JPH0564492B2 - - Google Patents

Description

Detailed Description of the Invention: "Industrial Application Field" This invention is an encoding rule identification method for mutually identifying different encoding systems existing in a communication network when audio signals are digitally encoded and transmitted. It is related to.

``Prior art'' In the future, as the digitalization of communication networks progresses, networks will be equipped with audio signals other than PCM CODEC (code/decoder).
It is expected that CODEC will be connected, sometimes transmitting audio in a frequency band more than twice the conventional telephone band.
Wideband voice CODECs, which encode at the same 64Kb/S bit rate as the PCM system, are expected to be widely used to meet the demand for better call quality.

However, for broadband voice communication services, interconnection with existing telephone services is essential.
Combining wideband CODEC with PCM encoding function,
It is necessary to switch the encoding/decoding functions depending on the type of CODEC to be used, that is, the coding rule.
At this time, the coding rule switching control methods include: (1) a method in which the network identifies and switches the terminal type via the D channel (control channel) supported by the digital network; (2) a method in which the network identifies and switches the terminal type via the D channel (control channel) supported by the digital network; Possible methods include transmitting and receiving low-level tone signals, identifying and switching terminal types based on the presence or absence of such in-band tone signals. However, these methods have the disadvantage that control is complicated and the burden on the network increases.

An object of the present invention is to provide an encoding rule identification method for realizing encoding rule switching control with less burden on the network.

"Structure of the Invention" According to the present invention, an audio signal is encoded using two types of encoding methods, and the encoded word lengths of the two types of encoding methods are the same or an integral multiple of one of the other, and In a method that selects and transmits one of the audio signals encoded using two types of encoding methods, the transmitting side transmits each code encoded using the two encoding methods within a predetermined time. For example, for one encoding rule, the most significant bit (MSB) of the encoded word, the most significant bit (MSB) of the encoded word for one encoding rule, and the According to the rules, the least significant bit (LSB) of the coded word is defined as a specific bit, and this specific bit is placed at a predetermined bit position in the transmission frame on the transmission path and transmitted. Then, the correlation value between adjacent frames within a certain period of time is detected for the bit at the predetermined bit position in the received code within one frame, and the encoding rule of the received code is determined depending on whether this correlation value is greater than or equal to a predetermined value. identify

"Embodiment" An embodiment of the encoding rule identification method according to the present invention will be described in detail below.

In different speech encoding methods (hereinafter referred to as method A and method B) in which the coded word length is L bits, the bit string of the code of the n-th coded word encoded with method A is a _0,o , Let a _1,o , a _2,o ...a _L-1,o .

a _0,o : MSB, a _L-1,o : LSB, a _i,o = 0 or 1,
i=0,1,...L-1 The bit string of the code of the nth coded word encoded by method B is b _0,o , b _1,o , b _2,o ...b _{L-1 ,o} .

b _0,o : MSB, b _L-1,o : LSB, b _i,o = 0 or 1 On the sending side, a _0,o , a _1,o , a _2,o , ...a _L-1,o and
For each bit of b _0,o , b _1,o , b _2,o ……b _L-1,o, let the bit correlation value between adjacent coded words within a certain time be a relatively large value. One specific bit is determined from a _00,o , a _1,o , ...a _L-1,o , and this is set as a _MAX , and the bit correlation value between adjacent coded words is relatively A specific bit indicating a small value is set to b ₀ ,
Determine one from b ₁ , ...b _L-1 , set it as b _MIN , and set it as the predetermined K-th (1≦K≦ L) bit position (predetermined bit position)
When using encoding method A, place a _MAX ,
Furthermore, when encoding method B is used, b _MIN is arranged and transmitted.

On the receiving side, the bit correlation value between adjacent frames within a certain time is calculated for the K-th bit of the received code of one frame, and this value is compared with a predetermined threshold to determine the encoding rule for the received code. Determine. In other words, if the calculated correlation value is larger than the threshold value, method A is determined, and if it is smaller, method B is determined.
It is determined that

To determine a _MAX and b _MIN , a _0,o , a _1,o , ... a _L-1,o and b _0,o ,
The bit correlation value between adjacent coded words may be determined for each of b _1,o . . . b _L-1,o , and the result may be statistically determined. In particular, a _MAX is the MSB of the sign, i.e. a ₀ and b _MIN is the LBS of the sign, i.e.
b It is effective to choose _L-1 . This means that in unvoiced speech signals, polarity reversals between adjacent coded words are extremely rare compared to cases where the same polarity is maintained.
On the other hand, since the LSB changes almost randomly for each coded word, the MSB has a high bit correlation value between adjacent coded words, and the LSB has a low bit correlation value between adjacent coded words.

The bit correlation value can be determined, for example, by negating the difference between a bit of a coded word and a bit of an adjacent coded word, and accumulating this for a certain period of time, for example, one second. That is, the i-th bit of the code of the n-th encoded word is a _i,o
Then, for a certain integer N, it is sufficient to find _N 〓 ⁿ⁼¹ ( _i,o+1 − _i,o ). (represents the negation of x).

Figure 1 shows a telephone band audio signal using the PCM method.
MSB of PCM code when encoded to 64Kb/S
Figure 2 shows the bit correlation value between adjacent coded words of Indicates the correlation value. The correlation value in FIG. 1 is greater than the correlation value in FIG. Therefore, on the transmitting side, in the case of PCM encoding, the MSB is placed at the same bit position of the transmission code of one frame, and in the case of subband ADPCM encoding, the LSB of the low-band ADPCM code is placed and transmitted. On the receiving side, the PCM encoding rule and the subband ADPCM encoding rule are distinguished by setting a threshold value for the cumulative correlation value of bits between adjacent frames within a certain period of time for bits at the same bit position in one frame of the received code. can.

A third example of the configuration of a coding rule identification circuit on the receiving side in a system to which the coding rule identification method of the present invention is applied is shown below.
As shown in the figure. A code bit string of code length L bits input from input terminal 1 is input to switch 2.
Switch 2 has the function of passing only the K-th bit in the code of one frame. For example, the MSB of the code is placed at the K-th bit position in the case of a code based on one encoding system, and the LSB of the code is placed in the case of a code based on the other encoding system. The bit selected by switch 2 is supplied to one frame delayer 3 and subtraction circuit 4. The subtraction circuit 4 calculates the difference between the K-th bit of the code of the previous frame, which is the output of the delay device 3, and the K-th bit of the current frame. This difference output is output by the inverting circuit 5
The output signal is supplied to the adder circuit 6 through the circuit 1, and is added to the output of the delay circuit 7. The addition output is supplied to the delay circuit 7 and delayed by one frame. circuits 6 and 7
constitutes an accumulation circuit 8, which calculates the accumulated value within a certain period of time, compares it with a threshold value in a determination circuit 9, determines its magnitude, and outputs it to an output terminal 10.

This encoding rule identification circuit can be realized using individual components such as adders, subtracters, and memories, or can also be realized by program control using a signal processing processor.

Although the above-mentioned embodiment shows an example in which the code lengths of the coded words of the two encoding methods are equal, the principle of this invention is that the code lengths of the coded words of the two encoding methods are integer multiples of each other. It can also be applied to identification methods. For example, when distinguishing between the 64Kb/SPCM encoding method (code length 8 bits) for telephone band audio and the ADPCM encoding method (code length 4 bits) for 7KHz audio, the MSB of the PCM code and 1 The LSB of the ADPCM code for every encoded word is
The bit sequence of the transmission code may be configured so that the bits are placed at the same bit position within one frame of the PCM code.

"Effects of the Invention" As explained above, the present invention can identify the encoding rules of opposing CODECs without using a special control channel or in-band control signals, so the network burden associated with encoding rule switching control is reduced. It has the great feature of being able to significantly reduce the amount of noise generated, and is extremely effective in promoting the introduction of broadband CODECs into networks.

[Brief explanation of the drawing]

Figure 1 shows telephone band audio at 64Kb/s using the PCM method.
Figure 2 shows the correlation value between adjacent coded words of the MSB of the PCM code when it is encoded using the subband ADPCM method. FIG. 3 is a block diagram showing a configuration example of a coding rule identification circuit on the receiving side when the present invention is applied. 1: input terminal, 2: switch, 3: delay circuit,
4: subtraction circuit, 5: negation circuit, 6: addition circuit,
7: Delay circuit, 8: Accumulation circuit, 9: Discrimination circuit, 1
0: Output terminal.

Claims (1)

[Scope of Claims] 1. An audio signal is encoded using two types of encoding methods, the coded word lengths of the two types of encoding methods are the same or an integral multiple of one of the other, and the two types of codes In a method that selects and transmits one of the audio signals encoded using the above two encoding methods, the transmitting side performs adjacent encoding within a predetermined time for each code encoded using the two encoding methods. One specific bit having different inter-word correlation values is defined in each coded word, and the specific bit is placed at a predetermined bit position in a transmission frame on a transmission path and transmitted. On the receiving side, the correlation value between adjacent frames within a predetermined time period is determined for the bit at the predetermined bit position in the received code within one frame, and this correlation value is compared with a predetermined threshold value. An encoding rule identification method characterized by identifying an encoding rule of a received code.