JPH0769717B2 - Speech encoder and decoder - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a waveform coding method, which is one of the coding methods for audio signals, and particularly to adaptive delta modulation.
ation (hereinafter abbreviated as ADM) to a speech coder and a speech decoder in an encoding system.

(Prior art and its problems) The ADM coding system is less affected by code errors during transmission, has easier synchronization than other coding systems, and is small and has low power consumption. It is applied to mobile radio.

FIG. 4 shows a conventional encoder 22 and decoder 23 of the ADM coding system.
It is a block diagram of. In the figure, is the analog input signal of the encoder, is the digital output signal encoded by the encoder 22, is the digital input signal of the decoder 23, and is the analog output signal decoded by the decoder 23. Encoder 22
, 1 is a subtractor (DEC), 2 is a quantizer (QTZ),
3 is an encoder (DGT), 4 is a step width controller (STP),
Reference numeral 5 is an adder (ADD), and 6 is a delay element (D). The digital output signal encoded by the encoder 22 is, for example,
It is sent to a wireless or wired transmission line at a transmission rate of 16kbps. Decoder 23 on the receiving side is a decoder (DE
T), 9 is a step width controller (STP), 11 is an adder (AD
D) and 12 are delay elements (D), which are decoded into analog waveforms by the reverse operation of the encoder 22 and output as analog output signals. As the data, a binary value corresponding to the number of quantized bits is used except for the digital transmission / reception data.

FIG. 2 is a waveform example diagram for explaining the operating principle of the circuit of FIG. The horizontal axis represents time and the vertical axis represents amplitude. 2 and 4
The basic operation will be described with reference to the drawings.

The analog voice input signal which is the input of the subtracter 1 is converted into X (n
T). The prediction signal Y (nT) obtained by quantizing the time-divided output Z (nT) from the subtractor 1 bit before by the quantizer 2 and adding it by the adder 5 is delayed by 1 bit by the delay element 6. It is input to the subtractor 1 with a delay Y (nT-T) by the time T. Here, comparison and subtraction are performed with the current input X (nT). Therefore, the output Z (nT) of the subtractor 1 is expressed by the following equation.

Z (nT) = X (nT) -Y (nT-T) This output Z (nT) and the step width Δ (nT) output from the step width controller 4 are added to the quantizer 2 and the quantizer 2 The output Q (nT) of 2 and the output B (nT) of the encoder 3 are Z
Depending on whether (nT) is positive or negative, it becomes as follows.

When Z (nT) ≧ 0 Q (nT) = + Δ (nT), B (nT) = 1 When Z (nT) <0 Q (nT) = − Δ (nT), B (nT) = 0 Sign The output B (nT) of the encoder 3 is applied to the step width controller 4, and the output step width Δ (nT) is increased or decreased by the output B (nT) of the continuous encoder 3.

That is, as shown in FIG. 2, the curve and the stepwise waveform Y (nT) are compared for each bit in which the analog voice waveform is time-divided, and repeated at delay time T and output as a transmission digital signal.

The predicted signal Y (nT), which is the output from the adder 5, has the same waveform as the decoded signal output from the decoder 23 described later, which is a condition for obtaining an output close to the original waveform.

On the other hand, the digital input signal of the decoder 23 is
0, the step width controller 9, the adder 11, and the delay element 12 perform the reverse processing of the encoder 22, and an analog output signal, that is, Y (nT) is output and a voice signal is reproduced.

Step width of encoders 22 and 23 Step width of controllers 4 and 9 Δ (n
For example, the value of T) becomes 1.4 times the step width Δ (nT) when the same code is continuously applied in a section of 3 to 4 bits, and becomes 0.98 times when the different code is used.

Now, problems in the conventional circuit based on the above-described operation principle will be described below.

FIG. 3 is a waveform example diagram for explaining the operation of the encoder 22 and the decoder 23. The horizontal axis represents time t, and the vertical axis represents amplitude. In the figure, the stepped waveform of the input audio signal is the output waveform of the receiving side decoder when the circuit and the transmission line are normal. When the transmission output digital signal obtained by encoding the input voice signal is [11011010100 ...], a code error was received during transmission because the condition of the transmission line was bad. When the bit of is changed from 0 to 1, in the decoder, the amplitude at that bit is increased and the output waveform becomes like the waveform a, and when it becomes an analog voice signal, it largely changes compared with the original waveform and the voice deteriorates. It shows that you do.

As described above, the conventional circuit configuration has a drawback that when a code error occurs during transmission between the digital transmission / reception data and the transmission path, the amplitude varies depending on the error occurrence state. If a code error occurs such that "0" of the transmission signal changes to "1" during transmission, there is a problem that the audio audibility is greatly deteriorated.

(Object of the Invention) An object of the present invention is to provide a speech coder and a speech decoder in the ADM coding system, which suppresses the deterioration of speech audibility due to an increase in amplitude when a code error occurs during transmission as described above. To do.

(Structure and Operation of the Invention) The present invention compares the signal Y (nT) decoded bit by bit with k times the step width Δ (nT) (k is arbitrarily set to about 2 to 4). The main feature is that Y (nT) is configured to be attenuated when the output signal Y (nT) is larger than Δ (nT) × k.

The present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram showing the circuit configuration of an embodiment of the present invention. In the figure, 7 and 13 respectively arranged in the encoder 20 and the decoder 21 are comparative attenuators (LM) for exhibiting the features of the present invention. 1 to 6 of the other encoders 20,
9 and 12 in the decoder 21 are the same as those of the conventional circuit shown in FIG.

FIG. 5 is a circuit block diagram showing an embodiment of the comparative attenuators 7 and 13. 14 is an amplifier for multiplying the amplitude of the input signal by k, 15
Is a comparator that compares two input voltages and outputs only the larger one, 16 is an attenuator, and 17 is a switch. That is, the comparison attenuators 7 and 13 input the output step width Δ (nT) of the step width controller 4 or 9 from one of the input terminals to amplify the amplifier.
Predicted signal (in the case of encoder) and decoded signal (in the case of decoder) input from the other terminal by multiplying the voltage by 14 by k
The absolute value of the amplitude of Y (nT) is compared with the comparator 15 and the switching device 17 is switched to perform the following attenuation control and output.

When | Y (nT) | ≧ Δ (nT) × k, the switch 17 is switched to c and Y (nT) is changed to Y (nT) × P for output.

When | Y (nT) | <Δ (nT) × k, the switch 17 is switched to d and Y (nT) is output as it is.

However, k is a coefficient arbitrarily set within a range of 2 to 4. Further, P is a coefficient for attenuating Y (nT), and is set to a value smaller than 1 and in the range of 0.8 <P <0.99.

This setting condition is that the predictive signal Y (nT) input to the comparison attenuator 7 in the encoder 20 and the comparison attenuator in the decoder 21
In order to make the decoded signal Y (nT) input to 13 have the same waveform, it is set in both comparison attenuators 7 and 13, respectively.

In FIG. 3, the waveform of Y (nT) mainly represents the positive part, and the code “0” changes to “1” and the amplitude increases,
When Y (nT) exceeds Δ (nT) × k, the comparison attenuator 13 operates to attenuate Y (nT) P times. However, when the sign "1" should come in and "0" comes in, the comparison attenuator
13 does not work. Since Y (nT) is controlled in terms of absolute value, it is clear that the same effect can be obtained even in the negative portion of the waveform.

Therefore, when a code error occurs during transmission between the digital data and the transmission path, the decoded signal has the effect of attenuating the amplitude increased positively or negatively due to the code error.

In FIG. 3, when a code error occurs during transmission and the third bit of the received input data becomes "1" where it should have been "0", the conventional circuit produces a decoded waveform a, and according to the present invention. In the circuit, the decoded waveform b is obtained. Here, the case where k = 3.0 and P = 0.9 is shown.

As described above, when the code error occurs in the direction in which the amplitude increases in either the positive or negative range of the received waveform, the conventional circuit causes a considerable amplitude increase like the waveform a, but according to the present invention. It can be seen that in the circuit, the amplitude increase is suppressed and a decoded waveform similar to the original waveform can be obtained as in the waveform b.

If no code error occurs, this comparison attenuator
13 has no effect.

(Effects of the Invention) As described in detail above, by carrying out the present invention, it is possible to reduce the disturbance of the decoded waveform due to the code error of the transmission line with a lot of noise and the like. When used for a line with a large error, it is possible to realize an economical speech decoder with little deterioration in sound quality by simply adding a small number of circuit parts to the conventional circuit, which has a great effect.

[Brief description of drawings]

FIG. 1 is a block diagram of a circuit configuration showing an embodiment of the present invention, FIG. 2 is a waveform example diagram showing the operating principle, FIG. 3 is a waveform example diagram for explaining the operation, and FIG. 4 is a conventional circuit. FIG. 5 is a block diagram showing a configuration, and FIG. 5 is a block diagram showing an embodiment of a comparative attenuator which is a main part of the present invention. 1 ... Subtractor, 2 ... Quantizer, 3 ... Encoder, 4,9
…… Step width controller, 5,11 …… Adder, 6,12 …… Delay element, 7,13 …… Comparison attenuator, 14 …… Amplifier, 15 …… Comparator, 16 …… Attenuator, 17 ...... Switcher, 20,22 …… Encoder, 2
1,23 …… Decoder.

Claims (2)

[Claims] 1. A speech coder for coding an analog input signal by an adaptive delta modulation coding method, wherein a prediction signal obtained by quantizing and adding the analog input signal is delayed by one bit and a step width controller. Of the step width output and the absolute value of the amplitude of the signal delayed by one bit and the step width output are multiplied by k (k = 2 to
4) The amplitude of the predicted signal is compared when the absolute value of the amplitude of the signal delayed by 1 bit is compared with the amplitude of the signal obtained by 1 bit and exceeds the amplitude of the signal obtained by multiplying the step width output by k. A speech coder, comprising a comparison attenuator that multiplies P times (0.8 <P <0.99). 2. A voice decoder for receiving a signal digitized by an adaptive delta modulation coding method and decoding it into an analog signal, wherein a signal obtained by delaying the decoded output signal by one bit and a step width control. For each bit of the received digital signal, the absolute value of the amplitude of the signal delayed by one bit and the amplitude of the signal obtained by multiplying the step width output by k (k = 2 to 4) are input. When the absolute value of the amplitude of the signal delayed by one bit exceeds the amplitude of the signal obtained by multiplying the step width output by k, the amplitude of the decoded output signal is multiplied by P (0.8 <P < 0.99) A speech decoder comprising a comparison attenuator such as