JPS5866440A - Waveform coding system - Google Patents

Abstract

PURPOSE:To cover a range of all input PCMs, by transmitting selectively the compressed PCM signals of the same number of bits when the maximum effective bit of the differential data between adjacent signals exceeds the expected level. CONSTITUTION:An effective maximum level signal is fed to a data compressing circuit 18 from an adaptive level setting circuit 16. Then the circuit 18 feeds compressed PCM signals ranging from an upper bit to the effective minimum level like a differential data compressing circuit 15 to a terminal A of a data selecting circuit 19. Therefore levels 3-0 are detected by the circuit 16, and a level detecting circuit 17 to which the effective minimum level signal is fed delivers ''0'' to the circuit 19 for selection of a terminal B and then delivers ''1'' to the circuit 19 for selection of the terminal A when receiving level 4 respectively. Thus the levels 3-0 are turned into ADPCMs as before. However, an output signal is not turned into a differential data signal but into a compressed PCM signal with level 4.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention is an Adaptive Differe system used for analysis and synthesis of speech in the PCM system, particularly in the audible band.
-ntial Pulsa Code Module
ation (ADPCM) method.

With the development of semiconductor technology, especially digital LSI, which is suitable for high-speed processing, the PCM system has become possible not only for PCM multiplex transmission on telephone lines and satellite communications, but also in the voice domain, allowing operation without deterioration in the transmission and processing system. It has become popular as a feature. That is, in the normal audio domain, the analog signal t
Since recorded sound, playback, and transmission are performed in ..., there is no way to absorb the data deterioration that occurs during analog transmission and processing. In addition to being able to easily apply playback processing, it also has functions that cannot be obtained with analog processing, such as a pitch conversion device that independently changes pitch and performance time using digital processing.

Furthermore, as shown in Figure 1(a), PCM samples the target input waveform at a certain frequency and quantizes it using, for example, an 8-bit binary sign, but the lower bits are truncated due to the allowable plexiform noise. - Signal compression is performed by widening the quantization width. FIG. 2 shows a block diagram of a conventional waveform encoding method ADPCM.

In FIG. 2, 11 is an A/D conversion circuit, 12 is a clock signal generation circuit, 13 is a difference data circuit, 14 is a data holding path, 15 is a difference data compression circuit, and 16 is an adaptive level setting circuit. The input waveform given to the conversion circuit 11 is sampled and held according to the reference clock signal generated by the clock signal generation scheme 12.
After being converted into an 8-bit code (PAM) signal, each signal is converted into a binary signal of, for example, an 8-bit code, and then converted into an 8-bit code (PCM) corresponding to Fig. 1(a).
The PCM signal (one of the 8 bits) is sent to the difference data circuit 13 as a signal and is sent to the data holding circuit 14 where it is converted to 1 in accordance with the reference clock signal from the clock signal generation circuit 12.
The zero difference data creation circuit 13, which is held for a clock time and sent to another terminal of the difference data circuit 13 in synchronization with the next clock signal, receives the 8-bit PCM (No. 1 to A The 8-bit PCM signal from the /D conversion circuit 11 is subtracted to obtain a difference PCM (difference data) signal with a polarity sign corresponding to wJ1 diagram (b).0 This difference data signal is input to the difference data compression circuit 15 and It is sent to the adaptive level setting circuit 16.The adaptive level setting Ni 16 that has received the difference data signal detects the highest valid data bit of the data signal, and sends the lowest effective level signal to the difference data compression circuit 15 from a preset number of valid bits. The 0-difference data compression circuit 15 outputs data from the upper bits up to the effective minimum level according to the effective minimum level signal from the adaptive level setting circuit 16, and anything below the effective minimum level is not output as an invalid information code. The output signal is the difference data compression circuit 1
Depending on the level input in step 5, the level becomes any of levels 5 to 0 shown in FIG. 1(C) to (h). A signal output from the adaptive level setting circuit 16 in addition to the above output is used as a control signal for related circuits. In this way, when ADPCM compresses a PCM signal, it calculates the difference between adjacent signals and calculates its value. Adaptive compression is performed using each quantization width corresponding to the maximum effective bit.

For example, the difference data of the PCM signal shown in FIG.
As shown in the code strings shown in FIGS. (c) to (h), it is expressed by codes (I, O) that correspond to positive and negative values and their absolute values. If we compress the difference data into a total of 4 bits, including the upper 3 bits and the sign, we get
5.4 Difference data to its effective 3-bit position
．． 3.2.It is expressed in six stages: 1 and 0. Normally, the signal obtained by sampling the input waveform does not have a large change between adjacent data, and the difference data is often concentrated in a low-level region, so the quantization noise must be set to a sufficiently narrow value as the quantization width. This is based on the idea that it can be suppressed to a range H that can be ignored in practical terms. The quantization width at each level is determined by the least significant bit of each level, but there are unavoidable drawbacks when compressing to a limited bit length. For example, the quantization width is 2" at level 5 and 2" at level 0. On the other hand, considering the allowable change width corresponding to sudden changes in the input waveform, the quantization width is is large, so it is possible to follow large changes, but at lower levels, it may be difficult to follow large changes.For example, the maximum display value is 1/2 of the input waveform maximum value 1.
At level 4, if the input waveform at the previous input changes from 1/2 value, the difference data will be within 1/2 even if it changes to 0 or 1 with the maximum change, so there is no synchronization, but if the previous input is extremely 0 Further, it has a drawback that the inverse direction from Fil causes quantization noise.

The present invention seeks to provide means to eliminate this drawback. Therefore, the present invention obtains a difference data signal for adjacent signals of PCM signals obtained by sampling and quantizing an input waveform, and adaptively compresses the quantization of the difference data signal to its maximum effective bit. In a waveform encoding method that outputs a polarity code, if the maximum effective bit of the difference data exceeds a preset level, compression of the same number of bits is performed instead of the compressed difference data signal and polarity signal.
It is characterized by selectively transmitting CM signals.

An embodiment of the present invention will be described below with reference to the drawings.

FIG. 3 shows a block diagram of a waveform encoding method according to an embodiment of the present invention. A/D conversion circuit 11, four-way signal generation circuit 12, difference data circuit w113, data holding circuit 14
, the differential data compression circuit 15, and the adaptive level setting circuit 16 are common to those of the conventional system, and the operation remains unchanged. However, it has other functions: 17 is a level detection circuit, 18 is a data compression circuit, and 19 is a selection concave path. Level detection times M&1
7 is set to a separately preset level, for example, level 4, and when the effective lowest level signal is sent from the adaptive level setting circuit 16 and matches level 4, the detection signal %1N is sent to the data selection circuit 19. A/D conversion circuit 11
Since the data compression circuit 18 which receives the PCM (i) corresponding to FIG. The signal up to the effective minimum level is inputted to the A terminal of the data selection circuit 19. Therefore, the adaptive level setting circuit 16 detects the signal level 3 to 0, and the level detection circuit 17, which receives the effective minimum level signal, inputs V. It outputs to the data selection circuit 19 to select the B terminal, and when level 4 is received, it outputs Y to the data selection circuit 19 to select the A terminal.

Therefore, at levels 3 to 0, the output signal becomes ADPCM as in the conventional case, but at level 4, the output signal does not become a difference data signal, as shown in FIG. 1 (1). The compressed PCM is changed to level 4 and the upper 4 bits are output.

As a result, the quantization width is 24 as in the conventional level 4, and in the level 4 code according to an embodiment of the present invention, an absolute value signal by compressed PCM is output, so all input P
CM range can be covered.

At the same time, level 5 becomes uncovered.

[Brief explanation of drawings]

FIG. 1 is a bit configuration diagram of a waveform encoding method according to a conventional method and an embodiment of the present invention, FIG. 2 is a block diagram of a conventional waveform encoding method, and FIG. 3 is a diagram of a waveform encoding method according to an embodiment of the present invention. FIG. 2 is a block diagram of the method. 11 is an A/D conversion circuit, 13 is a difference data creation circuit, 15
is a differential data compression circuit, 17 is a level detection circuit, and 18#'
i data compression circuit and 19 a data selection circuit. Figure 1

Claims (1)

[Claims] Pu1seCo obtained by sampling and quantizing the input waveform
In a waveform encoding method, a difference data signal is obtained for adjacent signals of de Modulation (PCM) signals, and the difference width of the nuclear ship data signal is adaptively compressed to its maximum effective bit, and the resultant signal is output together with positive and negative polarity codes. comprising means for detecting a difference data signal level and means for selecting an adaptively compressed PCM signal or an adaptively compressed difference data signal as an output signal, and when the maximum effective bit of the difference data signal exceeds a preset level, the compressed difference Compression Pul&e Co of the same number of bits instead of data signal and polarity code
A waveform encoding method characterized by selectively transmitting deModula (PCM) signals.