JPH098670A - Data transmission method and data transmission device - Google Patents

Abstract

(57) [Abstract] [Object] To provide a data transmission method and device for performing highly efficient data transmission using a communication network including a code converter. [Configuration] A data transmission device 3 (10) according to the present invention is
A data pattern set that can pass through the communication network (4 to 8) is predicted, and a different bit pattern is assigned to each data pattern of the set based on a predetermined rule. Then, the data pattern corresponding to the same bit pattern to be transmitted is output to the communication network. The receiving side predicts a data pattern that can pass through the communication network in the same manner as the transmitting side, and assigns a bit pattern to each data pattern based on the same rule as the transmitting side. Then, the bit pattern assigned to the same data pattern as the received data pattern is output.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data transmission method and apparatus, and more particularly to a data transmission method and apparatus for performing data transmission using a communication network including a code converter such as an ADPCM encoder and a decoder. It is a thing.

[0002]

2. Description of the Related Art Conventionally, 64k in a digital telephone network
Although bps (8 bits x 8 kHz) PCM signals were used, 32 kbps (4 bits x 8 kHz) ADPCM signals will be used in international communication lines and PHS systems in order to use the lines efficiently. Has become. In this ADPCM system, since the 8-bit PCM signal is compressed to 4 bits by utilizing the correlation characteristic of the audio signal on the time axis, the reproduced PCM signal does not completely match the original signal and other than the audio signal. When random data is input, the degree of mismatch becomes very large. Therefore, when data is normally transmitted using the ADPCM line, the data is transmitted by converting the data into a voice band analog signal using a modem and converting the data into an ADPCM signal.

[0003]

In the conventional data transmission system as described above, since an analog modem is used, a transmission rate of at most about 9.6 kbps can be obtained, and there is a problem that the transmission efficiency is poor. there were. SUMMARY OF THE INVENTION It is an object of the present invention to provide a data transmission method and apparatus for improving the above-mentioned problems of the prior art and performing high-efficiency data transmission using a communication network including a code converter. .

[0004]

According to the present invention, a data pattern set that can pass through a communication network is predicted, and different bit patterns are assigned to each data pattern of the set based on a predetermined rule. The data pattern corresponding to the same bit pattern as the bit pattern to be transmitted is output to the communication network. The receiving side predicts a data pattern that can pass through the communication network in the same manner as the transmitting side, and assigns a bit pattern to each data pattern based on the same rule as the transmitting side. The bit pattern assigned to the same data pattern as the received data pattern is output.

[0005]

The present invention predicts a data pattern set that can pass through a communication network, assigns different bit patterns to each data pattern of the set based on a predetermined rule, and determines a bit pattern to be transmitted. A data pattern corresponding to the same bit pattern is output to the communication network. The receiving side predicts a data pattern that can pass through the communication network in the same way as the transmitting side, assigns a bit pattern to each data pattern based on the same rule as the transmitting side, and receives the data pattern. The bit pattern assigned to the same data pattern as is output. Therefore, in a communication network of any method, it is possible to transfer data close to the limit of the amount of information that can be transmitted through the communication network without any modification to the communication network.
It enables extremely high-speed transmission compared to the conventional one.

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a block diagram showing an example of a communication network to which the present invention is applied. Digital telephone 1 or 9
And the data transmission device 3 or 10 to which the data terminal 2 or 11 such as a personal computer is connected is the digital exchange 4
Alternatively, it is housed in 8. The digital exchange 4 or 8 is a well-known 64 kbps (8 bits x 8 kHz) PC
Exchange M signals. The ADPCM converter 5 or 7 is
The PCM signal at the relay terminal (A or B) of the digital exchange 4 or 8 and the relay line 6 by a known method.
The above 32 kbps (4 bits × 8 kHz) ADPCM signal is mutually converted. The data transmission device 3 or 10 according to the present invention converts the data received from the data terminal into a 64 kbps signal by the method described later, decodes the received 64 kbps signal into the original data, and outputs the original data to the data terminal.

FIG. 6 is a block diagram showing the configuration of the data transmission device 3 (or 7) of the present invention. Based on the program stored in the ROM 31, the CPU 30 controls the entire data transmission device and performs transmission / reception processing described later. In addition to the programs, the ROM 31 also stores a bit pattern table as shown in FIG. The RAM 32 is used as a work area during processing or as a buffer for transmitted / received data. The terminal interface circuit 33
For example, it is a bidirectional data transfer circuit such as an RS-232C interface circuit, and exchanges transmission / reception data with a data terminal such as a personal computer. The line interface circuit 34 is an interface circuit of the digital subscriber line of the electronic exchange 4, and transmits / receives a signal of 64 kbps according to, for example, the well-known ISDN line specifications. The bus 35 connects each circuit in the data transmission device.

FIG. 2 is a block diagram showing the configuration of a well-known ADPCM encoder 28 and decoder 29. Adder 2
0 subtracts the prediction value which is the local reproduction value one sampling cycle before stored in the predictor 23 from the 8-bit input signal and outputs the prediction error signal. The adaptive quantizer 21 quantizes and encodes the prediction error signal to generate a 4-bit output signal. At this time, the quantization step width is adaptively controlled based on the previous output signal value and the like. The output signal is output to the transmission line 6 and is also transmitted to the quantizer 2 by the inverse quantizer 22.
A process reverse to that of 1 is performed, and a locally decoded error signal is output. The adder 24 adds the error signal and the output of the predictor 23 to update the predicted value. In the decoder 28 on the receiving side, the inverse quantizer 25, the predictor 27, and the adder 26 perform the same processing as the inverse quantizer 22, the predictor 23, and the adder 24 in the encoder 27 to reproduce. Get the value. Therefore, the values of the two predictors 23 and 27 are usually the same.
Note that such an ADPCM system is well known, and in the present invention, any quantization and coding system can be used.

FIG. 3 illustrates the principle of the present invention.
Input value to ADPCM encoder (point A in FIG. 1 or FIG.
3 is a graph showing the relationship between the input signal) and the reproduction value (point B in FIG. 1 or output signal in FIG. 2) output from the ADPCM decoder. The values of the predictors 23 and 27 in FIG. 2 are both S. When the input value is in the vicinity of S, the reproduction value becomes almost the same value.
When the input value is (S + 10) or more, all the reproduction values become (S + 11) when the value becomes -10) or less. Therefore, as shown in the figure, there are only 15 possible reproduction values when all patterns are input as input values. Since the most significant bit is used as the sign bit in the actual 8-bit PCM data, the graph has a characteristic different from that of the graph of FIG. 3, but the present invention can be applied to an encoder having an arbitrary characteristic.

In the first embodiment, every pattern is input as an input value to the local encoder-decoder for every transmission cycle to obtain all transmittable reproduction values, and when there are 15 reproduction values, , 15 playback values are assigned different bit patterns. Then, the reproduction value to which the bit pattern that matches the pattern of the data to be transmitted is assigned is transmitted. At the receiving side, all transmittable playback values are obtained by the same processing as at the transmitting side. If there are 15 playback values, different bit patterns are applied to 15 playback values by the same processing as at the sending side. assign. Then, the bit pattern assigned to the same reproduction value as the received reproduction value is output as actual data. ADPCM
If the signal is 4 bits, the maximum is 16 depending on the encoding method.
Although there is a possibility that a plurality of reproduction values will be obtained, when 1 bit is used as the sign bit, “+0” and “−0” have the same reproduction value, so that the maximum reproduction value is 15. Also, when the predicted value approaches the maximum or minimum possible value, the number of reproduction values decreases.

FIG. 5 is a diagram showing a bit pattern set assigned by each reproduction value number obtained by using the local encoder-decoder, and the data in this figure is stored in advance in both the transmitting and receiving data transmission devices. Has been done. In addition,
A number called a label is given to each bit pattern, and the order of giving is arbitrary.

FIG. 4 is an explanatory diagram showing how to allocate bit patterns to an arbitrary number. If the number is a power of 2 such as 2, 4 or the like, all patterns having corresponding bit numbers are sequentially assigned. In other cases, for example, 5
In the case of the number of bits, "00" is expanded into "000" and "001" of the lower L rows out of the four patterns of the K rows to form a bit pattern as shown in FIG. In this way, different playback patterns may be assigned different bit pattern lengths. Further, when the appearance frequencies of the bit patterns are biased, the transmission efficiency is improved by preferentially allocating the pattern having the higher appearance frequency.

FIG. 8 is a flowchart showing a transmission process in the CPU 30, which is executed every transmission cycle. In step S1, the reproduction value buffer is cleared and the variable n is reset to 0. In step S2, a prediction value stored in a predetermined address in RAM is subtracted from n to obtain a prediction error. And the difference is A
Adaptive quantization is performed by the same method as the DPCM encoder. That is, the possible range of the error signal value is divided into a plurality of ranges based on the quantization step width based on the immediately preceding output signal, and which range the error signal falls into is determined.

In step S3, the reverse process of step S2 is performed, and the local reproduction error which is a value representative of the range is output. In step S4, the stored predicted value and the local reproduction error are added to obtain the local reproduction value.
In step S5, it is determined whether or not the obtained local reproduction value matches any of the reproduction values stored in the reproduction value buffer. If the result is negative, the reproduction value is labeled. And save it in the playback value buffer. The labels may be added in order, for example, 0, 1, 2, ... In step S7, it is determined whether or not n has become 255, and if the result is negative, step S8
At n, 1 is added, and the process returns to step S2 to repeat the process.

In step S9, the number of reproduction values in the reproduction value buffer is counted. In step S10, for example, a bit pattern set corresponding to the number of reproduction values counted is read out from a table storing values as shown in FIG. In step S11, the read bit pattern set is searched for a bit pattern that matches the leading bit pattern of the bit string of the data to be transmitted, and the label number of the bit pattern is obtained.
In step S12, the local reproduction value corresponding to the label number obtained in step S11 is read from the reproduction value buffer, stored in the RAM as a predicted value, and output to the line as a transmission signal. This output value is the ADP
Both the predictors are adaptively quantized by the CM encoder as in step S2, encoded and transmitted, and dequantized by the ADPCM decoder as in steps S3 and S4 to generate a reproduction value. If there is a transmission error, the output value and the reproduction value on the receiving side will always match.

FIG. 7 is a flow chart showing the receiving process in the CPU 30. This process is also executed for each reception cycle. In step S20, the same processing as the processing of steps S1 to S9 in FIG. 8 is performed to obtain the reproduction value and the number thereof. In step S21, of the reproduction values in the reproduction value buffer, the label of the reproduction value that matches the received reproduction value is obtained. In step S22, the table having the same contents as the transmitting side is selected in step S2.
A bit pattern in which the number obtained in 0 and the label obtained in step S21 match is read and output as received data (added after the previous received data).
In step S23, the received reproduction value is stored in the RAM as a predicted value.

FIG. 9 is an explanatory diagram showing an example of data in the transmission process. If the predicted value stored in the RAM is 100, local reproduction values are calculated for all input values (n), and if duplication is eliminated, the 15 reproduction values shown in the reproduction value column of FIG. The value remains. Label numbers 0 to 14 are assigned to the reproduction values in the order of generation, for example. Since the number of reproduction values is 15, the bit pattern set corresponding to the number of reproduction values 15 is read from the table of FIG. Since a label is added to the bit pattern set, the reproduction value and the bit pattern have a one-to-one correspondence.
Next, a bit pattern that matches the beginning of the bit string of the actual data is searched. For example, the first part of the actual data is 01
101001 ..., label 6 “011
The bit pattern "0" matches. Therefore, a reproduction value of 99 corresponding to the bit pattern is output to the communication network.

On the receiving side, the data in the table as shown in FIG. 9 is generated by performing the same processing as on the transmitting side. When the received data is 99, the corresponding bit pattern is "0110". RAM as received data
Output to the internal buffer. Then, the stored received data is transferred to the data terminal device, for example, every 8 bits.
High-speed data transmission becomes possible by the above processing.

Next, measures against data errors will be described.
In the communication network as shown in FIG. 1, the pattern and data length of the received data may be erroneous due to an error during transmission. Further, as described above, the ADPCM system transmits only an error signal with respect to the predicted value, and therefore, when the predicted values on the transmitting side and the receiving side do not match due to a transmission error or the like, the error is continuously resolved. It has the characteristic that it is not done.

In the system of the present invention, the ADP on the receiving side
The predicted value (output value) in the CM decoder and the predicted value in the receiving side data transmission device always match, and since the bit pattern corresponds to each of the transmittable error signals, the predicted value is 1
If there are five areas, the data itself is likely to be transmitted correctly, even if the predicted values deviate to some extent. However, when the predicted value approaches the maximum value or the minimum value, the number of playback values may differ due to the deviation of the predicted value. In this case, the bit length of the transmitted data and the received data may change as the patterns do not match. It is possible that the bit lengths of will not match.

As a countermeasure against this, for example, a specific synchronization pattern in actual data, for example, "11001100110" is used.
The data is processed so that the same pattern as "0" does not appear (insert "1" before the last "0" of the same pattern in the data). The synchronization pattern is inserted at a predetermined interval, for example, every 1000 bits to obtain actual data. Then, the synchronization pattern in the received data is monitored, and if the data length is not a predetermined interval, the data is regarded as an error, and the data after the synchronization signal is received as new data. Or HD
A transmission control procedure may be adopted in which a fixed-length or variable-length packet added with error check information, such as an LC frame, is used and a retransmission request is made in the case of an error.

Further, in order to eliminate the deviation of the predicted value, for example, before or during the data transmission, from the transmission side data transmission device to the communication network, only for a period of time or more necessary for the predicted value to reach the maximum value. , Enter the maximum value you can enter,
After that, the minimum value is input only for a period of time or more required for the predicted value to reach the minimum value, and 0 (intermediate value) is further input for a predetermined period. Then, whichever predicted value on the receiving side deviates, it finally becomes 0, and the predicted values between the transmitting and receiving devices match.

Although the embodiment has been disclosed above, as a result of computer simulation, as a result of the method of the present invention, 30 kb is obtained.
It was confirmed that a transmission rate of ps or more can be achieved. In addition, the following modifications may be considered in the present invention. Although an example in which the present invention is applied to a communication network via an ADPCM encoder is disclosed as an embodiment, the present invention is applicable to a communication network via any code conversion method. If the encoding method is such that the output can be determined irrespective of the internal state of the encoder, the process of obtaining the local reproduction value and its number is executed only once in advance and the result is stored. Good. Furthermore,
Correct data transmission is possible even when the output signal of the data transmission device is directly input to the reception side data transmission device without passing through the encoder in the communication network.

As a result of the simulation, the number of reproduced values that occurred was all within the range of 11 to 15.
This is because, for example, even if the predicted value reaches the maximum value, there are always eight reproduced values on the minus side including the maximum value. Therefore, at least 3 bits (8 bits) can always be transmitted, and the bit pattern allocation may be fixed to 3 bits. In this case, the transmission rate is 32 kbps x (3 /
4) = 24 kbps, but there is almost no possibility of erroneous reception data length. Moreover, the error rate is further reduced by assigning the bit patterns to a plurality of adjacent reproduction values in a duplicated manner. In the table, only the number of reproduction values that can occur, for example, in the case of the ADPCM method as in the embodiment, only the bit pattern sets for 8 to 15 may be stored.

In the system of the embodiment, the bit patterns are assigned in ascending order of the local reproduction value. In this case, if the actual data is random, the probability that the predicted value (reproduction value) will randomly increase and decrease and is biased to one side is low, but if the appearance frequency of 0 of the actual data is more than 1, If the allocation as in the example is performed, the probability that the predicted value decreases toward the minimum value increases, and the error rate may increase due to the bias. As a measure against this, for example, in order to make the average value of the prediction error as close to 0 as possible,
Change the bit pattern allocation of, or compress the actual data, encrypt it, or make it redundant for error checking and correction so that the frequency of occurrence of any bit pattern is as uniform as possible. It is possible to convert the code.

In the embodiment, all the bit patterns (n) that can be input are input in order to obtain the local reproduction value, but as can be seen from FIG. 3, except for the vicinity of the predicted value S, All have the same playback value. In the present invention, the range that is known to have the same reproduction value in advance is 1
Since it is sufficient to obtain the reproduction value only for the number of times, if the characteristic is as shown in FIG. 3, in the flowchart of FIG. 8, the initial value of n is set to (S-10) in step S1 and the end determination condition of step S7 is n = It may be (S + 10). In this way, the processing load on the CPU 30 is reduced.

[0027]

As described above, the present invention predicts a data pattern set that can pass through a communication network, and determines different bit patterns for each data pattern of the set based on a predetermined rule. A data pattern corresponding to the same bit pattern as the one to be assigned and transmitted is output to the communication network. The receiving side predicts a data pattern that can pass through the communication network in the same way as the transmitting side, assigns a bit pattern to each data pattern based on the same rule as the transmitting side, and receives the data pattern. The bit pattern assigned to the same data pattern as is output. Therefore, in any type of communication network, it is possible to transfer data close to the limit of the amount of information that can be transmitted through the communication network without any modification to the communication network. There is an effect that the data transmission can be performed.

[Brief description of drawings]

FIG. 1 is a block diagram showing an example of a communication network to which the present invention is applied.

FIG. 2 is a block diagram showing a configuration of a known ADPCM encoder and decoder.

FIG. 3 is a graph showing the relationship between the input value to the ADPCM encoder and the reproduction value output from the ADPCM decoder.

FIG. 4 is an explanatory diagram showing how to assign a bit pattern to an arbitrary number.

FIG. 5 is a diagram showing a bit pattern set assigned for each number of reproduction values.

FIG. 6 is a block diagram showing a configuration of a data transmission device of the present invention.

FIG. 7 is a flowchart showing a receiving process in the CPU 30.

FIG. 8 is a flowchart showing a transmission process in the CPU 30.

FIG. 9 is an explanatory diagram showing an example of data in a transmission process.

[Explanation of symbols]

1, 9 ... Telephone, 2, 11 ... Data terminal, 3, 10 ... Data transmission device, 4, 8 ... Digital exchange, 5, 7 ... AD
PCM encoder, 6 ... Relay line

Claims (5)

[Claims] 1. A data transmission method including the following processing steps (1) to (6) for transmitting data using a communication network including code conversion means. (1) A step of predicting a data pattern set that can pass through the communication network. (2) Assigning different bit patterns to each data pattern of the predicted set based on a predetermined rule. (3) A step of outputting a data pattern corresponding to the same bit pattern to be transmitted to the communication network. (4) A step of predicting a data pattern that can pass through the communication network on the receiving side in the same manner as the transmitting side. (5) A step of assigning a bit pattern to each data pattern based on the same rule as the transmitting side. (6) A step of outputting a bit pattern assigned to the same data pattern as the received data pattern. 2. A data transmission method including the following processing steps (1) to (8) for transmitting data using a communication network including an encoder and a decoder. (1) On the transmission side, a data pattern that can be input to the communication network is input to a transmission side local encoder having the same characteristics as the communication network, and the output of the transmission side local encoder has the same characteristics as the communication network. Input to the decoder to obtain the output pattern set of the transmitting local decoder. (2) A step of counting the number of output patterns that do not overlap and assigning a number to each of the output patterns according to a predetermined rule. (3) A bit pattern set corresponding to the counted number is read out from a table in which bit pattern sets that are different from each other and are given numbers for each pattern number and equal to the number of patterns are stored in advance. Determining a bit pattern that matches the data to be transmitted. (4) A step of searching for an output pattern having the same number as the number corresponding to the determined bit pattern and outputting it to the communication network. (5) On the receiving side, all data patterns that can be input to the communication network are input to the receiving side local encoder having the same characteristics as the communication network, and the output of the receiving side local encoder is received having the same characteristics as the communication network. Input to the side local decoder to obtain an output pattern set of the receiving side local decoder. (6) A step of counting the number of non-overlapping output patterns and assigning a number to each of the output patterns according to a predetermined rule. (7) Retrieving an output pattern equal to the received data and determining the number corresponding to the output pattern. (8) A step of reading out a bit pattern corresponding to the counted number and number from a table in which the same bit pattern set as that on the transmitting side is stored in advance and outputting it as received data. 3. The encoder and the decoder are AD respectively.
A PCM encoder and an ADPCM decoder, wherein each of the processing steps (1) to (8) according to claim 2 is repeated for each data transfer cycle. 4. A data transmission device for transmitting data using a communication network including an encoder and a decoder, which has the same characteristics as the communication network and encodes a data pattern that can be input to the communication network. A transmitting side local encoding means having, a transmitting side local decoding means having the same characteristics as the communication network for decoding the output of the transmitting side local encoding means, and an overlap from the output pattern of the transmitting side local decoding means After exclusion, a numbering means for counting the number and assigning a number to each of the output patterns according to a predetermined rule, and a number different from each other for each output pattern and equal to the number of output patterns Of the bit pattern set stored in advance and the bit pattern set corresponding to the counted number are read out, and the bit pattern set corresponding to the beginning of the data to be transmitted is read out from the table. A data transmission device comprising: a determining unit that determines a bit pattern; and an output unit that retrieves an output pattern assigned with the same number as the number corresponding to the determined bit pattern and outputs the output pattern to a communication network. . 5. A data transmission device for transmitting data using a communication network including an encoder and a decoder, which has the same characteristics as the communication network and encodes a data pattern that can be input to the communication network. Receiving side local encoding means having, decoding the output of the receiving side local encoding means, receiving side local decoding means having the same characteristics as the communication network, and duplication from the output pattern of the receiving side local decoding means After the elimination, the numbering means for counting the number and assigning a number to each of the output patterns in accordance with a predetermined rule, and a bit pattern set having the same content as the transmitting side data transmission device for each output pattern are stored in advance. The table means, the determining means for searching the output pattern equal to the received data, and determining the number corresponding to the output pattern, and the counting means for counting the output pattern. A data transmission device comprising an output means for reading a bit pattern corresponding to a number and a number and outputting it as received data.