JPS6347800A - Voice encoding and decoding for delta modulation - Google Patents

Abstract

(57) [Abstract] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention] (Industrial application field) The present invention relates to a method for encoding and decoding audio signals.

(Prior Art) Conventionally, the delta modulation method, which is one of the voice encoding methods, converts the difference signal between the previous sample value and the current sample value into 1 b
An attempt is made to transmit an audio signal by quantizing it with the value of it.

FIG. 3 shows an example of a conventional delta modulation method. In FIG. 3, (a) is a sample value sequence υ of the input audio signal, and (b) shows the amplitude value of the difference from the previous sample value. (C) is a code string in which, in accordance with the delta modulation method, if the difference signal in (b) is a positive value, 1 is applied, if it is a negative value, 0 is applied, and if there is no change, one is determined and applied. (In the example in Figure 3, 0 was applied.) (
d) is a signal sequence decoded based on the code sequence of (C). The step size in this case is fixed. In other words, if the code string indicates 1, the step size is Δ
(delta), +Δ, and if it indicates 0, -Δ.

In this way, even with the above-mentioned conventional delta modulation method, if the change in the sample value sequence of the input audio signal is not rapid, and the input audio signal does not change very much at all, the above code can be used. The audio signal can be decoded from the sequence.

(Problems to be Solved by the Invention) However, in the conventional delta modulation method described above, although the waveform of the original voice hardly changes, the code string is
0,1,0,1. There have been problems in that so-called granular noise occurs, in which step-like changes such as .

The present invention solves the problem of the conventional interlayer, and aims to provide a speech encoding/decoding method that suppresses the above-mentioned granular noise and overload noise.

Furthermore, the present invention is capable of suppressing the above-mentioned granular noise and overload noise by simply using an off-the-shelf Dinotal processing LSI, and is capable of transmitting audio signals with excellent quality. The purpose is to provide a decryption method.

(Means for Solving the Problems) In order to achieve all of the above objects, the present invention provides that the difference signal generated between the previous sample value and the current sample value of the input audio signal is the absolute value of the sample value at that time. This method detects the proportion of the value and determines the step size each time based on this proportion.

To this end, the speech encoding/decoding method of the present invention includes the steps of: creating a sample value sequence of an input speech signal; detecting a difference signal between a previous sample value and a current sample value of the input speech signal; a step of calculating a ratio between the difference signal and the absolute value of the sample value at that time, a step of changing a step size based on this ratio, and a step of decoding the difference signal based on the change in the step size. It is possible to follow signals with little change and signals with rapid changes in the input audio signal.

(Operation) Therefore, according to the present invention, when the degree of change in the difference signal between the previous sample value and the current sample value of the input audio signal is minute, the absolute value of the difference signal and the sample value at that time Since the step size is determined by this ratio, the width of the step size becomes narrower, so the amplitude value of the decoded audio signal has little change, and the granularity that occurs in the conventional method is reduced. The noise has the effect that it can be prevented.

Further, according to the present invention, when the degree of change in the difference signal between the previous sample value and the current sample value of the input audio signal is rapid, the ratio between the difference signal and the absolute value of the sample value at that point is , becomes larger, and the step size is determined by this ratio, so the width of the step size becomes larger, so the amplitude value of the decoded audio signal changes greatly, and the overload noise that occurs in the conventional method is reduced. , it has the effect of being able to prevent.

(Embodiment) FIGS. 1 and 2 are diagrams for explaining an embodiment of the present invention. In FIG. 1, (a) shows a sequence of sample values of an input audio signal, and (b) shows the amplitude value of the difference from the previous sample value. (C) is a time series showing whether the value of the difference signal in (b) is correct compared to the absolute value of the sample value at that time. In this example, all values greater than or equal to 100% are expressed as 100%. (d) is a code string shown by the conventional delta modulation method. In Figure 2,
(a) shows the relationship between the step size and the ratio, showing how the step size is determined according to the ratio shown in FIG. 1(c). FIG. 2(a) is a signal sequence decoded by FIGS. 1(c) and (d) based on the functions shown in FIG. 2(,).

In this manner, according to the embodiment of the present invention, the step size is determined according to the degree of change in the input signal, so that an output signal sequence that corresponds to the change in the input signal can be obtained.

In the embodiment of the present invention, the relationship between the step size and the ratio is linear up to 100%, and is uniform for values above 100%. However, if the relationship is linear up to 200%, for example, the original audio will be more accurate. A waveform close to that can be obtained. Therefore, the relationship between the ratio of the difference signal to the absolute value of the sample value and the steno knob size is not necessarily unique.

Furthermore, each means for carrying out the present invention can be easily achieved by using an off-the-shelf digital processing LSI.

(Effects of the Invention) As is clear from the above embodiment, the present invention calculates the stend size in the conventional delta modulation method by using the difference signal between the previous sample value and the current sample value of the input audio signal, and the difference signal at that point. Since the input audio signal is determined from the ratio of the absolute value of the sample value to the absolute value of the input audio signal, even if there is little change in the amplitude or there is a sudden change, the decoded audio signal can be obtained by following the input audio signal. be able to.

Furthermore, according to the present invention, since a decoded signal can be obtained by following the changing input audio signal as described above, so-called granular noise and overload noise can be prevented.

Furthermore, the present invention can produce an audio signal with better quality than the conventional delta modulation method by simply adding means for calculating the ratio between the difference signal and the absolute value of the sample value at that time. will be restored.

[Brief explanation of the drawing]

1 and 2 show one embodiment of the present invention.
In the figure, (,) is a sample value sequence of the input audio signal, (b)
is the time series of the difference signal, (c) is the ratio of the sample value of the difference signal to the absolute value, and (d) is the code string. In Figure 2, (,) is the ratio of the difference signal to the sample value vs. The normalized step size (b) shows the decoded signal sequence. Figure 3 is an example of the conventional delta modulation method.
(a> shows the sampled sequence of the input audio signal, (b) shows the time series of the difference signal, (C) shows the code sequence, and (d) shows the decoded signal sequence. ) Nte issue 11 1 101000ilOO○
10 Figure 2

Claims (1)

[Claims] A voice encoding/decoding method in delta modulation, comprising: means for creating a sample value sequence of an input voice signal; means for detecting a difference signal between a previous sample value and a current sample value of the input voice signal;
means for calculating the ratio between the difference signal and the absolute value of the sample value at that time; means for changing the step size based on this ratio; and means for decoding the difference signal based on the change in the step size. A voice encoding/decoding method in delta modulation, which is capable of following signals with little change and signals with rapid changes in an input voice signal.