JPS6366097B2 - - Google Patents

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to a signal processing circuit, and more specifically, to converting an image signal into a digital signal, transmitting it,
The present invention relates to a signal processing circuit that converts an error into an error correction code and decodes and reproduces what is transmitted or recorded so that when an error occurs during recording and reproduction, it can be corrected.

[Background of the invention]

In digital magnetic recording devices such as digital VTRs that PCM encode television signals and record and reproduce them, error correction codes are used to correct code errors that occur during reproduction. However, with error correcting codes, there are limits to the code errors that can be corrected.
For errors exceeding this range, no correction is performed during error correction (hereinafter referred to as uncorrectable error), or a correction operation is performed but the error is incorrectly corrected (hereinafter referred to as erroneous correction). ) occurs. Among these, if it is uncorrectable, it can be determined that it is uncorrectable by, for example, dividing by the generator polynomial of the error correction code and checking the remainder.

Therefore, the code corresponding to the code word that has become uncorrectable is replaced with a signal from, for example, one scanning line earlier, by utilizing the redundancy of the television signal. On the other hand, when an error correction operation is performed, even if the correction is performed in error, it is impossible to determine this based on the corrected code. Therefore, if a code is incorrectly corrected during a correction operation due to an error that exceeds the scope of error correction, there is an error in this code and no correction is made. However, there is a problem in that this results in incorrect correction and deterioration of image quality.

Furthermore, in order to deal with burst errors where errors are concentrated in one place due to dropouts caused by scratches on the tape, etc., it is known that the code can be temporally interleaved between multiple code words of the error correction code. However, burst errors that occur in VTRs can have burst lengths exceeding several hundred bits in some cases. In order to enable correction of these long burst errors, it is necessary to interleave codes corresponding to multiple scanning lines. When performing interleaving on codes equivalent to multiple scanning lines in this way, if a longer burst error that exceeds the error correction range occurs, it is not possible to correct all the codes equivalent to multiple scanning lines, which may result in poor image quality. Since the deterioration of the code error is large, it is necessary to correct only the vicinity of the part with the code error using a correction code that has a strong correlation, such as a TV signal from one scanning line before.

In this case, when the range of errors that can be corrected is exceeded, as will be explained in detail later, several lines are corrected by the same correction signal (image), resulting in deterioration in image quality.

[Purpose of the invention]

Therefore, it is an object of the present invention to convert an image signal into a digital signal, add an error correction code to it, transmit or record it, and remove the erroneous corrections caused by erroneous corrections in a circuit that decodes and reproduces it. The objective is to realize a means for improving image quality.

[Summary of the invention]

In order to achieve the above object, the present invention includes the following steps before adding an error correction code to a digital image signal:
The present invention is characterized in that code conversion of the digital signal is performed. More specifically, after performing time series conversion of the code of a PCM encoded image signal, an error correction code is added, and a temporal interleaving process is performed in which a plurality of error correction codes are combined into one block. During playback, deinterleaving is performed in block units, error correction is performed after conversion to the original time-series error correction code, and this error-corrected code is further converted to time-series. When returning to the original PCM code, the error correction code word that has become uncorrectable is corrected. This modification can take various forms, as explained in the examples below.

[Embodiments of the invention]

Hereinafter, the present invention will be explained in detail with reference to Examples. FIG. 1 is a block diagram showing the configuration of an embodiment of a signal processing circuit according to the present invention, and FIG. 2 is a time chart for explaining the operation of the embodiment.

A PCM encoded television signal (input code) shown in FIG. 2a is applied to the PCM code time series conversion circuit 1. Here, PW ₁ , PW ₂ ,...PW ₄ are sample (picture element) signals, W ₁₁ to W ₁₆ , W ₂₁ to W ₂₆ ,
...W ₄₁ to _{W 46} are PCM codes. The signal a is time-series converted by the conversion circuit 1 according to a certain rule as indicated by the symbol in FIG. 2b. That is, PCM
The temporal order of the bit signals Wn, m (n, m are integers) constituting the code is rearranged by the conversion circuit 1 according to a certain rule. The encoder 2 adds error correction codes P _{1 .} . . P ₈ to the converted code b as shown in FIG. 2c. And W ₁ ,
One block is composed of four error correction codes W ₂ , W ₃ , and W ₄ . The interleave processing circuit 3 converts the plurality of error correction codes c into the code sequence shown in FIG. 2d. Signal d is recorded on the VTR.

On the other hand, during playback, the reproduced code from the VTR is converted by the deinterleave processing circuit 4 into a time-series signal that is the same as the original time-series error correction code c shown in FIG. Corrections will be made. The error-corrected code is sent to the delay circuit 6,
After being delayed by one code word of the error correction code, the PCM
In the time series conversion circuit 8, it is converted into the original time series PCM code as shown in FIG. 2g. However, if a so-called burst error, in which errors are concentrated in one place due to scratches on the tape, etc., occurs as shown by the arrow d, each code in the code series shown in e, which has undergone deinterleaving processing, This results in a correction error (shaded area) that exceeds the correctable range for the word. Therefore, the code after error correction is
This results in a mixture of codes that cannot be corrected and codes that have been incorrectly corrected. For example, f
As shown in , W ₁ , W ₃ , and W ₄ cannot be corrected, and W ₂ is incorrect and corrected. This corrected code is converted into a PCM code shown in g in the figure by time series conversion.

Therefore, the uncorrectable code word is determined in the uncorrectable detection circuit 7, and the PCM time series converter circuit 9, which has the same configuration as the PCM time series converter circuit 8, determines the uncorrectable code word corresponding to the PCM code. gives a sequence of codewords. The delay circuit 10 is
Give a delay equal to the number of PCM quantization bits, while
The counting circuit 11 counts the number of codes of uncorrectable code words included in the corresponding PCM code, and outputs a control signal of "1" if an uncorrectable code is included, and "0" otherwise. occurs.

Then, the selection circuit 12 selects the modified code when the control signal is 1, and selects the output of the delay circuit 10 when the control signal is 0. Therefore, selection circuit 1
As the output code of No. 2, a code sequence from which mis-correction has been removed is obtained.

The effects of the signal processing circuit according to the present invention will be explained in more detail below using the above embodiments.

In conventional methods that do not perform time-series conversion, even if uncorrectable codewords are corrected, the influence of erroneous correction becomes large. For example, in the conventional system that does not perform time series conversion, PW ₁ ,
Error correction codes are added to the PCM codes of PW ₂ , PW ₃ , and PW ₄ , respectively, to create W ₁ , W ₂ , W ₃ , and W ₄ . And when playing, W ₁ ,
Assume that W ₃ and W ₄ are uncorrectable, and W ₂ is incorrectly corrected. In this case, included in W ₁ , W ₃ , W ₄
PCM codes PW ₁ , PW ₃ , and PW ₄ have code errors that cannot be corrected, so they are corrected when decoding to PCM codes, but PCM code PW ₂ included in W ₂ corrects the errors made by W ₂ . Therefore, it has not been edited in any way and contains errors.

That is, in a code structure in which conventional time-series conversion is not performed for the same error, the influence of errors and corrections occurs. However, by performing time-series conversion as in the present invention, even if a code word that has been incorrectly corrected exists, when decoding to a PCM code,
This effect is almost eliminated.

Now, if the number of error correction codes included in a block is N, the number of quantization bits in PCM encoding is M, and the probability that the error correction code will incorrectly correct an error exceeding the correction range is Pe, then , when a code error exceeding the correction range occurs, in the conventional method that does not include time-series conversion, the probability P of incorrect correction while the block still contains the incorrect correction is P=1-(1-Pe) ^N ≒NPe. On the other hand, according to the present invention, the probability P of incorrect correction in a block is given by P=Pe〓 α=MIN(N, M). In this case, N=4, M=6, and
Pe depends on the type of error correction code, but is approximately 0.1 to 0.2
Then, in the conventional method, P≒0.4 to 0.8, while in the present invention, P≒10 ^-4 to 1.6 × 10 ^-3 . According to the present invention, the rate of incorrect correction due to incorrect correction can be reduced by time series conversion. It is possible to reduce the amount by several orders of magnitude compared to the conventional method that does not perform this process.

FIG. 3 is a block diagram showing the configuration of another embodiment of the signal processing circuit according to the present invention, and FIG. 4 is a time chart for explaining its operation. This is effective when performing interleave processing across multiple lines.

A PCM encoded image signal a is applied to the time series conversion circuit 1 as an input code. input code a
is the code for 3 scan lines. In the time series conversion circuit,
This code a is subjected to time series conversion into a signal indicated by b.
Error correction codes P ₁ , P ₂ , ...P ₆ are added to the encoder 2 for each fixed interval of the code sequence b, and a plurality of error correction code words W ₁ , W ₂ , ...W ₆ are made into one block. After converting to code sequence c, interleaving is performed by interleaving processing circuit 3 and converting to code sequence d.
Record on VCR.

During playback, playback codes from the VTR and dropout information corresponding to long burst errors exceeding the correctable range are passed through deinterleave processing circuits 4 and 13, and are converted into a code sequence as shown in FIG. This is code error information corresponding to the code sequence. Here, the shaded part of the signal e indicates a code error due to a burst error shown in the signal d,
All of these are considered to be beyond the range of correctable errors. The error correction circuit 5 controls error correction using code error information caused by a dropout that is separately detected by a dropout sensor, etc., and performs correction for errors within the correctable code error range. perform the following actions. Therefore, no correction operation is performed for errors that exceed the correction range.

In the case of FIG. 4, all errors are beyond the correctable range, so no correction is performed.

Therefore, after decoding, the code sequence shown in Fig. 4f is obtained, and the time series is further converted by the time series inverse conversion circuit 7 to obtain the same time sequence as the input code shown in Fig. 4g. Convert to sign. and,
This conversion causes code errors to be concentrated. On the other hand, the time series conversion circuit 8 generates a control signal that is "1" if there is an uncorrected code error, as shown in FIG. 4h, and "0" otherwise. Then, the selection circuit 9 selects and outputs the modified code when the control signal is "1", and selects and outputs the output of the time series conversion circuit 7 when the control signal is "0".

Therefore, W ₂₃ to _{W 210} of line 2 is line 1
In W ₁₃ to _{W 110} , W ₃₁ and W ₃₂ of line 3 are modified using W ₂₁ and W ₂₂ of line 2. That is, the erroneous codes on each line are corrected with different correction codes. In order to further clarify the effects of this embodiment, the case of a conventional method in which time-series conversion of input PCM codes is not performed will be explained.

FIG. 5 is a time chart illustrating the operation of a circuit according to the conventional system, and each symbol a, c...f, h corresponds to the same symbol in FIG. Since there is no signal corresponding to the signals b and g obtained by the time-series conversion in FIG. 4, it is necessary to modify the shaded portion of the decoded signal f by replacing it with a modified code from one scan earlier, for example. That is, W ₁₃ , W ₁₄ , W ₁₈ , and W ₁₉ in line 1 are replaced with W ₁₃ ′, W ₁₄ ′, W ₁₈ ′, and W ₁₉ ′ in the previous line. Next, in line 2, W ₂₃ , W ₂₄ ,
Replace W ₂₈ and W ₂₉ with W ₁₃ , W ₁₄ , W ₁₈ , and W ₁₉ of line 1, that is, W ₁₃ ′, W ₁₄ ′, W ₁₈ ′, and W ₁₉ ′. Also, W ₃₃ and W ₃₈ of line 3 are W ₂₃ of line 2,
Replace with W ₂₈ , that is, W ₁₃ ′ and W ₁₈ ′. Therefore, since lines 1, 2, and 3 are corrected using the same correction code in a portion with a code error, deterioration of image quality becomes a problem. However, in the embodiment shown in FIG. 3, by performing time-series conversion, multiple lines with erroneous codes are not corrected with the same correction code as in the conventional method described above. Plans are being made to improve image quality.

In addition, in the explanation of the example shown in FIG. 4, the explanation was about a time series conversion in which uncorrected code errors are concentrated in one place in the code sequence after decoding, but in the code sequence after decoding, A similar effect can also be obtained by performing time-series conversion such that undetected code errors are placed at different positions in two consecutive scanning lines of a plurality of scanning lines included in the block. For example, in the code sequence g that is finally decoded by performing time series transformation as shown in _FIG .
W ₁₅ , W ₁₆ , W ₁₇ , and in line 2 W ₂₁ , W ₂₂ ,
W ₂₃ , W ₂₉ , and on line 3, W ₃₆ and W ₃₇ will be corrected. However, the portions to be corrected on each line are located at different positions on two consecutive lines. Therefore, the time-series conversion according to the present invention causes much less deterioration in image quality during retouching than the conventional method that does not perform time-series conversion.

[Brief explanation of drawings]

1 and 3 are block diagrams showing the configuration of an embodiment of the circuit according to the present invention, FIG. 2 is a time chart for explaining the operation of the embodiment shown in FIG. 1, and FIGS. This figure is a time chart for explaining the operation of the embodiment shown in FIG. 3, and FIG. 5 is a time chart for explaining the operation of the conventional circuit. 1... Time series conversion circuit, 2... Encoder, 3...
Interleaving processing circuit, 4, 13... Deinterleaving processing circuit, 5... Decoder, 6, 10...
Delay circuit, 7... Uncorrectable detection circuit, 8, 9...
Time series inverse conversion circuit, 11... Counting circuit, 12...
selection circuit.

Claims (1)

[Claims] 1. A means for adding an error correction code to a PCM code;
A signal processing circuit comprising means for interleaving and recording/reproducing the code signal to which the error correction code has been added, means for deinterleaving the reproduced code signal, and means for error correcting and modifying the deinterleaved code signal. In the above
Before the means for adding an error correction code to the PCM code, a circuit is provided to perform time series conversion by rearranging the temporal order of the constituent bit signals of the PCM code according to a certain rule, and the deinterleaved code signal is A signal processing circuit comprising a circuit that performs inverse time-series conversion of the above-described time-series conversion. 2. In the signal processing circuit described in paragraph 1, the modifying means detects an error correction code in which error correction is not performed and a code error exists in the PCM signal obtained by the inverse transformation, and correlates the PCM code. A signal processing circuit consisting of a circuit that replaces with a strong modification code.