JP2660365B2 - Voice packet assembler / disassembler - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multimedia packet communication technique for transferring real-time information such as voice information in packets.

[0002]

2. Description of the Related Art FIGS. 16 and 17 show a "voice packet system" (Saito et al., Nippon Telegraph and Telephone Corporation, Tsuken Real Report, 38, N
o3, pp 231-237, 1989). In FIG. 16, voice terminals (1) such as telephones and faxes are PBXs (2)
The PBX (2) is connected to a voice PAD (4) via a TTC standard 2 Mb / s I / F (3) in which 30 channels are time-division multiplexed. The audio PAD (4) is connected to a packet multiplexer (6) via an X21 I / F (5), and the packet multiplexer (6) is connected to another packet multiplexer via a high-speed digital line (7). It is connected to the.

FIG. 17 is a diagram showing the configuration of a conventional audio PAD (4). In the figure, audio PAD (4) is I / F (3)
The PBX interface (8) connected to the PBX and the code / code provided for each channel connected to the PBX interface (8)
A decoder (CODEC) (9), a PAD unit (10) connected for each CODEC (9) and assembling and disassembling a voice packet,
Layer 2 protocol between a multiplexing unit (11) for multiplexing voice packets from a PAD unit (10) for each channel, a data memory (12), a CPU (13) and a packet multiplexing device (6) Line control unit (14) that terminates and sends and receives voice packets
And a packet multiplexing unit (15).

Next, the operation will be described. Referring to FIG. 16, a packet multiplexing apparatus receives a voice packet from another packet multiplexing apparatus via a high-speed digital line (7).
(6) transfers the voice packet to the voice PAD (4) via the X21 I / F (5). In the audio PAD (4) on the transmitting side, only the audio part of the audio signal is embedded ADPCM.
The audio data is encoded, packet interleaving is performed to packetize the encoded audio data into a packet in which the upper bits are collected and a packet in which the lower bits are collected, an audio packet is assembled, and a plurality of channels are multiplexed. ).

A detailed description of the above packet interleaving process will be described with reference to FIG.
0) to M11 to M18, block decomposition register (221) to M21
～M28, the deblocking register (222) M21 ~M38, the deblocking register (223) M41 ~M48, the deblocking register (224) writes the M51 ~M58.

Next, block disassembly registers (220) to (22)
4) Set middle bits M11, M21, M31, M41, M51 to FIFO ( 230 )
And finally the least significant bit of each register
M18, M28, M38, M48, and M58 are arranged in series in the FIFO ( 230 ) to perform packet interleave processing.

Therefore, the audio PAD (4) receiving the audio packet from the packet multiplexer (6) via the X21 I / F (5)
Decomposes a packet that collects the upper bits and a packet that collects the lower bits that are packet-interleaved for each channel, restores the encoded audio data from the data in the two packets, and encodes the encoded audio data into a CODEC ( 9) decrypts it and sends it to the voice terminal (1) via the I / F (3) and the PBX (2).

Next, the detailed operation of the voice PAD (4) will be described with reference to FIG. The line controller (14) receives the voice packet from the I / F (5) according to the layer 2 protocol and stores it in the data memory (12). The CPU (13) reads out the voice packet stored in the data memory (12) and sends it to the multiplexing unit (11). The multiplexing unit (11) separates the received voice packet based on the channel number stored in the header part, and sends it to the PAD unit (10) of the corresponding channel.

The PAD unit (10) waits for the arrival of both a packet in which the upper bits are collected and a packet in which the lower bits are collected. When the two packets arrive, the two packets are decomposed, and the data in the packet is decomposed. And restores the coded audio data from CODEC (9). CODEC (9) decodes the received encoded audio data to reproduce the audio signal,
Send to I / F (3) via XI / F (8).

[0010]

The conventional voice packet assembling / disassembling apparatus requires a PAD unit for decomposing and reproducing voice packets for each channel, so that the hardware configuration becomes large-scale and the apparatus is expensive. There was a problem of becoming.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and has as its object to provide a voice packet assembling / disassembling apparatus which guarantees the quality of voice and which is small in size and inexpensive.

[0012]

A voice packet assembling / disassembling apparatus according to the first invention collects a plurality of samples of coded voice data in which only a voice signal in a voice section is embedded-coded on a voice signal transmitting side. In the case where the collected packets are collected for each bit at the same bit position to form a plurality of blocks, and the plurality of blocks are collectively packetized to receive voice packets for a plurality of channels from a packet switching network. Calculates the sum of the time required for decomposing the audio packets and decomposing the blocks for each channel at the start timing of the predetermined time period, and arbitrarily determines that the disassembly, block decomposition, and decoding of the packets of all the channels are completed within the predetermined time period. For channels of the same channel, discard the block that collected the lower bits and collect the higher bits that are important for quality. Processing time calculation means for determining a channel for decomposing only a channel, and separating and decomposing received voice packets of a plurality of channels for each channel, restoring embedded coded voice data in the packet, and codecs of the corresponding channels. And for the channel that decomposes only the block that collected the upper bits according to the result of the processing time calculation means, decomposes only the block that collected the upper bits that are important for quality, and It is provided with a packet decomposing means for decomposing a block, restoring it into encoded voice data, and sending it to a codec.

Further, in the voice packet assembling / disassembling apparatus according to the second invention, the voice signal transmitting side collects a plurality of samples of encoded voice data obtained by embedding only the voice signal in the voice section and embedding the voice signal. For each sample collected for each bit at the same bit position and configured into a plurality of blocks, and when a plurality of channels of voice packets are received from the packet switching network for a plurality of channels, the fixed time At the start of the cycle, the sum of the time required for voice packet decomposition and block decomposition for each channel is calculated, and all channels are decomposed and decoded so that packet decomposition and decoding for all channels are completed within a fixed time period. Channel for decomposing only the block that collected the upper bits important for quality, upper bits and 1 or 2 Processing time calculation means for classifying and determining any of a channel for decomposing a block in which lower bits are collected and a channel for decomposing all blocks; and only a block in which upper bits are collected in accordance with the judgment of the processing time calculation means for each channel. For the channel that decomposes
For a channel that decomposes only the block that has collected the upper bits and decomposes the block that has collected the above bits and one or two lower bits, decomposes the block that has collected the upper bits and one or two lower bits, and For the channel (1), a packet decomposing means for decomposing all blocks and restoring them into coded voice data is provided.

Further, the voice packet assembling / disassembling apparatus according to the third aspect of the present invention encodes a packet which cannot be discarded, such as a push button signal, and an encoding method other than the embedded encoding of a modem, a fax signal, etc. In the case of receiving both the packet that cannot be discarded and the audio signal packet that can be discarded in a block, at a certain time period timing, the sum of the time required for the disassembly and block disassembly of each channel is calculated, Packets that cannot be discarded and packets that are not discarded and encoded using a method other than embedded coding are used so that packet disassembly, block disassembly, and decoding for all channels are completed within a certain time period. For channels other than the channel that is receiving a possible signal, discard the block that collects the lower bits, and Processing time calculation means for determining a channel for decomposing only the blocks for which the upper bits have been collected, and for the channel for decomposing the blocks for which the upper bits have been collected, only the blocks for which the upper bits have been collected, according to the judgment of the processing time calculation means. And packet decomposing means for decomposing all blocks and restoring the coded audio data by decomposing all blocks for the other channels.

Further, the voice packet assembling / disassembling apparatus according to the fourth aspect of the present invention encodes a packet which cannot be discarded, such as a push button signal, and a coding method other than the embedded coding of a modem, a fax signal, etc. in those receiving both of destruction and discarding impossible packets and blocks disposable voice signal packets, when the start of the predetermined time period, to end the packet decomposition and decoding of all the channels at a given time period In addition to the channels that are receiving signals that cannot be discarded, such as modems, faxes, and push-button signals, channels that collect only the high-order bits that are important for quality into coded voice data and channels that are important for quality For restoring the block that collects the most significant bits and some of the lower blocks into coded audio data. And the processing time calculation means for determining a channel, according to the judgment of the processing time calculation means, only for decomposing the channel block to collect the upper bits,
For the channel for decomposing only the block in which the upper bits are collected and the block in which the upper bits and 1 or 2 lower bits are collected, the upper bits and 1 or 2 are used.
And a packet decomposing means for decomposing a block in which the lower bits are collected and decomposing all blocks to decompose all blocks to restore encoded voice data.

[0016]

In the voice packet assembling / disassembling apparatus according to the first aspect of the present invention, at the start timing of a certain time period, the processing time calculating means calculates the total time required for decomposing the voice packet for each channel and decomposing the blocks. In order to shorten the processing time so that the packet decomposition and block decomposition of all channels are completed within a certain time period, the number of channels and the number of channels for decomposing only the blocks in which upper bits important for quality are collected are calculated. And the packet disassembly means
For all channels, the voice packets received for each channel are decomposed.

For a channel determined by the above means to decompose only the block in which the high-order bits have been collected, the embedded coded voice data is restored only from the block in which the high-order bits in the packet have been collected. With regard to (5), the embedded coded audio data is restored from all the block data, and the embedded coded audio data corresponding to the audio signal for a predetermined time period is transmitted to the codec.

Further, in the voice packet assembling / disassembling apparatus according to the second aspect of the present invention, a channel for disassembling only a block in which upper bits important in quality are collected for all channels, an upper bit and lower bits of 1 or 2 are collected. Processing time calculating means having a function of classifying and deciding into a channel for decomposing a divided block and a channel for decomposing all blocks, calculates a total sum of time required for decomposing audio packets and decomposing blocks for each channel. In order to complete decomposition and block decomposition of all channels within a certain time period, collect only the blocks that have collected the upper bits important for quality and the number of channels, and collect the blocks and lower bits that have collected the upper bits Channel and number of channels for decomposing one or two blocks, and channels for decomposing all blocks Determining Le and the number of channels, respectively.

In accordance with the determination of the above means, the block decomposing means decomposes only the block from which the upper bits have been collected for the channel which decomposes only the block from which the upper bits have been collected, and decomposes the block from which the upper bits have been collected and the lower bits. For a channel for decomposing one or two collected blocks, a block in which upper bits and one or two lower bits are collected is decomposed, and for a channel in which all blocks are decomposed, all blocks are decomposed to coded audio data. Restore.

The voice packet assembling / disassembling apparatus according to the third aspect of the present invention encodes a packet which cannot be discarded, such as a push button signal, and a coding method other than the embedded coding of a modem, a fax signal, etc. Processing time calculation means having a function of determining a channel for decomposing only a block in which upper bits important in quality are excluded from channels other than a channel which is reproducing a non-disposable packet which is the calculated degradation and total time required to block degradation of voice packets, as degradation and deblocking of packets of all channels within a certain time period is completed, the block discard impossible packet <br/> bets and For channels other than the channel that is playing back non-discardable packets encoded using a method other than embedded coding, , To determine the decomposing channels only blocks collected significant upper bit quality.

The packet decomposing means decomposes only the block from which the upper bits have been collected for the channel which decomposes the block from which the upper bits have been collected in accordance with the judgment of the processing time calculating means for the channel which can discard the block. For the channel of, all blocks are decomposed, restored to coded voice data, and transmitted to the codec.

The voice packet assembling / disassembling apparatus according to the fourth aspect of the present invention encodes a packet such as a push button signal, which cannot be discarded, and an encoding method other than the embedded encoding of a modem, a fax signal, etc. for channels except the channel being reproduced waste <br/>棄不possible packets, channel degrades only the blocks collected significant upper bit quality for all channels, the lower bits of the upper bits and 1 or 2 Processing time calculation means having a function of classifying and determining the channel into which the block into which the collected data is collected and the channel which decomposes all the blocks, calculates the sum of the time required for decomposing the audio packet and the block decomposition for each channel. Calculate and calculate the upper order so that packet decomposition and block decomposition of all channels are completed within a certain time period. Tsu DOO channel degrades block only collect, upper bits and 1 or channel degrades block a collection of lower bits of 2, classifies determined channel degrades all blocks.

According to the judgment of the processing time calculation means for a channel in which block discarding is possible, for a channel in which only the block in which the upper bits are collected is decomposed, only the block in which the upper bits are collected is decomposed. Alternatively, for a channel that decomposes a block in which two lower bits are collected, a block in which the upper bits and one or two lower bits are collected is decomposed. For a channel in which all blocks are decomposed, all blocks are decomposed.
Restore to coded audio data.

[0024]

An embodiment of the first invention will be described below with reference to the drawings. In FIG. 1, (16) shows a case where a plurality of samples of embedded ADPCM encoded voice data are collected for each bit at the same bit position to form a plurality of blocks, and a voice packet obtained by packetizing the plurality of blocks together is A packet decomposing unit that decomposes and restores encoded audio data of a plurality of samples, (17) is a two-port memory.

FIG. 2 shows a packet format used in the embodiment of the present invention.
9) is the data part of the voice packet.
(190) is an M1 block that collects the MSB, (191) is an M2 block that collects the MSB-1,
Similarly, (192) is an M3 block, (193) is an M4 block,
(194) is the M5 block.

FIG. 3 is a block diagram of the packet decomposing portion (16), In the figure, (20) performs control of Write-packet decomposing read packets from the two-port memory (17) CPU, (2
1) is a processing time calculating means for calculating the sum of the processing time for each channel, (22) is a block decomposition register for decomposing a plurality of blocks in which bits at the same bit position are collected, and restoring encoded data of a plurality of samples, Reference numeral (23) denotes a FIFO for holding a plurality of samples of the coded data restored by the block decomposition register (22) and transmitting the coded data to the CODEC (9) at a fixed timing, and is prepared for each channel.

Reference numeral (24) denotes a control information FIFO for storing speech / silence discrimination and encoding information data, which is prepared for each channel similarly to the FIFO (23), and transmits encoded data from the FIFO (23). And sends it to CODEC (9).

FIG. 4 is a diagram showing a method of decomposing a data portion of a voice packet into blocks in the packet decomposing unit (16). FIG. 5 is a diagram showing a timing chart of the packet decomposing unit (16) and the CODEC (9) and a timing chart of the I / F (3).

FIGS. 6 and 7 show only the blocks in which the upper bits have been collected, which are used in the present embodiment, so that the audio packet disassembly processing for all channels accommodated within a certain time period is completed. 5 is an example of an algorithm for determining a channel to be restored to data and the number of channels. FIG.
FIG. 9 is a conceptual diagram showing the effect of shortening the block decomposition processing time by the processing time calculation means (21) in the present embodiment.

Next, the operation will be described. In FIG. 1, a line control unit (14) receives a voice packet transmitted from a packet multiplexing device (6) via an I / F (5) and writes it into a data memory (12). And the packet multiplexing unit (15)
The CPU (13) transfers the voice packet from the data memory (12) to the two-port memory (17). The packet decomposer (16)
Receives voice packets of multiple channels stored in port memory (17), separates voice packets for each channel,
The voice packet is decomposed for each channel and transmitted to the CODEC (9) of the corresponding channel.

FIG. 2 shows the format of the voice packet used in this embodiment. The header (18) stores bits indicating whether block discarding has been performed in the packet network, a sequence number, and the like. In addition, the blocks that collected bits important for quality were the M1 block (190) and the M2 block (191), and the other M3 blocks (192), M4 blocks (193), and M5 blocks (194) were missing. Even M1
(190) and M2 (191) blocks guarantee the minimum necessary quality.

FIG. 3 is a diagram showing a block configuration of the packet decomposing unit (16).
Receives voice packets of multiple channels stored in port memory (17), separates voice packets for each channel,
The voice packet is decomposed for each channel, and one byte is taken out from each block of the packet and transmitted to the block decomposing register (22). The block disassembly register (22) stores the coded audio data after the block disassembly in a FIFO provided for each channel. This operation is performed at regular time intervals.

The operation of the block decomposition register (22) will be described with reference to FIG. The CPU (20) calculates M11 to M18, M21 to M28, M31 to M38,
M41 to M48 and M51 to M58 are taken out and written into the corresponding registers of the block disassembly register (21). That is, M11 to M1
8 to the block disassembly register (220), M21 to M28 to the block disassembly register (221), M31 to M38 for the block disassembly register (222), and M41 to M48 for the block disassembly registers (223) and M51 to M58. The block decomposition register (2
24) Write to.

After the writing is completed, the CPU (20) creates control information (sound, coding type information),
When the start instruction is performed by setting the control information in the start port of (22), the block disassembly register (22) sets M11,
The M21, M31, M41, and M51 bits are taken out, written into the FIFO (230) of the corresponding channel (channel 1 in this embodiment), and then the same processing is performed on M12 to M52... M18 to M58 for eight samples. Is written into the FIFO (230).

The interface between the CODEC (9) and the packet decomposer (16) has a frame structure with a period of 125 μS as shown in FIG. 5 and identifies the channel at the time slot position in the frame. When the time slot corresponding to the channel number is reached, one sample of coded voice data is transmitted from the FIFO (230) to the CODEC (9). The CODEC (9) decodes the received coded audio data of one sample and sends out 8-bit PCM data to a time slot corresponding to the channel number of TTC2MI / F.

In order to transmit one sample of encoded audio data from the FIFO (23) every 125 μS and to restore eight samples of encoded audio data by one activation of the block decomposition register (22), CPU (20) is 125 μS × 8 = 1
Within mS, it is necessary to take out the next one byte from each block of the voice packet and write it to the block disassembly register (22).

If no voice packet has arrived at the two-port memory (17), the block disassembly register (22)
The control information (voice / silence discrimination = silence) is set to the startup port without writing data to the, and the CODEC (9) is instructed to output background noise.

On the other hand, the processing time calculation means (21) has a CODEC
A synchronous 1 mS interrupt is inserted at the beginning of the frame of the interface between (9) and the packet decomposing unit (16). Upon the interruption, the processing time calculation means (21) determines whether there is a sound (a packet to be reproduced arrives in the two-port memory (17)) for each channel, a sound (no packet), or a sound. In the case of, whether or not the block is discarded in the packet network is determined for all the channels based on the header (18) of the voice packet, and the previously calculated voice packet disassembly processing time in each case is calculated. The number of channels for which block discarding is performed is determined by comparing the sum of the values for the channels and a certain time period (1 mS in this embodiment). The determination method will be described with reference to FIGS.

First, the channel pointer is set to 1 in S1, and the total processing time (T _TOTAL ) for all channels is set to 0 in S2. Then S3
Then, it is determined whether or not the channel indicated by the channel pointer is silent during the time period. If no, the process proceeds to S4, and if there is a sound, the process proceeds to S5. In S4, the silence processing time is added to the total processing time (T _TOTAL = T _TOTAL + T _N ).

On the other hand, in S5, it is determined based on the header information whether or not the packet to be reproduced during the corresponding time period of the corresponding channel has been discarded in the packet switching network. Add the packet replay processing time when the block is discarded to the total processing time (T _TOTAL = T
_TOTAL + T _BD), the process proceeds to S7 Without block discard, adds the packet reproduction processing time without block discard _{(T TOTAL = T TOTAL + T} BND) the total processing time. In S8, the channel pointer is incremented, and it is determined whether or not the totalization of the processing time for all channels has been completed (S9) .If all channels have been completed, the process proceeds to S10.
Step S7 is repeated.

At S10 in FIG. 7, the channel pointer is set to 1 again.
It is determined whether or not the total processing time (T _TOTAL ) for all channels calculated in S11 to S9 in S11 is equal to or longer than a predetermined time period (T _A = 1 ms). If T _TOTAL > T _A , block discard processing is performed. the process proceeds to step S12 to perform, because in the case of T _{TOT AL} ≦ T _a there is sufficient capacity to start the block decomposition of the voice packet exits the flowchart. In S12, it is determined whether or not there is no packet to be reproduced for the channel indicated by the channel pointer (silence processing is being performed).

Then, the process proceeds to step S16 if no sound is being processed, and to step S13 if no sound is being processed. It is determined whether or not the packet reproduced by the channel indicated by the channel pointer has been discarded in the network. If S16
If not, the process proceeds to S14.

In S14, it is determined that the packet of the corresponding channel is to be discarded, and the total processing time T for all channels is determined.
_{From TOTAL} , the difference between the disassembly processing time (T _BND -T
_BD ) Subtract the total processing time T _{TOTAL for} all channels
And proceed to S16.

In S16, the channel pointer is incremented, and in S17, it is determined whether or not processing of all channels has been completed.
If not completed, the process returns to S11, and the processes of S11 to S16 are performed.
Repeat until _TOTAL ≦ T _A, starts blocks degradation of voice packets through the flowchart if finished.

When exiting this flow at S17, T _TOTAL > T _A is _maintained and the processing time is short. However, the number of channels to be processed is limited based on the processing capacity of the CPU (20). , S17 are set in advance so as not to escape from the flow.

The CPU (20) determines from the result of the determination made by the processing time calculating means (21) based on the flowcharts shown in FIGS.
And data of M2 block only
Write in 2).

In that case, CODEC (9) contains 25 in FIG.
It is transmitted in the bit pattern shown in b. When the data of the M1 to M5 blocks are written into the block decomposition register (22), they are transmitted in the bit pattern of 25a.

The effect of using the processing time calculation means (21) will be described with reference to FIG. FIG. 8A shows all channels (FIG.
In this example, the packet disassembly process is performed without block discard for eight channels). In this case, the processing capability of the packet disassembly unit (16) is insufficient, so that a certain time period is exceeded. On the other hand, when the block B is discarded for all eight channels, the CPU (20) and the block disassembly register (22) process only the data of the M1 and M2 blocks. .

Therefore, the maximum number of channels is the maximum number of channels within a fixed time period when the packet disassembly processing time when all channel blocks are discarded. On the other hand, in C, since there are a plurality of silent channels, a sound channel is completed within a certain time period without discarding blocks.

[0050] Also, since the packet decomposing time in D exceeds the certain time period, FIG. 6, the packet channel 1,2,4 accordance with the flowchart of FIG. 7 performs block disposal, street constant shown in E suppress obtain the processing time within the time period.

In the embodiment of the first invention, FIG.
In S14, the processing time calculation means (21) decides to discard the packet of the corresponding channel, and the CPU (2
In (0), according to the judgment of the processing time calculation means (21), only the M1 and M2 blocks in which the upper bits important in quality are collected are restored to the encoded voice data.

However, as shown in the second invention, the processing time calculating means (21) uses M1 and
The processing time may be shortened by determining a channel for restoring the M2 block and a part (M3 or M3 + M4) of the block in which the lower bits are collected into the encoded audio data.

FIGS. 9 and 10 show the processing flow of the processing time calculation means (21) in that case, and S1 to S9 are the same as those in FIG. M5 is set to the block to be discarded first in S10 of FIG. The processing of S11 to S18 is almost the same as S10 to S17 of FIG. Different locations determines the disposal of M _i blocks in S15, the processing time when discarded until M _i blocks when not blocked discarded and the M5 blocks from the total processing time T _TOTAL of all the channels in S16 difference T _BND - _Subtract T _BDM1 ,
This is the part that updates T _TOTAL .

[0054] except the channel of block disposal in silence and reticuloendothelial for all channels, also discard the M _i blocks for the channel, the total processing time T _TOTAL of all channels
If the processing time exceeds a certain processing time period T _A , as shown in S19, the blocks to be discarded are expanded to M4 and M3, and the M4 and M3 blocks are discarded until the processing time is within the certain time period. It may be.

Further, as in the third invention, when a packet which cannot be discarded such as a modem signal, a fax signal and a push button signal is received, the processing time calculating means is used.
(21), as shown in the processing flowcharts of FIGS. 11 and 12, except for the channel receiving the packet whose block cannot be discarded (S13), discard the M3 to M5 blocks in which the lower bits are collected (S14). , S15), the processing time may be shortened.

Further, as shown in FIG. 13 to FIG. 15, the processing time of all the channels is constant except for the channel receiving the block non-discardable packet (S14). M1 that collects the upper bits important for quality so that it is less than the time T _A
M2 block and part of the block that collects lower bits (M3
Alternatively, (M3 + M4) may be determined to be restored to the encoded audio data (S15, S16, S17), and the CPU (20) may perform block decomposition according to the result to reduce the processing time.

[0057]

As described above, according to the first aspect of the present invention, the channel and the number of channels for decomposing only the block in which the upper bits important for quality are collected are determined, and only the block in which the upper bits are collected is decomposed. By restoring the embedded coded voice data from only the block in which the upper bits in the packet are collected for the channel determined to be performed, the packet disassembly processing time can be shortened.
Therefore, the channel that can be accommodated in the voice packet assembling / disassembling device
Can increase the multiplicity of the
The structure of the packet assembler / disassembler can be reduced in size and cost
A device with good performance can be obtained.

According to the second aspect of the present invention, for a channel for decomposing only a block in which upper bits important in quality are collected, only a block in which upper bits are collected is decomposed, and a block in which upper bits are collected and a lower bit are collected. For a channel that decomposes one or two blocks from which bits have been collected, decompose the block that collects the high-order bits and one or two low-order bits. For a channel that decomposes all blocks, decompose all the blocks. By restoring to data, in addition to the effect of the first invention, decryption
This has the effect of improving the quality of the reproduced audio .

According to the third aspect of the present invention, a packet such as a push button signal which cannot be discarded and a packet which is coded using a coding method other than the embedded coding such as a modem signal and a fax signal and cannot be discarded are used. For the channels excluding the channel that is playing back, determine the channel that decomposes only the block that collected the upper bits that are important for quality, and for the channel that decomposes the block that collected the upper bits, only the block that collected the upper bits By decomposing all blocks for the other channels and restoring the encoded audio data, in addition to the effect of the second invention, the encoded data is encoded by the embedded encoding method.
Signals that cannot be discarded and embedded codes
Signal that cannot be discarded because it is encoded using an encoding method other than
Issue quality can be guaranteed.

[0060]

According to the fourth aspect of the present invention, non-discardable packets, such as push-button signals, which cannot be discarded and modems, fax signals, etc., which are coded using a coding method other than the embedded coding, are used. Regarding channels other than the channel for which the packet is being reproduced, a channel for decomposing only a block in which upper bits important in quality are collected for all channels, a channel for decomposing a block in which lower bits of upper bits 1 or 2 are collected, and all blocks are referred to. For a channel that is classified into any of the channels to be decomposed and decomposes the block in which the upper bits are collected, only the block in which the upper bits are collected is decomposed and the block in which the lower bits of upper bits 1 or 2 are collected is a block that is decomposed. Is a block that collects the high-order bits and one or two low-order bits. In addition to the effects of the third invention, by decomposing all blocks and decomposing all blocks and restoring the coded audio data, a channel for decomposing all blocks is used. Can be packetized and transmitted, so if the load is constant,
It is possible to maintain a certain quality, and to obtain a voice packet assembling / disassembling apparatus having good cost performance and voice quality.

[Brief description of the drawings]

FIG. 1 is a block diagram of a voice packet assembling / disassembling apparatus showing an embodiment of the first invention.

FIG. 2 is a format diagram showing a format of a voice packet used in the embodiment.

FIG. 3 is a block diagram of a packet decomposing unit.

FIG. 4 is an explanatory diagram showing a method of decomposing a data portion of an audio packet into blocks in a packet decomposing unit (16).

FIG. 5 is a timing chart showing a timing chart of the packet decomposing unit (16) and the CODEC (9) and a timing chart of the I / F (3).

FIG. 6 shows a channel and a channel for decomposing only blocks in which upper bits are collected and restoring encoded data so that voice packet decomposing processing of all channels accommodated within a certain time period used in the present embodiment is completed. FIG. 9 is an explanatory diagram illustrating an example of an algorithm for determining the number.

FIG. 7 shows a channel for decomposing only a block in which upper bits are collected and restoring encoded data so that voice packet decomposing processing for all channels accommodated within a certain time period used in the present embodiment is completed. FIG. 4 is an explanatory diagram illustrating an example of an algorithm for determining the number of channels.

FIG. 8 is a conceptual diagram showing the effect of reducing the block processing time by the processing time calculation means (21) in the present embodiment.

FIG. 9 is a processing flow chart of a processing time calculation means (21) in one embodiment of the second invention.

FIG. 10 is a processing flowchart of a processing time calculating means (21) according to the embodiment of the second invention.

FIG. 11 is a processing flowchart of a processing time calculating means (21) according to an embodiment of the third invention.

FIG. 12 is a processing flowchart of a processing time calculating means (21) according to the third embodiment of the present invention.

FIG. 13 is a processing flowchart of a processing time calculating means (21) according to an embodiment of the fourth invention.

FIG. 14 is a processing flowchart of a processing time calculating means (21) according to an embodiment of the fourth invention.

FIG. 15 is a processing flowchart of the processing time calculating means (21) according to the fourth embodiment of the present invention.

FIG. 16 is a system configuration diagram showing a conventional voice communication system.

FIG. 17 is a configuration diagram showing a configuration of a conventional voice packet assembling / disassembling apparatus.

[Explanation of symbols]

(1) Voice terminal (2) PBX (3) TTC standard I / F (4) Voice PAD (5) X21 I / F (6) Packet multiplexer (7) High-speed digital line (8) PBX interface (9) CODEC (10) PAD section (11) Multiplexing section (12) Data memory (13) CPU (14) Line control section (15) Packet multiplexing section (16) Packet decomposing section (17) 2-port memory (18) Voice packet (19) A block that collects bits at the same bit position in multiple samples of embedded ADPC encoded audio data (20) CPU (21) Processing time calculation means (22) Block decomposition register (23) FIFO (24) Control information FIFO (25) One sample of coded audio data

 ──────────────────────────────────────────────────続 き Continuation of the front page (72) Inventor Masatsugu Yano 5-1-1, Ofuna, Kamakura City Mitsubishi Electric Corporation Communication Systems Laboratory (56) References JP-A-2-7643 (JP, A) Hei 2-205141 (JP, A) JP-A-61-156949 (JP, A)

Claims (4)

(57) [Claims] An audio signal transmitting side collects a plurality of samples of encoded audio data obtained by embedding only an audio signal in an audio section and collects all the collected samples for each bit at the same bit position. A voice packet assembling / disassembling apparatus configured into a plurality of blocks and receiving the voice packets obtained by packetizing the plurality of blocks for a plurality of channels from a packet switching network. Calculate the sum of the time required for decomposition and block decomposition, and discard the block in which lower bits have been collected for any channel so that the decomposition and block decomposition and decoding of packets of all channels are completed within a certain time period. During processing to determine the channel that decomposes only the block that collected the upper bits important for quality Inter-computing means, and separates and decomposes the received voice packets of a plurality of channels for each channel, restores the embedded encoded voice data in the packets, and sends them out to the codecs of the corresponding channels,
According to the determination result of the processing time calculation means, for a channel that decomposes only the block that collected the upper bits, decomposes only the block that collected the upper bits that are important for quality, and decomposes all blocks for the other channels. An audio packet assembling / disassembling apparatus, comprising: packet decomposing means for restoring encoded audio data and sending it to a codec. 2. A voice signal transmitting side collects a plurality of samples of coded voice data obtained by embedding only a voice signal in a voice section and collects all the collected samples for each bit at the same bit position. A voice packet assembling / disassembling apparatus configured into a plurality of blocks and receiving the voice packets obtained by packetizing the plurality of blocks for a plurality of channels from a packet switching network. The sum of the time required to decompose the audio packet and the block decomposition was calculated, and the upper bits important for quality were collected for all the channels so that the decomposition and decoding of the packets for all the channels could be completed within a certain time period. The channel for decomposing only the block, and the block where the upper bits and one or two lower bits are collected are separated. Processing time calculation means for classifying and determining any of the channel to be resolved and the channel for decomposing all blocks; and a channel for decomposing only the block of which upper bits have been collected, according to the judgment of the processing time calculation means for each channel, For a channel that decomposes only the block that has collected the upper bits and decomposes the block that has collected the above bits and one or two lower bits, decomposes the block that has collected the upper bits and one or two lower bits, and A voice packet assembling / disassembling apparatus, which comprises packet decomposing means for decomposing all blocks for each of the channels (1) and (4) to restore coded voice data. 3. A push-button signal block discard possible packets and modem etc., using an encoding scheme other than Emubededdo encoding such facsimile signal encoded waste <br/>棄不possible packets and blocks In a voice packet assembling / disassembling apparatus which receives both discardable voice signal packets, at a certain time period timing, a total time required for decomposing and decomposing a voice packet for each channel is calculated, and a certain time period is calculated. So that the disassembly and block disassembly and decoding of the packets of all channels are completed within
For channels other than channels that are receiving non-discardable packets and non-discardable signals encoded using a method other than embedded coding, discard the block that collects the lower bits and discard the upper bits that are important for quality. Processing time calculation means for determining a channel for decomposing only the block in which the bits have been collected, and according to the determination of the processing time calculation means,
For channels that decompose blocks that have collected the upper bits, only the blocks that have collected the upper bits are decomposed,
A voice packet assembling / disassembling device comprising: packet decomposing means for decomposing all blocks for other channels to restore encoded voice data and sending the data to a codec. 4. A push-button signal block discard possible packets and modem etc., using an encoding scheme other than Emubededdo encoding such facsimile signal encoded waste <br/>棄不possible packets and blocks In a voice packet assembling / disassembling apparatus which receives both discardable voice signal packets, a modem, a facsimile apparatus and a facsimile apparatus are provided so that at the start of a certain time period, packet decomposition and decoding of all channels are completed within a certain time period. For channels other than the channel that is receiving a block non-discardable signal such as a push button signal, a channel for restoring only a block in which high-order bits important for quality are collected into encoded voice data and a high-order bit for quality Processing time calculation to determine the collected block and the channel for restoring some lower blocks to coded audio data For the channel that decomposes only the block in which the upper bits are collected according to the determination of the stage and the processing time calculation means, only the block in which the upper bits are collected is decomposed, and the block in which the upper bits and one or two lower bits are collected. Packet decomposing means for decomposing a block in which high-order bits and 1 or 2 low-order bits are collected for a channel for decomposing, and decomposing all blocks for a channel for decomposing all blocks to restore encoded voice data. Voice packet assembly
Disassembly device.