JPS6359293B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a device for encoding and transmitting (communication, recording, storage, etc.) a signal.

2. Description of the Related Art Devices that encode audio signals and perform recording/reproduction or communication are being used in various fields in place of conventional devices that perform signal processing at an analog level. As such a signal transmission device, a PCM recording and reproducing device is known, which converts an audio signal into a PCM signal, records it on a magnetic tape, magnetic disk, etc., and reproduces the signal to demodulate it again into an audio signal. . The number of bits of the PCM signal in such a device is, for example, 7 bits in the case of an audio signal, but in order to have a sufficiently large dynamic range so that quantization noise does not increase even with a small level signal, it is necessary to For example, 12 bits are required.

As a method to reduce the number of bits of such a PCM signal, DPCM is a method that encodes and transmits the difference between the previous sample value and the current sample value.
(Differential PCM) method is known. According to this method, for example, 8 bits and 12 bits
Almost the same accuracy or dynamic range as the PCM signal can be obtained.

Furthermore, as a method to expand the dynamic range, rough level information or range information of the original signal is transmitted together with the DPCM signal, so that even when the signal level is high or low, the 8-bit DPCM is always transmitted. The difference between the sample values can be encoded using all bits of the signal.
A method called DPCM-AQ (Adaptive Quantization) is also known. That is, when the signal level is high, quantization is performed with a quantization step width such as 8 bits/1V, and when the signal level is low, quantization is performed with a quantization step width such as 8 bits/0.1V. Information about the quantization step width is transmitted as a range signal (for example, 2 bits) at regular time intervals, and the receiving end uses the range information as a range signal.
This is a DPCM method that uses a type of non-uniform quantization method that accumulates DPCM signals. With this method, the quantization accuracy is extremely high even in areas where the signal level is small, so the PCM method and DPCM described above
The quantization noise is even lower than that of the conventional method, and the dynamic range is wider.

However, regarding the stability of reproduction or demodulation,
It gets worse in the order of PCM, DPCM, and DPCM-AQ.
This is because the recorded or transmitted code contains bits that are poorly read, and in the PCM system, the playback level changes instantaneously, which appears as click noise in the playback sound. Furthermore, in the DPCM method, addition of differentially encoded signals is performed sequentially during demodulation, so if a defective bit or error bit occurs once, the influence of this defective bit continues.
The playback sound is noticeably distorted. Furthermore, in the DPCM→AQ method, if the range information contains error bits, the range of the reproduced sound changes and the original signal cannot be obtained correctly, and in extreme cases, the range of the D/A converter may be changed. There is also the inconvenience that if the disc comes off, there will be no reproduced sound at all.

The present invention has been made in view of the above-mentioned problems, and it is an object of the present invention to provide a signal encoding/transmission device with extremely good stability in reproduction or demodulation.

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

Figure 1 shows an audio DPCM to which the present invention is applied.
- It is a block diagram of the recording system of the AQ disk recording and reproducing device. The input audio signal a is converted into a sample pulse b of, for example, 30KHz in the A/D converter 1.
is sampled and converted to a 12-bit PCM signal. This PCM signal is supplied to the shift register 2, where, for example, the upper two bits of the PCM signal (representing 1/2 and 1/4 levels of the full range of the audio signal) are sent to the shift register 2 as range information.
Extracted by. Note that the transmission cycle of the range information is, for example, 0.1 sec, and the register 3 extracts, for example, the maximum value in one cycle by a shift pulse c corresponding to this transmission cycle. This range information indicates the quantization step width.

The output of shift register 2 is supplied to shift register 4. The outputs of registers 2 and 4 are respectively supplied to a subtracter 5, where the previous sample value (output of register 4) and the current sample value (output of register 2) are
The difference between Therefore, the 7-bit DPCM signal (differential encoded signal) RD is obtained from the output of the subtracter 5.
A 1-bit code signal SI indicating the sign of the difference is obtained.

The above-mentioned DPCM signal RD, code signal SI, range signal AQ output from the shift register 3, and PCM signals RS0 to RS11 output from the A/D converter 1 are supplied to the recording section 6 and recorded on the disk 7.

FIG. 2 is a diagram showing the format of the recording signal in this embodiment. As shown in Figure 2, the 7-bit DPCM signal RD and the 1-bit code signal SI
is recorded on the disk 7 at approximately every sampling period of 33 μsec. Also, 2-bit range signal
AQ is recorded every 0.1sec. In this example,
For the range signal AQ, the same signal is recorded twice in succession as AQ-1 and AQ-2, so that a reading failure of one of the signals during reproduction can be detected.

Furthermore, in this embodiment, along with the range signal AQ,
12-bit PCM signal RS of A/D converter 1 output
is recorded approximately every 0.1sec, and the DPCM signal is recorded during playback.
Even if RD and range signal AQ contain error bits, conversion (demodulation) from DPCM to PCM on the playback side is possible.
is always performed with reference to the PCM signal RS at the correct level. This PCM signal RS is
Upper 8 bits RS-1A and lower 4 bits RS-1B
The same data as the upper 8 bits RS-1A is recorded as RS-2A after RS-1B, and the PCM signals RS-1A and RS-1A during playback are
It is possible to detect a reading failure on one side of 2A. Note that the entire 12-bit PCM signal may be recorded twice or more consecutively to check for reading errors. As shown in Figure 2, the upper 8 bits are RS-1A.
If the information is recorded twice as RS-2A and RS-2A, the number of transmission bits may be smaller.

FIG. 3 is a block diagram of the reproducing system of the disk recording/reproducing apparatus according to the present invention. 7-bit DPCM signal RD0 obtained from disk 7 via playback unit 8
~RD6 and 1-bit code signal SI are DPCM→
The signal is supplied to a PCM converter 9, where the DPCM signal is demodulated into a PCM signal. The converter 9 is a differential data accumulator, and is a circuit that successively adds or subtracts DPCM signals, which are differential data, according to the code signal SI, as described later.

On the other hand, the first range signal AQ-1 and the second range signal of 2 bits each obtained from the reproducing section 8
AQ-2 is the data accumulation pulse CP1, CP2 formed corresponding to the bit position of each data in FIG.
is stored in shift registers 10 and 11. The respective outputs of registers 10 and 11 are output from comparator 1.
2, and when the two match, the output of the comparator 12 becomes high level. When the output of comparator 12 is at a high level, it is assumed that the range information does not contain any error bits, and the set pulse is
SP is applied via gate 15 to the clock input of shift register 16. As a result, the correct range signal AQ-1 is stored in the register 16,
The range information AQ output from this register 16 also operates the DPCM→PCM converter 9.

Note that when the inputs AQ-1 and AQ-2 of the comparator 12 do not match, the comparator output becomes a low level, the set pulse SP is not supplied to the register 16, and the previous range information in the register 16 is saved. state.

On the other hand, the data obtained from the reproduction section 8 every 0.1 sec.
PCM signal (RS-1B, RS-2A, RS-1A)
are the corresponding shift registers 17, 18, 19, respectively.
are stored according to data storage pulses CP3 to CP5. From the output of register 17, the first PCM
The lower bit (RS-1B) of signal RS-1 is obtained. Also, from the output of register 19, the first
The upper 8 bits (RS-1A) of the PCM signal are obtained, and the second PCM signal (upper 8 bits) RS-2A is obtained from the output of the register 18.

The outputs of registers 18 and 19 are supplied to comparator 20, and when they match, comparator 2
0 output becomes high level. When the output of the comparator 20 is at a high level, it is assumed that the PCM signal does not contain any error bits, and the load pulse LD for one clock is applied to the gate 21 and line selector 2.
2 to the DPCM→DCM converter 9. The converter 9 receives PCM signals RS-1A (8 bits) and RS-1B output from registers 19 and 17.
(4 bits) is a 12-bit PCM signal RS0 to RS1
1, and when the load pulse LD is supplied, the addition shift register (accumulator register) in the converter 9 outputs the PCM signals RS0 to RS0.
It can also be loaded or preset on RS11.
This ensures that the converter 9 always operates with reference to the PCM signal, which represents the correct signal level.

FIG. 4 shows a circuit diagram of the DPCM→PCM converter 9 of FIG. DPCM signal RD0~ in Figure 3
RD6 is supplied to the range switching circuit 26, distributed to an appropriate range according to the output AQ (range signal) of the register 16, and inputted to the input A0 of the adder/subtractor 27.
~ A9 is supplied. The other inputs B0 to B11 of this adder/subtractor 27 are the output Q of the shift register 29.
0 to Q11 are supplied. Therefore, the adder/subtractor 27
, B+A or B depending on the code signal SI
-A calculation is performed and the calculation result is output Y0~Y1
Obtained from 1. The outputs Y0 to Y11 are normally supplied to the inputs D0 to D11 of the register 29 via a switch circuit 28 connected to the contact a side.
For each 30KHz data sending pulse DS supplied to its clock input CK, the register 29
The addition results or subtraction results Y0 to Y11 are accumulated. Therefore, the register 29 accumulates the DPCM signals, which are differential data, one after another and outputs the normal PCM signal.
Data demodulated into a signal is obtained. This data is supplied to the D/A converter 23 in FIG. 3, and the same audio signal as the input is reproduced.

Next, when the load pulse LD is obtained as described above, all the switch circuits 28 are switched to the b side, and the true value recorded on the disk 7 every 0.1 sec.
PCM data RS0 to RS11 (outputs of registers 19 and 17 in Figure 2) is input to register 29 D0
~D11. Therefore, the register 29 is
The true PCM data is preset or loaded every 0.1 sec, and thereafter, addition and subtraction of the difference data DPCM is performed using the adder/subtractor 27 and the register 29 based on the true PCM data.

Note that when the true PCM data RS-1A and RS-2A do not match, the output of the comparator 20 in FIG. 3 becomes a low level and the load pulse LD is not supplied to the switch circuit 28 in FIG. The register 29 is never preset.

As a result, the range information AQ and the code signal SI contain reading errors, and these are false information, resulting in a situation where the data converted from the DPCM signal to the PCM signal does not reproduce the original waveform at all. However, since the true PCM data is corrected every 0.1 seconds, extremely stable demodulation can be performed. In particular, it is possible to prevent an accident in which the demodulated PCM signal goes out of the conversion range of the D/A converter 23 due to false information contained in the range signal AQ, resulting in no reproduced sound at all. Similarly, even if the differential data DPCM contains error bits, the demodulated PCM signal is corrected with the true PCM data, which prevents distortion from occurring in the reproduced sound and allows extremely precise demodulation. It can be performed.

Furthermore, immediately after the range signal is transmitted and the range is switched, the inside of the shift register 29 is replaced with absolute level PCM data RS0 to RS11.
This PCM data is interpolated at each discontinuity point with a different quantization step width, thereby alleviating the discontinuity and significantly reducing the distortion that occurs each time the range is switched.

Note that in the above embodiment, the same PCM data is
The PCM data is recorded twice (RS-1A and RS-2A), and it is determined that the PCM data contains an error bit based on the discrepancy between the two.
Error bits may be checked by a combination of PCM data RS-1A and parity bits. Other bit check methods (CRC code method, IBM method) may also be used.
Alternatively, only one piece of PCM data may be recorded every predetermined period.

In addition to the method described above that encodes and transmits the difference between the previous sample value and the current sample value, the DPCM method also encodes the difference between the predicted value and the sample value (prediction error). There are also known predictive coding methods (generally called
The present invention can also be applied to the DPCM method.

As described above, the present invention provides an adaptive differential coding transmission device that transmits range information about a quantization step width that is varied according to an input level at predetermined time intervals, in which the absolute level of the input is indicated together with the range information. Since encoded signals are transmitted, when cumulatively decoding differentially encoded signals based on range information at the transmission receiving end, even if there is an error in the range information, the cumulative value is immediately corrected by the encoded signal. Therefore, the cumulative value will not diverge abnormally. In addition, each time the quantization step width changes, the encoded signal indicating the absolute level is interpolated into the cumulative value, so the quantization error corresponding to the quantization step width is not propagated one after another every time the range is changed. Therefore, discontinuous distortion caused by intermittent transmission of range information is greatly alleviated.

Also, the same range information is transmitted multiple times in a row,
When the range information does not match each other, differentially encoded signals are accumulated according to the previous value retention information of the range information, so accumulation in the wrong range almost never occurs, and stable decoding can be performed. .

[Brief explanation of the drawing]

Figure 1 shows an audio DCPM to which the present invention is applied.
- A block diagram of a recording system of an AQ disk recording/reproducing apparatus; FIG. 2 is a diagram showing the format of a recording signal according to the present invention; FIG. 3 is a block diagram of a reproducing system of a disk recording/reproducing apparatus to which the present invention is applied; FIG. 4 is a block diagram of the DPCM→PCM converter of FIG. 3. In addition, in the symbols used in the drawings, 1...
...A/D converter, 5...Subtractor, 7...Disk, 9...DPCM→PCM converter, 12...Comparator, 15...Gate, 16...Shift register,
26...Range switching circuit, AQ-1, AQ-2...
...This is a range signal.

Claims (1)

[Claims] 1. In an adaptive differential encoding transmission device that varies the quantization step width at predetermined time intervals in accordance with the level of an input signal, and transmits the difference between an input sample value and its previous value or predicted value, A pair of a differential encoded signal DPCM, a range signal AQ corresponding to the quantization step width, and an encoded signal PCM indicating the absolute level of the input signal one sample after the encoded signal PCM from which the range signal was detected. is transmitted at the predetermined time interval, and a plurality of pieces of the same information regarding the range signal are transmitted in succession,
At the time of decoding, the differentially encoded signal DPCM is accumulated based on the range signal AQ, and the cumulative value is replaced with the encoded signal PCM every time the range signal AQ is transmitted, so that the plurality of range signals An adaptive differential encoding transmission device that performs the above-mentioned accumulation based on a previous value holding signal of a range signal when the range signals do not match.