Error detection/correction and fault detection/recovery – Pulse or data error handling – Digital data error correction
Reexamination Certificate
1999-07-16
2002-02-12
Decady, Albert (Department: 2784)
Error detection/correction and fault detection/recovery
Pulse or data error handling
Digital data error correction
C714S794000, C714S795000, C375S262000, C375S265000, C375S341000
Reexamination Certificate
active
06347390
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to data coding method and device, data decoding method and device, and data supply medium, and particularly to data coding method and device, data decoding method and device, and data supply medium which are effectively used for cases where video data, audio data or other digital data are recorded in a recording medium such as a magnetic disc, a magnetic tape, an optical disc, a magnet optical disc, a phase-variation disc or the like and these data are reproduced from the recording medium.
2. Description of the Related Art
When video data and audio data are digitally recorded or reproduced into or from a recording medium such as a magnetic disc, an optical disc, a magnet optical disc, a magnetic tape or the like, it is required that the data can be recorded in as high density as possible and with high reliability. It is known that a PRML (Partial Response signaling with Maximum Likelihood detection) system which is implemented by combining a partial response system and a maximum likelihood decoding (detection) system is preferably used in order to satisfy the above requirement. According to PRML, data can be recorded in higher density and with higher reliability. In general, a partial response (1,1) or a partial response (1,0,−1) is used for the recording of digital data, and Viterbi decoding (Viterbi detection) is usually used for the maximum likelihood detection system.
There has been known a technique in which by combining the partial response system with the coding technique, the squared free euclidean distance d
2
free
is increased and SNR (Signal to Noise Ratio) is enhanced on the output of a partial response channel, whereby data can be recorded in high density and with high reliability. This technique is called as TCPR (Trellis Coded Partial Response), and a code produced by this technique is called as a Trellis code.
Here, the squared free euclidean distance d
2
free
is the minimum euclidean distance between two different paths which start from a common state and end to a common state on a Trellis diagram representing the output sequence of the partial response channel (hereinafter referred to as Detector Trellis. The Viterbi detection is carried out on the basis of the Detector Trellis). The starting and ending states may be different from each other.
For example, in the partial response (1,0,−1) (hereinafter abbreviated as PR
4
), if d
2
free
of the conventional bit-by-bit detection system is set to 1, d
2
free
can be set to 2 by performing the Viterbi detection (PR
4
ML). Here, PR
4
ML means a combination system of PR
4
and the maximum likelihood decoding (Viterbi detection). SNR can be more enhanced as the value of d
2
free
is increased, and it means that the recording can be performed in higher density and with higher reliability. Further, with respect to PR
4
, a practical Trellis code which sets d
2
free
to 4 is known (TCPR
4
). Here, TCPR
4
is a combination system of PR
4
and the Trellis code.
There is known a theory in which the value of d
2
free
can be increased by making coincidence between the null point of power density of codes and the null point of the transfer function of a transfer path, and the Trelllis code formed on the basis of the theory is called as MSN (Matched Spectral Null) code.
For example, in PR
4
, the transfer function becomes null at a DC component and a frequency component (so-called Nyquist frequency) of a half of a recording rate (1/T
c
, T
c
represents the time width (bit period) of one code bit). Accordingly, d
2
free
can be increased by making one or both of the DC component of the code and the Nyquist frequency null.
Further, for example, in the partial response (1,1) (hereinafter abbreviated as PR
1
), the transfer function is null at the Nyquist frequency component. Accordingly, d
2
free
can be increased by making null the Nyquist frequency component of the power density of the code.
Here, symbols “+1” and “−1” are allocated to code bits “1” and “0” respectively, and the sum of all the symbols from a symbol at a start time point (start point) or an infinite past of a code sequence to a symbol at the current time point, that is, RDS (Running Digital Sum) is an index for estimating the above DC component. If RDS is limited to a value within a fixed range, it means that the DC component of the power density of the code becomes null.
As in the case of RDS, the sum of all the values obtained by multiplying the symbol “+1” or “−1” allocated to a code bit by “−1” every other bit from a start time point or infinite past to the current time point, that is, ADS (Alternating Digital Sum) is an index for estimating the Nyquist frequency component as described above. If ADS is limited to a value within a fixed range, it means that the Nyquist frequency component of the power density of the code becomes null.
When the transfer function has a DC component like PR
1
, it is generally required to make the DC component of the power density of the code null. That is, it is required to prevent a DC component from being contained in a code which is a recording signal, for example in order to prevent occurrence of an error due to fluctuation of a reference level when a code is detected from a reproduction signal by the Viterbi detection in magnetic recording/reproduction having differential characteristics in a reproducing system, and in order to prevent occurrence of variation of various error signals such as a tracking error signal, etc. in servo control of a disc device in recording/reproduction of an optical disc or a magnet optical disc.
Therefore, in 8/10 conversion code (Rate 8/10 code) adopted in digital audio tape recorders (DAT), DSV (Digital Sum Variation) which is the amplitude value of RDS (the maximum value of RDS—Minimum value of RDS) is made infinite so that the power density of the code has no DC component, and also DSV is reduced to as small a value as possible. Here, the smaller DSV is, the more the low frequency band component of the power density of the code is suppressed. The code which is subjected to DSV control is called as “DS free code”.
In EFM (Eight-to-Fourteen Modulation) used for compact disc (CD) players, although the complete DC free coding is not performed, the control is performed so that DSV is reduced as much as possible in order to suppress the low frequency band component of the power density thereof.
Further, in the case where the PRML system is applied, there occurs a problem in length of a path memory for the Viterbi detection operation. The path memory is a storage device for storing a temporary judgment value of the detection until the Viterbi detection result is settled, and it needs a length (storage capacity) proportional to the time interval until the detection result (decoding result) is settled.
The time interval until the Viterbi detection result is settled, that is, the length of the path memory is normally controlled by constructing codes so that the Quasi-Catastrophic sequence (hereinafter referred to as “QC sequence”) is nullified and also the maximum length of the minimum-distance error events is shortened as much as possible.
Here, the minimum-distance error events generally indicate error events caused by a sequence making d
2
free
on detector Trellis. The QC sequence means two or more different paths whose squared euclidean distances are not accumulated on the Trellis diagram and thus exist (continue) infinitely. For example, when paths having state transitions of 111 . . . , 333 . . . , 555 . . . respectively on the Trellis diagram continue infinitely for a code sequence 101010 . . . , these three paths are called as QC sequences. Since the distance between the respective paths is not accumulated at any time and thus equal to zero, it cannot be judged which QC sequence of these sequences is right, that is, the path cannot be settled. Therefore, when a QC sequence occurs, the result of Viterbi detection cannot be settled.
When a device for performing the
De'cady Albert
Dooley Matt
Oblon & Spivak, McClelland, Maier & Neustadt P.C.
Sony Corporation
LandOfFree
Data encoding method and device, data decoding method and... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Data encoding method and device, data decoding method and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Data encoding method and device, data decoding method and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2967600