KR100377571B1 - Apparatus for processing voice chatting data - Google Patents

Abstract

Disclosed is a multi-party voice chat data processing device capable of increasing voice data processing efficiency by significantly reducing voice data traffic on a network by synthesizing voice data of multiple users who participated in a voice chat and then using a single bandwidth. do.

According to the present invention, voice data from multiple users is synthesized into single synthesized voice data, and the final synthesized data transmitted to each chat participant becomes a single channel with a single bandwidth, thereby significantly reducing the data processing load of the chat participant terminal. In addition, all chat participants can process voice data with only a single bandwidth, so there is no need to limit the number of chat participants.

Description

Multi-party voice chat data processing device {Apparatus for processing voice chatting data}

The present invention relates to an apparatus for processing multi-party voice chat data, and more particularly, by synthesizing voice data of a plurality of users who participated in a voice chat and transmitting / receiving using a single bandwidth to significantly reduce voice data traffic on a network. The present invention relates to a multi-party voice chat data processing device capable of increasing data processing efficiency.

When transmitting and receiving voice over a digital network including the Internet, the voice, which is an analog signal, cannot be directly transmitted, so it is converted into digital voice data converted into digital data and transmitted. In this case, the bandwidth required for the transmission of digital data is determined by how much the sampling rate necessary for converting the analog signal into digital data is required. In general, a bandwidth of 64 kbps is required. It must be high, which means an increase in the digital voice data being transmitted, thus requiring higher network bandwidth. Therefore, various types of digital compression methods are adopted to secure stable transmission / reception bandwidth of data in a network that provides a fixed speed like the Internet, and some international standards that define the methods are 32kbps ADPCM, G.729, CELP 3.2a. Examples include LPC-10, G711 / 721/723, G.728 LD-CELP, GSM 6.10, and MPEG-1 Layer3.

1 to 3 show a conventional voice chat configuration method.

1 is a 1: 1 voice chat configuration, which is the most basic form of chat, and a representative example is an internet telephone.

This configuration uses the most basic and general type of voice signal transmission and reception, and the queuing method for stable voice playback by securing a fixed bandwidth in the Internet that provides digital voice compression and non-fixed communication efficiency. This is used, and due to the nature of the 1: 1 method, processing problems due to an increase in voice data traffic do not occur.

2 and 3 is a voice chat configuration method in which three or more people simultaneously participate, FIG. 2 is called a chat server directory method, and FIG. 3 is called a chat server relay method.

In the chat server directory method shown in FIG. 2, the chat server merely provides a directory function to inform each user of only the address on the Internet of the user participating in the chat (dashed arrow portion of FIG. 2), and the actual data flow of the voice chat. The distribution is composed between each user and distributed and synthesized (ie, the server only connects users on a network and does not participate in the distribution and synthesis of data). In this case, each user terminal should transmit / synthesize all of its own voice data to everyone else participating in the chat, and also receive / synthesize all voice data received from each user. Therefore, as the number of users participating in the chat increases, the transmission bandwidth and the reception bandwidth required for each user increase in proportion to the number of members. One solid arrow in FIG. 2 represents the bandwidth (12kbps) allocated to each user. As shown in FIG. 2, when four chat participants are present, the bandwidth required for each user is 48kbps, which is four times 12kbps in a 1: 1 manner, for both transmission and reception. In general, a user terminal (typically a communication card and a voice processing card embedded in a PC) usually has a finite bandwidth for data processing, so the number of simultaneous chat participants is very limited. For example, if a voice data transmission has a bandwidth of 12 kbps, only four people can simultaneously chat using a 56 kbps PSTN network.

The chat server relay method shown in FIG. 3 is a method in which the chat server receives voice data from all users participating in a chat and relays voice data to all users. In this case, all user terminals fulfill the task of transmitting voice data by transmitting their own voice data to the chat server (i.e., transmission is possible only with a single bandwidth). However, since the voice data relayed by the chat server is repeatedly transmitted to all other users, each user terminal needs to receive / synthesize the voice data sent by everyone else and reproduce the original voice. In this case, the transmission bandwidth requires a constant size of one channel (12kbps) for each user. However, as for the reception bandwidth, as the number of users participating in the chat increases, the reception bandwidth on the Internet required for each user is increased. It will increase in proportion to the number. In addition, data traffic throughout the network grows exponentially as the number of chat users grows, requiring a tremendous amount of network throughput.

Therefore, the existing multi-party real-time voice chat has a problem that the number of concurrent participants is very limited. In addition, in the case of the existing multi-party chat, even in the case of high-speed leased lines, an increase in exponential network traffic in proportion to the number of users causes a situation in which simultaneous chats of several dozen or more people are impossible due to a slight increase in simultaneous participants.

Therefore, the present invention was devised to solve such a problem, and an object of the present invention is to synthesize voice data of a plurality of users who participated in a voice chat and to relay using a single bandwidth, thereby significantly reducing voice data traffic on the network. It is an object of the present invention to provide a multi-party voice chat data processing apparatus capable of increasing processing efficiency.

1 is a block diagram of a 1: 1 chat method.

Figure 2 is a configuration of the chat server directory method of the chat method through the existing voice chat server.

3 is a configuration of the chat server relay method of the chat method through the existing voice chat server.

4 is a device configuration diagram and a network configuration diagram of the present invention.

5 is a conceptual configuration in which the present invention is implemented.

In order to achieve the above object, the present invention, by combining a plurality of users participating in the voice chat and the voice data transmitted from each of the users to generate a single-band synthesized voice data to send the chat to each user In the multi-party voice chat data processing device comprising a server: The chat server, the incoming data amount of the received synchronization signal of the device for asynchronous correction to each of the user-specific voice data flowing asynchronously through the network The excess factor value is given to voice data that is larger than the data amount normally flowing during the interval, and the delay factor value is given to voice data that is smaller than the data amount normally flowing during the receiving synchronization signal interval of the device. Modification factor value by the method Voice data correction / queuing module for queuing and then synchronizing to a voice reception queue, a decompression module for decompressing the queued voice data for each user, and generating synthesized voice data by synthesizing the decompressed voice data for each user. A final data synthesis module extracting the final synthesis data by erasing the self speech data from the synthesized speech data, a compression module compressing the final synthesis data, and a corrected transmission synchronization signal interval reflecting the correction factor value It characterized in that it comprises a synchronization module.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Figure 4 shows the device configuration and network configuration of the present invention.

Since only the configuration of the voice data processing part is presented in the user terminal 100 and the presented components are known in the art, each function will be briefly described.

The voice digitization processing unit 101 is a part for digitizing the analog voice of the user. Digitized data is in the form of PCM (Pulse Code Modulation) codes.

The voice data compressor 102 compresses the PCM data described above, and the compression method is one of several standards mentioned above.

The external communication unit 105 is a portion for transmitting and processing compressed data or receiving and receiving voice data from the chat server 200, for example, a communication card usually built in a PC.

The voice data decompressor 104 restores the voice data received from the chat server 200.

The voice analog processing unit 103 generates a voice directly heard by the user as a part of analogizing the received received voice data.

The voice data correction / queuing module 201 is a part for queuing the voice reception queue 202 after synchronously correcting by giving a correction factor value to voice data for each user asynchronously introduced through the network 300. Queuing is performed in the voice reception queue 202 by correcting the asynchronousness of user-specific voice data due to the delay of data transmission occurring within the network 300 itself and the performance difference of each user terminal.

In general, the above-mentioned queuing technique is widely used in the processing of data flowing asynchronously into the server, which may guarantee the inflow of the same amount of data from all users due to a delay in the network or a slight difference in the data generation synchronization signal of the user terminal. As a result, the server keeps data for a certain amount of time by placing a receiving queue of a certain size for each channel allocated to each user in the server to buffer the data asynchronously due to delays in the network. It can be a technique. However, since the difference in the data generation synchronization signal between each user accumulates over time, the shortage or excess of data occurring at a predetermined interval cannot be completely eliminated by simply securing a reception queue.

Therefore, in the present invention, in addition to such a queuing technique, in order to solve a lack or excess of data, a correction factor value (a delay factor value or an excess factor value) is added to each user's voice data to solve the asynchronousness of the voice data flowing into the server. I would like to.

First, the voice reception queue 202 is allocated to a predetermined size for each user channel. The predetermined size is the amount of voice data that should normally flow into the server 200 at the time interval of the voice data reception synchronization signal of the server 200. In this interval unit, the server 200 receives the voice data of each user from the voice reception queue 202. ).

The voice data correction / queuing module 201 records (queues) the voice data of each user in the voice reception queue 202 and sets one of the two parameter values in the voice data for each user. The voice reception queue 202 is provided with an excess factor value for data received faster than the voice data reception synchronization signal interval (data amount larger than the data amount normally introduced during the reception synchronization signal interval) and the corresponding voice data is recorded. Delete the oldest data from and record the newly imported voice data.

At this time, the excess factor value is given as a time value. For example, if the reception synchronization signal time interval of the server 200 is 10 and the synchronization signal time interval of rapidly received data is 9, the excess factor value is 1.

The voice data correction / queuing module 201 also provides a delay factor value for data received slower than the voice data reception synchronization signal interval of the server 200 (data amount less than the data amount normally introduced during the reception synchronization signal interval). To give.

At this time, the delay factor value is given as a time value as in the excess factor value. For example, if the reception synchronization signal time interval of the server 200 is 10 and the synchronization signal time interval of slowly received data is 12, the delay factor value Becomes -2.

The voice data correction / queuing module 201 assigns an excess factor value or a delay factor value to the voice data of each user and records (queues) it in the voice reception queue 202.

The decompression module 203 reads the compressed voice data of each user from the voice reception queue 202 and decompresses the decompressed data. The decompressed data passes through the voice digitization processing unit 101 of the user terminal 100. A code, i.e., original audio data of a user having a PCM code form, is recorded in the original speech queue 204.

The voice data synthesizing module 205 synthesizes each expanded voice data to generate synthesized voice data and queues the synthesized voice cues 206. The synthesis is performed by adding PCM codes of the original voice data to each other. The synthesizing module 205 first reads the original audio data from the original audio queue 204 at predetermined time intervals. If the PCM code of each original audio data is an integer 16-bit format, the synthesized speech data obtained by adding the respective original audio data is also integer 16. Bits allow transmission to each user using a single bandwidth. If 16 bits are exceeded as a result of synthesis, clipping is performed based on the maximum value of integer 16 bits to generate integer 16 bit synthesized speech data. In case of speech, it is different in discriminating individual data after synthesis. Unlike form data (images, texts, etc.), this synthesis is possible because it is known that there is no difficulty in recognizing individual voices even after synthesis.

The final synthesized data extraction module 207 erases the self-voice data from the synthesized speech data and then extracts the final synthesized data to be transmitted to each user and then synthesizes it from the entire synthesized speech queue 206 as a part of queuing to the final synthesized speech queue 208. The PCM code of the voice data is read, and the PCM code of the user-specific voice data is read from the original voice queue 204, and the final synthesized data is extracted as the difference between the two PCM codes, and recorded in the final synthesized sound queue 208. If the synthesized voice data is transmitted to each user without erasing the self voice data, since the data includes any user's own voice, a howling phenomenon occurs in which the user hears his voice like an echo. The voice data will be erased.

The user's voice data may be received slower or faster than the voice data reception synchronization signal interval of the server 200, and the delay factor may be delayed by the voice data correction / queuing module 201 as mentioned above to correct this situation. Value or excess factor value. The synchronization module 209 reflects these factors, calculates a corrected transmission synchronization signal interval reflecting the delay factor value and the excess factor value, and reads out user-specific composite data from the final synthesis tone queue 208 at the corrected transmission synchronization signal interval. Send to compression module 210. The corrected transmission synchronization signal interval is specifically calculated as shown below.

Corrected transmission synchronization signal interval = Transmission synchronization signal interval of the server itself + (average of delay factor value + average of overfactor value) * correction factor value

Here, the average value of the delay factor is the average value of the delay factor values of the voice data given the delay factor value, and the average value of the excess factor is the average value of the excess factor values of the voice data given the excess factor value. The correction factor value is experimentally extracted as a time conversion of the deviation between the two factor values. If the correction factor value is too large, the change of the synchronization signal becomes very severe, and an appropriate value should be set. In particular, by setting the maximum value and the minimum value of the correction transmission synchronization signal interval in advance, it is possible to prevent distortion of the synthesized sound that may occur when the correction factor value is very large.

For example, if the transmission synchronization signal interval of the server itself is 10, the delay factor value average is -2, the excess factor value average is 1, and the correction factor value is 0.2, then the corrected transmission synchronization signal interval is 9.8 according to the above equation. If the minimum value of the synchronization signal interval is assumed to be 9.9, the calculated synchronization signal interval is corrected to 9.9.

The final synthesized data transmitted to the compression module 210 is transmitted to each user by the transmission module 211 through a predetermined compression scheme, expanded by the voice data extension unit 104 of each user terminal, and the expanded data is voiced. The analogization processing section 103 converts the analog speech into speech to reproduce the speech.

In summary, the multi-party voice chat data processing apparatus of the present invention is basically implemented in the configuration of FIG. 3, that is, a chat server relay method, but synthesizes a single bandwidth by real-time synthesizing voice data transmitted from each user inside the server. This creates a voice channel, which is different from the existing chat server. That is, the existing chat server merely plays the role of receiving voice data from chat participants and relaying them without any manipulation, but the chat server of the present invention receives voice data from multiple users and synthesizes them into a single synthesized voice data. Since the final composite data transmitted to the chat participants becomes a single channel with a single bandwidth, all chat participants can receive the conversation contents voice data using only one channel bandwidth. Therefore, even in the case of a chat with almost unlimited chat participants, each chat participant can enjoy the chat with only a single channel bandwidth.

FIG. 5 is a conceptual diagram illustrating the implementation of the present invention in correspondence with FIG. 3.

While preferred embodiments of the invention have been described in detail above, those of ordinary skill in the art will appreciate that the invention may be modified without departing from the spirit and scope of the invention as defined in the appended claims. It will be appreciated that modifications or variations may be made. Thus, modifications to the foregoing embodiments of the invention will not depart from the teachings of the invention.

Use of the present invention has the following effects.

Voice data from multiple users is synthesized into a single synthesized voice data, and the final synthesized data transmitted to each chat participant becomes a single channel of a single bandwidth, thereby significantly reducing the load of processing the received data of the chat participant terminal. In addition, all chat participants can process voice data with only a single bandwidth, so there is no need to limit the number of chat participants.

This feature of the present invention can be applied not only to chat but also to teleconferences between large corporations or groups.

Claims (5)

And a chat server configured to synthesize a plurality of users participating in a voice chat and voice data transmitted from the respective users, thereby generating a single bandwidth of synthesized voice data and transmitting the synthesized voice data to the respective users. In: The chat server, In order to asynchronously correct each of the user-specific voice data flowing asynchronously through the network, an excess factor value is given to the voice data having a larger amount of incoming data than the data amount normally flowing during the reception synchronization signal interval of the device. In addition, a correction factor is given to the voice data in such a manner that a delay factor is given to the voice data having a smaller amount of the incoming data than the data flowing normally during the reception synchronization signal interval of the device, and then queued in the voice reception queue. voice data correction / queuing module for queuing; A decompression module for decompressing the queued voice data for each user; A voice data synthesizing module for synthesizing the decompressed voice data of each user to generate synthesized voice data; A final synthesized data extraction module for extracting final synthesized data by erasing self speech data from the synthesized speech data; A synchronization module for calculating a corrected transmission synchronization signal interval in which the correction factor value is reflected; A compression module for compressing the final composite data; And And a transmitting module for transmitting the compressed final synthesized data to the respective users at the transmission synchronization signal intervals. delete The apparatus of claim 1, wherein the synthesized speech data is generated by adding PCM codes of the voice data of each user. The apparatus of claim 1, wherein the final synthesized data is generated as a difference between a PCM code of the synthesized voice data and a PCM code of the voice data for each user. 6. The method of any one of claims 1 to 5, wherein the corrected transmission synchronization signal interval is an average of the transmission synchronization signal interval and the delay factor value of the device itself, the average of the excess factor values, and the deviation of the two factor values. ) Is calculated by the following equation by reflecting a correction factor value which is a value converted into time. Corrected transmission synchronization signal interval = transmission synchronization signal interval + (average of delay factor value + average of overfactor value) * correction factor of the apparatus itself.