JPS61284133A - Waveform interpolation method - Google Patents

Abstract

PURPOSE:To suppress unnatural sound due to long time repetition of a waveform by applying weighting to an output waveform, attenuating gradually an output signal starting interpolation and restoring gradually the output signal when the interpolation is finished. CONSTITUTION:A variable attenuator 11 uses a control signal inputted from an error detector 12 to weight an output signal. The relation of y(t)=k(t)x(t) is established, where x(t) is an input signal to an attenuation section 11, y(t) is an output signal and k(t) is a weight coefficient at the attenuation section 11, and when the detection section 12 detects a decode disable error in an input data from an input terminal 8, a waveform repetition section 14 starts interpolation. Then the attenuation section 11 applies weighting to squeeze the output signal gradually and when the input data is restored normally, the interpolation is finished. In this case, the attenuation section 11 applies weighting as shown in equation, where t2 is an end time of interpolation and A2 is a constant, and the amplitude of the output signal is restored gradually.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a waveform interpolation method for interpolating waveforms in sections that cannot be decoded due to transmission errors in digital transmission of waveforms such as audio signals.

BACKGROUND OF THE INVENTION FIG. 3 is a block diagram showing the configuration of an apparatus for implementing a conventional waveform interpolation method. In FIG. 3, 1 is an input terminal, and this input terminal 1 is connected to a decoding section 2 and an error detection section 4. The decoding section 2 is connected to the switch 3 and the waveform storage section 5. Switch 3 is connected to output terminal 7. The error detecting section 4 is connected to the waveform storage section 5, the waveform repeating section 6, and the switch 3. The waveform storage section 5 has a waveform edge turning section 6.
The waveform repeating section 6 is further connected to the switch 3.

Next, the operation of the above conventional example will be explained. In FIG. 3, in a normal state, the switch 3 connects the decoding section 2 and the output terminal 7, and data input from the input terminal 1 is decoded by the decoding section 2 and outputted to the output terminal 7. At the same time, the output of the decoding section 2 is temporarily stored in the waveform storage section 5. The stored contents of the waveform storage section 5 are updated sequentially in the normal state. When data containing an error is input to the input terminal 1, the error detection section 4 detects the error and determines that the decoding section 2 cannot decode it, the switch 3, the waveform storage section 5, the waveform A control signal is sent to the edge turning section 6. When the error detection section 4 detects an error, the waveform storage section 5 stops inputting from the decoding section 2 and inputs the waveform stored so far to the waveform edge return section 6. The waveform edge repeating section 6 repeats the waveform input from the waveform storage section 5 to generate a signal for interpolation.

The switch 3 connects the waveform repeating section 6 and the output end 7 according to a command from the error detecting section 4, and interpolates the output of the repeated waveform produced by the waveform repeating section 6.

FIG. 4 (to) shows the output waveform when no error occurs in the conventional example, and FIG. 4 (B) shows the output waveform when an error occurs and interpolation is performed. Note that G) in FIG. 4 CB) indicates an interval in which interpolation is performed.

In this manner, the conventional waveform interpolation method described above can interpolate noise generated due to undecodable transmission errors to the immediately preceding waveform, thereby preventing large noise from appearing in reproduced audio.

The problem that the invention seeks to solve However, in the conventional waveform interpolation method described above.

When errors in input data occur continuously, interpolation using repeated waveforms continues for a long time, causing the problem that the reproduced audio becomes an unnatural continuous sound.

The present invention solves these conventional problems,
It is an object of the present invention to provide an excellent waveform interpolation method that can suppress the occurrence of continuous sounds in the output even when errors occur continuously in input data.

Means for Solving the Problems In order to achieve the above object, the present invention weights the output waveform of the conventional example, gradually attenuates the output waveform during the interpolation period, and gradually reduces the amplitude of the output waveform when the interpolation is completed. This is to make it return to normal.

Therefore, according to the present invention, since the amplitude of the output waveform is reduced when interpolation continues for a long time, it is possible to suppress the output of unnatural continuous sounds when errors in input data occur continuously. .

Embodiment FIG. 1 shows the configuration of an embodiment of the present invention. In FIG. There is. The decoding section 9 is connected to the switch 10 and the waveform storage section 13. The switch 10 is connected to a variable attenuation section 11. The variable attenuation section 11 has an output end 15 .

It is connected to the. In addition, the erroneous detection section 12 is connected to the waveform storage section 1.
3, the waveform edge reversing section 14, the switch 10, and the variable attenuation section 11. The waveform storage section 13 is connected to a waveform repeating section 14, and the waveform repeating section 14 is connected to the switch 10.

Next, the operation of the above embodiment will be explained. In the above embodiment, the variable attenuation section 11 weights the output signal according to the control signal input from the error detection section 12.

One method of weighting is shown below using a formula. The input signal to the variable attenuation section 11 is x(t), and the output signal is y(
υ, the weighting in the variable attenuation section 11 expresses the relationship k (t)
Then, the output signal y(1) is y(t)
It is expressed as = k (t) x (t). In a normal steady state, k(t)=1, and the variable attenuation section 11 outputs the signal decoded by the decoding section 9 as it is to the output terminal 15. When the error detection section 12 detects an undecodable error in the input data from the input terminal 8, the waveform repeating section 14 starts interpolation. Assuming that the interpolation start time is H and A1 is a constant coefficient, the variable attenuation section 11 performs weighting as shown below to gradually reduce the output signal. When the input data from the input terminal 8 returns to normal, the interpolation ends. Assuming that tlA2 is a constant coefficient for the interpolation end time, the variable attenuation section 11 performs weighting as follows, and gradually restores the amplitude of the output signal.

FIG. 2 shows an example of waveform interpolation according to the above embodiment.

The second country is the output waveform when no error occurs, Figure 2 C
B) is an output waveform when an error occurs and interpolation is performed. A in FIG. 2 is the section where interpolation is performed.

As described above, according to the above embodiment, when the data inputted from the input terminal 8 contains a series of undecodable errors, the variable attenuation section 11 reduces the output signal by repeating a series of waveform repetitions. It has the advantage of being able to suppress unnatural continuous sounds caused by Further, according to the above embodiment, when the error rate of the input from the input terminal 8 is very large, the average attenuation amount in the variable attenuation section 11 becomes large, and it is possible to suppress noisy output due to decoding of erroneous data. Can be done.

Effects of the Invention As is clear from the above embodiment, the present invention gradually attenuates the output signal when interpolation starts and gradually restores it when interpolation ends. This has the advantage of suppressing unnatural sounds caused by time repetition. Furthermore, when the error rate of input data is very large, the amount of attenuation to the output is large on average, which has the effect of suppressing the output of noise due to decoding of erroneous data.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a device that implements the waveform interpolation method according to an embodiment of the present invention, FIGS. 2A and B are output waveform diagrams according to the same embodiment, and FIG. 3 shows a conventional waveform interpolation method. A block diagram of the apparatus to be implemented, and FIGS. 4A and 4B are output waveform diagrams of the apparatus. 8... Input end, 9... Decoding section, 10... Switch, 11... Variable attenuation section, 12... Error detection section, 13
... Waveform storage section, 14... Waveform repetition section, 15...
・Output end. Name of agent Written by patent attorney Toshio Tsuchinakao and one other person 1 Figure 2 Figure 3 Figure 4

Claims (1)

[Claims] Detects whether there is an error that cannot be decoded in the input data, decodes the input data and stores the decoded signal if there is no error, and stops decoding if there is an error and repeats the stored decoded signal. A waveform interpolation method characterized in that the amplitude of the decoded signal is gradually restored when the input data is free from undecodable errors.