JPH027643A - Audio multiple packet transmission system - Google Patents

Abstract

PURPOSE:To prevent rapid deterioration in the audio quality due to packet abort in the transmission by providing a means generation an audio packet header including abort object designation information and applying composition/ decomposition of audio packet and aborting the abort object designation packet. CONSTITUTION:An abort object block edit/reproduction section 300 stores a converted 8-bit code 201 in dual memories I1-Im having a prescribed capacity 301 corresponding to a channel, uses the content of a memory occupied by the prescribed capacity 301, that is, a coded data 202, utilities, e.g., an embedded code, accumulates the code bit not causing rapid deterioration of quality to form an abort object packet and integrates the coded data of each channel synchronously with each other into one and same packet. Thus, even if the abort object packet is aborted by congestion restriction of a packet network or the like, the unbalance between the rapid quality deterioration in the audio equality and plural audio channels is prevented.

Description

[Detailed Description of the Invention] "Industrial Application Field" This invention is a system that encodes multi-channel audio signals with high quality and transmits them as packets. An audio multiplex packet transmission method that prevents audio signals and easily realizes synchronization between multiple channels.
~? , 64kbit/00011z band audio/music
High quality encoding method using S (64 kb/s SB -
The DPCM method) is recommended by CCITT International Standard G, 722, and is used in teleconferencing systems that achieve high quality audio and a sense of realism.

If audio encoded data using this SB-ADPCM method is simply transmitted in packets, large noise will occur in the audio reproduction portion corresponding to the discarded encoded data due to packet discard during transmission. Since it is a PCM system, it has the disadvantage that the audio reproduction portions of other packets are affected, and this causes rapid deterioration of audio quality.

On the other hand, as a method of transmitting stereo signals,
A sum signal (L+R signal) and a difference signal (L-R signal) of the right channel and left channel are generated, and the current audio channel (first
The sum-difference method transmits the sum signal in a new audio channel (channel) and the difference signal in a new second audio channel (second channel), and the separation method transmits left and right channel signals independently in separate channels (first and second channels). is known (Journal of the Television Society, Vol. 34; No. 4, PP326-327.198
0).

Although both methods are capable of transmitting audio packets and realizing the sense of presence due to stereo sound, the former has the disadvantage that it is not possible to realize the sense of presence of sound of three or more channels with two channels. On the other hand, the latter has the advantage that by increasing the number of channels, it is possible to realize a 3-channel system that achieves a sense of realism on a large TV screen, and a 4-channel system that achieves a three-dimensional sound effect, but on the other hand, the transmission When the above packet is discarded, each channel independently assembles and disassembles the packet, so it is necessary to identify each packet dropout and synchronize between channels for packets that are not dropped. This has the disadvantage that the synchronization control for this becomes complicated. Another disadvantage is that during synchronization, packets may be missing or missing in each channel at a certain point, resulting in unbalanced audio quality between channels.

The purpose of this invention is to prevent the rapid deterioration of voice quality caused by packet discard during transmission and the imbalance of voice quality between multiple voice channels. Utilizing the embedded code used in the quality encoding method, 2. It is an object of the present invention to provide an audio multiplex packet transmission system in which a packet to be discarded is formed by accumulating code bits that do not cause severe quality deterioration, and in which encoded data of each synchronized channel is incorporated into the same packet.

[Means for Solving the Problem] According to the present invention, coded data is collected in order to form a packet to be discarded by accumulating coded bits that do not cause sudden deterioration of voice quality due to omissions of the coded data. A fixed capacity is stored, and block division is performed by editing and reproducing means to generate packets to be discarded and packets not to be discarded, and a plurality of block data generated for each channel are synchronized by multiplexing and dividing means. The header part (encoded header) consisting of the block playback number, synchronization control number, and channel multiplexing number of the block data is generated and read by the encoded header processing means, and includes discard target designation information. An audio packet header is generated, the audio packet is assembled and disassembled, and when packets are discarded during transmission, packets designated to be discarded are discarded.

Figure 2 shows high quality encoding method (SR-ADPCM method)
This shows the 8-bit code structure of G,722 used in The 8-bit code 201 is a high frequency (4 to 8 kHz) code 2
10 and a low frequency (0 to 4 kHz) code 220. Furthermore, the high-frequency code 210 has code pings) 211, 212.
, the low-band code 220 has code bits 221, 222, 2
It consists of 23, 224, 225° and 226.

The 8-bit code 201 uses embedded encoding, and the embedded encoding uses code bits 225
226 is missing, there is no sudden deterioration of voice quality (IEICE, IN
37-81. PPl-6, December 24, 1987)
.

Taking advantage of this point, G, 722 MODE 2 (56
kb/s audio + 8kb/s data) or MODE3 (4 kb/s
8 kb/s voice + 16 kb/s data) is used.

FIG. 3 is an example showing the features of this invention. The data in which the 8-bit code 201 is stored in a certain storage capacity 1301 is called encoded data 202), code pick l-225
, 226 for transmission block data 31 to be discarded.
0 and the code 211° 212. Transmission block data 320, 330, 340 that are not to be discarded are generated by dividing the data consisting of 221 to 224 into the same block size as the block data 310 to be discarded (Figure a). , these transmission block data 310, 320, 330°34
0 is incorporated into the audio packet 601. Figure 3(b) shows
When transmitting multiple channels simultaneously, (a) transmission block data 310, 320, 330, 3 generated in each channel (1st channel to mth channel) as shown in the figure.
40 is multiplexed into the same audio packet 601 in order to transmit them synchronously.

As described above, the present invention forms packets to be discarded by accumulating code bits with little quality deterioration due to omissions from the encoded data of the embedded code, and also integrates the encoded data of each synchronized channel into the same voice. It is characterized by being incorporated into a packet, and when a packet is discarded, it balances the voice quality of multiple channels and minimizes voice quality deterioration. In the conventional technique, a plurality of channels are individually packetized without distinction, and when a packet is discarded, it is dealt with uniformly, which has the disadvantage of causing rapid quality deterioration and variations in voice quality.

Embodiment FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, in which 10 is an input terminal for the first channel to the mth channel, 20 is an output terminal for the first channel to the mth channel, 30
is a packet network input/output terminal, lOO is an analog/digital (A/D) converter, 200 is an audio digital encoding/decoding converter, 300 is a discard block editing/reproducing unit, and 400 is a channel multiplexing/separating unit. , 500 is an encoded header processing unit, 600 is an audio packet assembly/disassembly unit, 70
0 is a control unit.

To operate this, according to a command from the control section 700, an audio signal input manually from the input terminal 10 is converted into an en+bedded code 8 through the analog/digital conversion section 100 and the audio digital encoding/decoding conversion section 200. It is converted into a bit code 201 for each channel (++ to 1.). The discard target block editing/reproducing unit 300 converts the converted 8-bit code 201 into a certain capacity 301 corresponding to the channel.
Dual memory ■, ~Memory 1. Accumulate in.

When the fixed capacity 301 on one side of the dual memory is filled, the discard target block editing/reproduction unit 300 stores the blocks on the remaining side of the dual memory. Using the contents of the memory that has filled a certain capacity 301, that is, the encoded data 202, the discard target block editing/reproducing unit 300 takes advantage of the characteristics of the embedded code to remove code bits that do not cause rapid quality deterioration, such as low-frequency bits. Low 2 bits of code 220 225.2
26 into one block and discard block data 31
0, and the remaining data is divided and divided into blocks to have the same block size as the block data 310 to be discarded to form block data 320, 330: 340 not to be discarded (Fig. 3). reference).

After the formation is completed, the block editing/reproduction unit 300 to be discarded
is the formed block data 310.320° 33
0,340 in a certain order, the channel multiplexing/demultiplexing unit 400
Transfer to. At the time of transfer, a synchronization control number (CN) 351 (see also FIG. 4), a block playback number (BN) 352, and additional information specifying whether or not the block is to be discarded are also transferred to each block.

The synchronization control number (CN) 351 represents the transfer order of blocks, and the block reproduction number (BN) 352 represents the location number of the block on the memory surface of the discard target block editing/reproduction unit 300. The block data transferred from each channel of the block editing/reproducing unit 300 to be discarded is synchronized based on the synchronization control number (CN) 351 and the block reproduction number (BN) 352. At the same time, assembling block data of multiple channels into single multiplexed data 401, synchronization control number (CN) 351 and block playback number (BN)
352, the number of channels to be multiplexed (m) 451, and additional information specifying whether or not to be discarded, and the multiplexed data 401 is transferred to the encoded header processing unit 500.

After the transfer is completed, the encoded header processing unit 500 generates a synchronization control number (CN) 351 and a block playback number (BN) 35.
2 and the number of channel multiplexes (m) 451.
The audio data 501 consisting of 01 and accompanying additional information specifying whether or not to be discarded are transferred to the audio packet a construction/disassembly unit 600. The audio packet assembly/disassembly unit 600 generates discard target designation information (Q) 35 that specifies a discard target based on additional information that designates whether or not the audio packet is to be discarded.
3 is incorporated into the audio packet header 602, and an audio packet 601 composed of the audio packet header 602 and audio data 501 is output to the packet network input/output terminal 30. Note that the additional information and the like may be transferred using a separate signal line (boat) or the same signal line.

The above describes the operation of transmitting voice packets, but
When the transmitted voice packet passes through the packet network, the packet HM accesses the voice packet through the ladder 602 depending on the packet network congestion state, and based on this, the packet HM accesses the voice packet with the discard target designation information (Q) 353. Audio packets designated as "to be discarded" may be discarded. Therefore, upon reception, voice packets designated as "to be discarded" may or may not be received.

Next, the operation of receiving voice packets will be described. According to the command from the control unit 700, the voice packet assembly/disassembly unit 60
0 receives audio packets 601 sent from the packet network via the packet network input/output terminal 30, reads the ladder 602 into the audio packets, rearranges them in the sending order at the time of transmission based on the reading result, and According to the sending order,
Audio data 501 is transferred to encoded header processing section 500. The encoded header processing unit 500 reads the encoded header 502 of the transferred audio data 501 and generates a synchronization control number (CN) 351 and a block playback number (BN) 35.
2 and the data of channel multiplexing number (m) 451 are added and multiplexed data 401 is transferred to channel multiplexing/demultiplexing section 400. The channel multiplexing/demultiplexing unit 400 has a channel multiplexing number m
The multiplexed data 401 is separated into block data for each channel based on the synchronization control number (CN) 3 for each channel.
51 and a block reproduction number (BN) 352, the block data is transferred to the discard target block editing/reproduction unit 300. The discard target block editing/reproducing unit 300 determines the storage location of block data on the memory based on the block reproduction number (BN) 352 for each channel, and stores the block data.

After the storage is completed, if the synchronization control number (CN) 351 is the initial setting number (CNo), the discard target block editing/reproducing unit 300 stores the block data having the initial setting number (CNO) on the other side of the dual memory. At the same time, the encoded data stored on one side of the dual memory that has been accumulated up to now is outputted to the audio digital encoder/decoder converter 200, analog/digital by adjusting the output timing between channels for each 8-bit code 201. Analog audio is output to the output terminal 20 via the converter 100.

On the other hand, the synchronization control number (CN) 351 is the initial setting number (C
If not (NO), the block editing/reproducing unit 3 to be discarded
00 is continuously stored in the same memory.

FIG. 4 is a configuration diagram of an audio packet, in which an audio bucket 601 is composed of an audio packet header 602 including discard target designation information 353 and audio data 501. The audio data 501 is composed of a coded header 502 including a synchronization control number 351, a block playback number 352, and a channel multiplex number 451, and multiplexed data 401.

Multiplexed data 401 includes block data equal to the number of channels multiplexed 451 (for example, 310.320°330.
340).

FIG. 5 shows an example of a blocking method when editing and reproducing blocks to be discarded. Figure 5(a) is similar to Figure 3(a).
) is supported. In the same figure (b), the blocks that are not to be discarded are divided into one block, code bits 211 and 212 of the high frequency code 210, and code bits 221, 222 and code bits 223 and 224 of the low frequency code 220, respectively. It is a method.

Also, FIG. 2(c) shows code bits 211 .
In this method, code bits 212 and 225 and 226 of the low-frequency code 220 are treated as one block to be discarded. Even if the high-frequency code 210 is discarded, it has relatively little effect on voice quality. Compared to methods (a) and (b), this method has the advantage of reducing the amount of memory and reducing the delay time during analog audio reproduction.

FIG. 6 shows a conceptual diagram of a method for reproducing block data. In this figure, two-channel stereo sound (L signal, R signal) is explained as an example. In the reproduction method, the block reproduction number BN is used as mapping information to the memory, and the synchronization control number CN is used for synchronizing a plurality of channels when reproducing analog audio.

The L, signal, and R8 signal are multiplexed to form an audio packet 601.
is created, and its block playback number, that is, the memory storage area of the discard target block editing/playback unit 300 is BN=1.
The 0th order audio packet has block playback number BN=2
The R8 signal and the R8 signal are multiplexed. Similarly, the next audio packet is 0, which is created by the R3 signal and the R3 signal.
The next audio packet is a block reproduction signal BN=4, and since the packet is to be discarded, the audio packet head is set to Q=1. This is a case where audio packets are received one after another on the receiving side, but audio packets with Q=1 are discarded and packet loss occurs. These received audio packet signals are stored on one side of the first channel dual memory as L1 signal, L2 signal, and L signal on the other side of the second channel dual memory based on the block reproduction number BN. The signals of the 0th-order received voice packets to which they are distributed are respectively stored on the other side of the first channel dual memory, and the R2 signal is stored on the other side of the second channel dual memory. These dual memories are based on the synchronization control number, that is, when CM=O arrives, the data stored up to that point is separated for each channel, and the L signal and R signal are separated.
Read out as a signal.

As described above, regarding synchronization during playback using a synchronization control number, synchronization is performed in which, upon reception of the initial setting number CN o, the encoded data accumulated up to the time of reception, excluding the received packet itself, is sent out. said. This method is effective when the packet to be discarded is sent in addition to the last one of r series of packets. Another possible synchronization method is to add the packet to be discarded to the beginning of r series of packets and send them.

This type of method is a method in which encoded data including the received packet itself is sent out upon reception of the initial setting number CN6, and this method is also included in the present invention. Furthermore, the case of one channel where m=1 can also be easily inferred from this method, and similarly rapid deterioration of voice quality when packets are discarded can be prevented. In this case, channel multiplexing/demultiplexing section 400 becomes unnecessary.

Using the embedded code as an example, we have described that a packet to be discarded is generated by accumulating code bits that do not cause rapid voice quality deterioration, but in addition to the e++bedded code, A method of distinguishing between codes of unimportant frequency bands and discarding the latter, and a method of distinguishing between audio codes (4 bits) and image codes (4 bits).
In the case of an 8-bit code that is converted into multimedia, there is a method of integrating 4 bits of the image code into a packet to be discarded.

"Effects of the Invention" As explained above, according to the audio multiplex packet transmission system according to the present invention, a packet to be discarded is formed by integrating code bits that do not cause rapid quality deterioration, for example, by utilizing an embedded code. At the same time, since the encoded data of each synchronized channel is incorporated into the same packet, even when packets to be discarded are discarded due to packet network width comparison regulations, rapid deterioration of voice quality, etc.
This has the advantage of preventing unbalanced audio quality between multiple audio channels. Furthermore, by incorporating encoded data to be synchronized into the same packet, there is an advantage that synchronization control processing can be simplified.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, FIG. 2 is a diagram showing the configuration of a G,722 8-bit code (embedded code), and FIG. 3 is a diagram showing the features of the present invention.
FIG. 4 is a diagram showing an example of the structure of an audio packet, FIG. 5 is a diagram showing an example of a blocking method, and FIG. 6 is a conceptual diagram of a method for reproducing block data. Patent applicant Nippon Telegraph and Telephone Corporation

Claims (1)

[Claims] (1) In a packet transmission method that generates encoded data for multiple channels of audio signals and assembles and disassembles audio packets using the encoded data, sudden audio loss due to omissions in the encoded data In order to form packets to be discarded by accumulating encoded bits that do not cause quality deterioration, the encoded data is stored in a certain capacity and divided into blocks to generate packets to be discarded and packets not to be discarded. It consists of an editing/reproducing means, a multiplexing/dividing means for synchronizing and assembling a plurality of block data generated for each channel into the same packet, and a block reproduction number, a synchronization control number, and a channel multiplexing number of the block data. an encoded header processing means that generates or reads a header part (encoded header) of the block data; generates an audio packet header that includes discard target designation information;
1. An audio multiplex packet transmission method, comprising means for assembling and disassembling audio packets, and when discarding packets during transmission, discarding the packets designated as being to be discarded.