JPS5868339A - Waveform encoding system on adpcm system - Google Patents

Abstract

PURPOSE:To prevent malfunction by providing an encoding circuit with a decision level transmitting circuit, and comparing an input decision level with a voice level from an A/D converter and stopping the output of an up/down signal transmitting circuit when a sampled value is greater than the decision value. CONSTITUTION:A voice waveform from an input terminal 10 is sampled by an A/D converting circuit 1 to convert the waveform into a digital signal, which is latched by a latch circuit 3. The last sampling output of the circuit 3 is compared by a comparing circuit 4 with the output of the circuit 1 and their difference data is compressed by a difference data compressing circuit 5 according to the level of an adequate level transmitting circuit 7. The output of the circuit 5 is applied to an up/down signal transmitting circuit 8 and the outputs of the circuits 7 and 8 are summed up by an adding circuit 6 to control the circuit 7 by the output of the circuit 6. A decision level transmitting circuit 9 is added to this encoding circuit to compare a decision level from an input terminal 13 with the voice level from the converter 1, thereby stopping the output of the circuit 8 when the sampled value is greater than the decision level.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to improving the waveform encoding method in the ADPCM method used to efficiently compress audio data. Differential PCM (DP
When transmitting normal audio signals, the DPCM method is replaced by the PCM method, which improves the S/N ratio by about 6 dB. In addition, in the general DPCM system, an adaptive DPCM called an ADPCM system is used in which the quantization step is determined according to the amplitude of the signal, taking into account that the amplitude of the audio signal changes significantly. ADP-C
By using the M method, both large-amplitude and small-amplitude audio signals can be transmitted with the same S/N ratio. However, when the difference data value of the transmission signal one sample before the adaptive level of the signal to be transmitted is larger than the level-up setting value, the adaptive level of the signal to be transmitted can be set one step higher than the adaptive level of the transmission signal one sample before (for example, quantization Step is doubled), and if it is smaller than the level down setting value, the adaptive level of the transmission signal #1 sample before is reduced by 1 step (for example, the quantization step is reduced to 1/2).
The ADPCM method, which performs adaptive compression of differential data, has the disadvantage that it cannot represent the differential data of the transmitted signal, which is formed by the relationship between the audio waveform and the sample points, and this results in large errors.

For example, in the audio waveform shown in Figure 1, AlBXC
, D are the sample points, and A', B', σ, and U are the difference data at the sample points A, B, and CSD, respectively.The adaptation level at the sample point is because the difference data C is very small. (less than the level down setting value), the adaptation level is decreased by one step. Therefore, although the difference data U is the same as the difference data, the adaptation level is one step smaller than that of the sample point B, so that it cannot represent the difference boot, and a large error occurs.

FIG. 2 is a block diagram showing a conventional ADPCM system.

The operation shown in FIG. 2 will be explained. Transmission signal waveform input terminal 1
The audio waveform inputted from 0 is sent to the sampling signal generator (CG) in the analog-digital switching circuit (An)
) 2 at the sampling period, converted into a digital signal, and sent to the soft circuit 4 and latch circuit 3. The latch circuit 3 holds the digital signal of one sample before and sends it to the comparison circuit 4. Comparison circuit 4 is ADCI
The difference between the signal of Y and the signal of latch circuit 3 is sent to the differential data compression circuit 5 as differential data. The differential data compression circuit 5 compresses the differential data according to the adaptive level from the adaptive level output path 7 and sends it to the adaptive pressure #I differential data output terminal 11 and also to the up/down signal sending circuit 8. The up-down signal sending circuit 8 compares the level-up or level-down setting value applied to the up-down judgment level input terminal 12 with the compressed difference data, and the compressed difference data determines the level-up or level-down setting. The 0 adaptive level sending circuit 7 sends a signal that increases the adaptive level by one step if the value exceeds the value and increases by one step, or a signal that decreases the adaptive level by one step if the value decreases. However, depending on the magnitude of the difference data and the direction of its increase or decrease according to the output of the adder circuit 6, the adaptive level is changed and sent to the difference compression circuit 5.

In the conventional ADPCM method explained above, when the difference data is obtained by sampling the audio waveform as shown in FIG. If it is made too small, a large error will occur.

SUMMARY OF THE INVENTION An object of the present invention is to set a judgment level in order to eliminate the above-mentioned drawbacks, and to prevent the means for reducing the adaptation level by one step when the sample value exceeds the judgment level to cope with a large change in differential data. In other words, the level indicated by Since the value exceeds the judgment level prevent.

FIG. 3 is a block diagram of a circuit showing an embodiment of the present invention.

The operation shown in FIG. 93 will be explained. Parts with the same functions as those in Figure 2 are numbered the same. Figure 3 is the same as Figure 2 with a judgment level sending circuit 9 added, and the judgment level sending circuit 9 is connected to the judgment level input terminal 13. The judgment level (the level indicated by . Therefore, as mentioned in the operation stirrup in Fig. 2, the adaptation level of the next differential boot button is set to 1.
It is possible to prevent the steps from becoming smaller and maintain the adaptation level as it is, thereby preventing the occurrence of large errors. By setting the judgment level (the level indicated by X in FIG. 1) appropriately for each adaptation level, more fine-grained adaptive differential compression data can be obtained.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the relationship between sample points and differential data for a voice waveform, FIG. 2 is a block diagram of a conventional ADPCM system, and FIG. 3 is a block diagram of a circuit showing an embodiment of the present invention. In the figure, 1#i is an analog-to-digital conversion circuit, and 2 is a sampling signal generator. 6 is an adder circuit. Figure 1

Claims (1)

[Claims] When the adaptive level of the signal to be transmitted is larger than the level-up setting value, which is the difference data value of the transmitted signal one sample before, the adaptive level of the signal to be transmitted is set to be one step larger than the adaptive level of the transmitted signal one sample before. In the ADPCM method, which performs adaptive compression of differential data by reducing the adaptive level of the transmission signal one sample before when the value is smaller than the adaptive level of the transmission signal one sample before, if the sample value of the previous sample exceeds the judgment level, the adaptive level of the transmission signal one sample before is reduced by one step. A waveform code in an ADPCM system characterized by providing means for maintaining the adaptive level of a signal to be transmitted even if the difference data value of the previous transmission signal is smaller than the level down setting value. method.