame classifier module; The audio frame classifier module is used to classify the sent audio frame according to the content, obtain the type information of the audio frame, and send the type information to the frame encapsulation module; The frame encapsulation module is configured to receive the type information of the audio frame sent by the audio frame classifier module and the encoding result of the audio frame sent by the audio encoder module, and encode the type information of the audio frame and the audio frame The results are packaged and sent out. 17. A receiver for audio stream error concealment, characterized in that the receiver includes a frame type discrimination module and an error concealment module, The frame type discrimination module is used to determine the type information of the audio frame obtained when the lost audio frame is classified according to the content, and send the type information to the error concealment module; The error concealment module is configured to use a corresponding error recovery strategy to reconstruct the audio frame according to the type information of the received lost audio frame. 18. The receiver according to claim 17, further comprising an error detection module and an audio decoder module, The error detection module is used to receive audio frames from the channel, send the received audio frames to the frame type discrimination module, and detect whether frame loss occurs, and if frame loss occurs, notify the frame type discrimination module ; The frame type discrimination module is further used to forward the audio frame to the audio frame decoder module; The audio frame decoder module is used to decode the audio frame. 19. The receiver according to claim 18, wherein the frame type discrimination module comprises a discrimination submodule and a storage submodule, The discrimination submodule is used to determine the type information of the audio frame obtained when the lost audio frame is classified according to the content, and send the type information to the storage submodule, and is also used to receive the error detection After the frame loss notification sent by the module, the type information of the lost frame is sent to the error concealment module, and the received audio frame is forwarded to the audio frame decoder module; The storage sub-module is used for saving the type information of the audio frame. 20. The receiver according to claim 18 or 19, wherein the error concealment module comprises a strategy decision submodule and an error concealment submodule, The strategy judgment submodule is used to receive the type information of the lost frame sent by the frame type judgment module, and judge the error recovery strategy adopted according to the type information, and send the result to the error concealment submodule; The error concealment submodule is configured to reconstruct the lost audio frame according to the error recovery policy decision result sent by the policy decision submodule. 21. An audio stream error concealment system, characterized in that the system comprises: a transmitter and a receiver; The transmitter is configured to classify the transmitted audio frame according to the content, obtain the type information of the audio frame, and package the type information of the audio frame and the encoding result of the audio frame to send to the receiver; The receiver is configured to determine the type information obtained when the lost audio frame is classified according to content when a frame loss occurs, and use a corresponding error recovery strategy to reconstruct the audio frame according to the type information. 22. The system according to claim 21, wherein the transmitter includes an audio encoder module, a frame packing module and an audio frame classifier module, and the receiver includes a frame type discrimination module and an error concealment module; The audio frame classifier module is used to classify the sent audio frame according to the content, obtain the type information of the audio frame, and send the type information to the frame encapsulation module; The frame encapsulation module is configured to receive the type information of the audio frame sent by the audio frame classifier module and the encoding result of the audio frame sent by the audio encoder module, and encode the type information of the audio frame and the audio frame The result is packaged and sent out; The frame type discrimination module is used to determine the type information of the audio frame obtained when the lost audio frame is classified according to the content, and send the type information to the error concealment module; The error concealment module is configured to use a corresponding error recovery strategy to reconstruct the audio frame according to the type information of the received lost audio frame. 23. The system of claim 22, wherein the receiver further comprises an error detection module and an audio decoder module, The error detection module is used to receive audio frames from the channel, send the received audio frames to the frame type discrimination module, and detect whether frame loss occurs, and if frame loss occurs, notify the frame type discrimination module ; The frame type discrimination module is further used to forward the audio frame to the audio frame decoder module; The audio frame decoder module is used to decode the audio frame; The frame encapsulation module is configured to encapsulate and package the type information of the audio frame and the encoding result of the audio frame and send them to the error detection module. 24. The receiver according to claim 22 or 23, wherein the error concealment module comprises a policy decision submodule and an error concealment submodule, The strategy judgment submodule is used to receive the type information of the lost frame sent by the frame type judgment module, and judge the error recovery strategy adopted according to the type information, and send the result to the error concealment submodule; The error concealment submodule is configured to reconstruct the lost audio frame according to the error recovery policy decision result sent by the policy decision submodule.

Images