ArticlePDF Available

DC-free Multimode Code Design Using Novel Selection Criteria for Optical Recording Systems

Authors:

Abstract and Figures

DC-free run-length limited codes have been the cornerstone of all three generations of optical recording, CD, DVD and BD. Research into very efficient coding methods is paramount for the upcoming fourth generation. Guided Scrambling (GS) is an efficient coding method that has been reported in the literature. Under GS rules, a user word is translated into a plurality of possible candidate words, and among the candidate words the encoder selects the codeword with the least low-frequency spectral content. In our paper, we will present results of our attempts to improve the performance of GS-based codes. We will present new selection criteria and evaluate their performance and complexity. Specifically, we will evaluate the new selection criteria to the 2/3(1,7) parity preserving code used in Blu-Ray Disc.
Content may be subject to copyright.
J. Lee and K. A. S. Immink: DC-free Multimode Code Design Using Novel Selection Criteria for Optical Recording Systems
Contributed Paper
Manuscript received December 1, 2008 0098 3063/09/$20.00 © 2009 IEEE
553
Abstract — DC-free run-length limited codes have been the
cornerstone of all three generations of optical recording, CD,
DVD and BD. Research into very efficient coding methods is
paramount for the upcoming fourth generation. Guided
Scrambling (GS) is an efficient coding method that has been
reported in the literature. Under GS rules, a user word is
translated into a plurality of possible candidate words, and
among the candidate words the encoder selects the codeword
with the least low-frequency spectral content. In our paper, we
will present results of our attempts to improve the performance
of GS-based codes. We will present new selection criteria and
evaluate their performance and complexity. Specifically, we
will evaluate the new selection criteria to the 2/3(1,7) parity
preserving code used in Blu-Ray Disc.
Index Terms — Selection Criterion, GS, DC-free RLL code
I. INTRODUCTION
The design of codes for optical recording is essentially the
design of the combined DC-free and runlength-limited
(DCRLL) codes. An RLL constraint in optical recording plays a
crucial role for the reduction of channel impairments and clock
recovery. The DC-free property is for circumventing or reducing
interaction between data written on disk and the servo systems
that follow the track. In literatures [1][2], the design of DCRLL
codes can be accomplished by several design techniques and
has been mostly concentrated on byte-oriented DCRLL code
with the small codeword length. The data recording industry has
been moving towards detection scheme that can function well at
high code rate such as 16/17, 24/25 and 32/33. For certain
application, it is desirable that the code rate and codeword
length of a modulation code for optical recording are even
higher. However, unfortunately, the design of high-rate DCRLL
code satisfying all of two constraints is far from obvious, and
severely hampered by the large number of states of the finite-
state machine (FSM) which models the channel constraints
[3][4].
One possible solution for achieving high capacity adopts a
weakly constrained code, instead of perfectly RLL constrained
code, for designing recording code. Recently, multimode coding
scheme is issued for next-generation optical recording [4][5].
The coding scheme is one of methods for constructing weakly
constrained code with DC-free property. In multimode codes,
each source word is translated into codewords of a selection set
Jun Lee is with Digital Storage Research Laboratory, LG Electronics in
Korea (e-mail: leejun28@lge.com).
Kees A. Schouhamer Immink is with Institute for Experimental
Mathematics, Ellernstrasse 29-31, Essen, Germany (e-mail:
immink@turing-machines.com).
consisting of L > 3 codewords. The encoder evaluates the
quality of each codeword in the selection set, and then transmits
the codeword with the least DC-contribution. There are two key
elements for multimode code design with DC-free property.
One is a scrambler for translating source words into their
corresponding selection sets, and the other is a good criterion for
evaluating the quality of the candidate codewords. The spectral
performance of the code greatly depends on both issues.
The best multimode coding scheme reported is guided
scramble (GS) [6]. Originally, GS scheme is designed for fiber
optic communication system required to DC-suppression, and
its application is limited to the transmission system over fiber
cables. In recent, its application is moving toward optical
recording system requiring DC-free property. GS scheme
exploits the linear shift feedback register as scrambler and
augmenting step for generating the distinct candidate codewords.
The DC-control of GS scheme can be achieved by developing
selection criterion, which is one of the key elements of
multimode code. The criteria developed can also be extended to
any multimode coding scheme. Immink and Patrovics [4]
assessed the spectral performance of the GS scheme under
conventional criteria. With the same redundancy, the simulation
results show that the performance of GS scheme with short
codeword length is almost the same irrespective the selection
criteria, while that with long codeword length is very sensitive
to the selection criteria. This fact reveals that the selection
criterion is indispensable for multimode code design with DC-
free property and certain application.
The criteria reported for evaluating the quality of the
candidate codewords are minimum running digital sum
(MRDS), minimum squared weight (MSW) and minimum
threshold overrun (MTO) [1][4][9]. The MRDS criterion that
selects a codeword with minimum absolute RDS at the end of
each codeword requires the simplest complexity, while its
spectral performance is degraded as the length of codeword is
increased. The MSW criterion that selects a codeword with the
minimum variance of RDS among candidate codewords can
achieve the best performance irrespective of the codeword
length, while its complexity is large because it requires the
squaring operation. The MTO criterion simply counts the
number of times that the absolute value of RDS in the codeword
is larger than the threshold predetermined by trial and error.
Then, it selects the codeword with minimum overrun. The
structure is simple compared to MSW, while it can select
codeword with bad quality because it randomly chooses one if
there are codewords with the same penalty. This paper suggests
the minimum peak RDS (MPRDS) criterion that is simple to
implement while its efficiency approaches that of the MSW
criterion. The scheme does not require the exhaustive search of
optimal threshold unlike MTO, and achieves the reasonable
DC-free Multimode Code Design
Using Novel Selection Criteria for Optical Recording Systems
Jun Lee and Kees A. Schouhamer Immink, Fellow, IEEE
Authorized licensed use limited to: Kees Immink. Downloaded on August 11, 2009 at 00:48 from IEEE Xplore. Restrictions apply.
IEEE Transactions on Consumer Electronics, Vol. 55, No. 2, MAY 2009
554
performance irrespective of the number of candidate codewords
like MSW, and only requires simple components (adder and
comparator) while MSW requires squaring operator in addition
to that of conventional criteria for implementation. For
improving the performance of criteria, we also propose the sign
change (SC) aided criteria. SC criterion is appended to the
proposed and conventional criteria as sub-criterion, and then SC
aided criteria reduce the probability selecting codewords with
large DC-contribution among codewords with the same penalty.
Thus, SC aided criteria can achieve better DC-suppression than
those without SC. We also suggest the absolute RDS criterion
(ABSRDS). This criterion is the simplified version of MSW
because it uses absolute operation instead of the squaring
operation of MSW for computing RDS variation. The scheme
achieves the best performance among the proposed schemes. In
addition, we introduce the complexity reduction (CR) versions
of new and conventional criteria, which are CRMSW,
CRABSRDS, CRMPRDS and CRMTO. CR criterion sparsely
evaluates each codeword instead of checking at each bit of the
codeword. The scheme requires definitely less complexity than
the conventional schemes, and the performance loss is not
noticeable in the range investigated. For displaying certain
application of the proposed criteria, we apply the proposed
criteria to 2/3(1, 7) parity preserving (PP) code [7], and GS
scheme satisfying RLL constraints. For better DC-control of (1,
7) PP code, its encoding exploits look ahead (LA) algorithm that
looks ahead some codewords. From the simulation results, we
identify that the spectral performance of the proposed criteria
outperforms that of conventional criteria.
This paper is organized as follows. We firstly start with some
preliminaries in Section II. In Section III, we introduce the
conventional and proposed criteria, and analyze their spectral
performance. Section IV applies the proposed criteria to two
DCRLL coding for optical recording systems. Finally, Section V
remarks conclusion.
II. PRELIMINARIES
DC-free codes translate binary source sequences into binary
channel sequences with spectral null at zero frequency. The
amount of DC-content in channel sequences depends on the
range of running digital sum (or digital sum value), in short
RDS, of channel sequences. Let xi = {. . ., x-1, x0, . . ., xi, . . .}, xi
{-1, 1} be a binary sequence. The RDS zi is defined as
1
i
ijii
j
z
xz x
=−∞
==+
. (2.1)
The literatures show that if zi is bounded with small value, its
power spectral density (PSD) vanishes at zero frequency
[1][8]. Let the RDS zi of sequences meet the condition N1
zi N2 at any instant i, where N1 and N2 are two (finite)
constants, N2 > N1. The digital sum variation (DSV, N) is
defined as N = N2 - N1 + 1, and then sequences have DC-free
property if N is sufficiently small.
The channel capacity C(N) of maxentropic DC-free code
can be easily computed if N is given [1], and is the important
parameter for computing the efficiency of the implemented
code. The other quantity is sum variance. Sum variance plays
a significant role in the evaluation of the spectral property of
the sequences. The reason is because the smaller the sum
variance of sequences, the smaller its DC-content at low
frequency.
Code efficiency [1] is given by
2
2
(1 ( ))
(1 )
R
DS
CN
ERs
σ
=
(2.2)
where 1C(N) and σ2RDS are the redundancy and sum variance
of maxentropic DC-free sequences, respectively, and R is the
code rate of the implemented code, and s2 is variance of RDS
values obtained at every bit position of sequences produced by
implemented code. It means how redundancy-sum variance
product of implemented code approaches to that of
maxentropic sequence. Maxentropic DC-free sequences
satisfy the following relationship between the sum variance
and the redundancy [1]
0.2326 <(1C(N)) σ2RDS 0.25. (2.3)
Since the codeword length that we are targeting is very long
(greater than 50 bits), DSV is around 30. For large N, the
redundancy-sum variance product of maxentropic sequences
is approximately constant and equals 0.2326. Thus, E is given
by
2
0.2326
(1 )
ERs
. (2.4)
The efficiency in the equation (2.4) is used for discussing the
performance of the proposed criteria.
The paper applies the proposed criteria to GS scheme,
which is a multimode coding scheme, for examining their
performance. Figure 1 represents the operation procedure of
GS algorithm. In Figure 1, symbols T and denote the shift
register and modulo-2 addition, respectively. The operation
process of GS algorithm is executed as follows. In the first
step, the source word X = {x1,…,xm}, xi {0, 1}and i = 1,...,m,
is preceded by all the possible binary sequences of length
(r1) to generate the L' = 2r-1 vectors of B={b1,…,bL'}, bj =
{b1,...,bn} and j = 1,...,L'. In the second step, each vector of
length n=m+r1 consisting of B is provided to linear shift
feedback register. Then, set B' ={b'1,...,b'L'} is produced, In the
third step, the vectors in B' are preceded by both a one and a
zero, and are shuffled by the scrambler with polynomial x+1
again. Then, selection set C ={c1,...,cL}, L=2r, is produced.
Each vector of C is composed by modulo-2 addition values
between the current input and previous output of scrambler.
This step embodies the polarity bit principle. In the fourth step,
the given criterion selects and transmits codeword (cbest) with
the least DC-contribution out of candidate codewords. At
receiver end, the codeword is firstly descrambled by using x+1
polynomial, and then after removing the first bit, it is
descrambled. The original source word X is eventually
reconstructed by removing r-1 bits.
III. SELECTION CRITERIA
There are two key components for multimode code
design with DC-free property. The first one is the scrambler
for converting source words into their corresponding
selection sets, and the second one is the selection criterion,
or metric, for evaluating the quality of the candidate
codewords. The DC-control of multimode code can be
Authorized licensed use limited to: Kees Immink. Downloaded on August 11, 2009 at 00:48 from IEEE Xplore. Restrictions apply.
J. Lee and K. A. S. Immink: DC-free Multimode Code Design Using Novel Selection Criteria for Optical Recording Systems 555
achieved by developing a better criterion. This section
offers an overview of the conventional criteria, it
introduces new criteria, and evaluates their performance.
A. Conventional Criteria
MRDS: MRDS is often called word-end RDS (WRDS).
MRDS is criterion that selects a codeword with minimum
absolute RDS at the end of each codeword. The scheme
computes the absolute RDS values of candidate codewords
and selects one with the minimum absolute RDS if there is
no other codeword with the same minimum absolute RDS.
Otherwise, it randomly selects one. The scheme requires
the simplest complexity, while its spectral performance is
degraded as the length of codeword is increased.
MSW: MSW is criterion that selects a codeword with
minimum squared weight. The squared weight (wsq) is
defined as the expectation of the squared RDS values at
each bit position of the codeword. The smaller wsq, the
smaller its DC-content at low frequency. MSW criterion
can achieve the optimal performance irrespective of the
codeword length, while its complexity is large because it
requires the squaring operation.
Fig. 1. The operation procedure of GS algorithm.
MTO: MTO criterion utilizes the parameter, RDS
threshold, denoted by T > 0, predetermined by trial and
error. MTO criterion counts the number of times that the
absolute value of the RDS is larger than T. Here, the
counted value is termed as the penalty of the codeword.
Then, it selects the codeword with minimum overrun. If
two or more codewords have the same penalty, one of them
is randomly chosen and transmitted. The scheme is simple
compared to MSW, while it can select codeword with bad
quality because it randomly chooses one if there are
codewords with the same penalty. Besides, the scheme also
needs the exhaustive search for a good optimal threshold.
B. Proposed Criteria
B.1. Two Main Criteria and One Sub-Criterion
MPRDS: In optical recording system, the width of the
spectral notch region from the zero frequency must be as
large as possible for servo signal, and the DSV value is
inversely proportioned to the notch width [1]. Thus, for
minimizing the DSV, MPRDS criterion selects the
codeword having minimum peak RDS. The peak RDS is
the maximum absolute RDS value in a codeword. This
scheme achieves good performance irrespective of the
length of candidate codewords like MSW and requires
simple operation (compare and update MPRDS value) for
implementation unlike MSW (square RDS value at each bit
position and calculate average).
ABSRDS: This criterion is the simplified version of MSW.
It exploits the absolute operation instead of the squaring
operation of MSW. Firstly, this criterion calculates the integral
of absolute RDS of bipolar recording sequence up to a given
position. As a result, the codeword with minimum absolute
RDS variation is selected. The scheme achieves better
performance that is close to that of MSW than MPRDS, and
its complexity is definitely simple compared to MSW.
SC: The (conventional and proposed) main criteria
randomly select and transmit one codeword if there are two
or more codewords satisfying the given criterion. The
random selection has the possibility that the codeword with
larger DC-content out of codewords with the same penalty
is chosen. If a criterion reducing this possibility is
collaborated with the main criteria, the performance
improvement of conventional criteria is obvious.
Frequent sign change of RDS in a codeword means that
the RDS values are near from the zero, in other words, the
DSV is small with high probability compared to the case of
less sign change. Thus, if the SC criterion is applied to main
criteria for investigating the quality of candidate codewords,
the criteria with SC can control DC content better than those
without SC. Figure 2 shows the operation procedure of
MPRDS/SC as an example. This criterion can be performed
by two steps. Firstly, the scheme computes the PRDS and
counts the number of SC of candidate codewords. Secondly,
the criterion selects one with large SC if there are codewords
with the same MPRDS, otherwise, it selects codeword with
MPRDS. The reason that SC is selected as a sub-criterion is
because it cannot guarantee that the DSV value is small
without the main criteria. The reasonable thought is
supported by simulation results.
B.2. Complexity Reduction Method
We introduce the complexity reduction (CR) versions of the
proposed and conventional criteria, which are CRMSW,
CRABSRDS, CRMPRDS and CRMTO. The CR criterion
evaluates the selection metric at regular intervals instead of
each bit position in the candidate codewords. If the interval is
equal to one, then it is just the criterion without CR. Thus, we
can save the computation time for selecting codeword using
the CR. The performance loss is not noticeable at the surveyed
intervals. The CR criterion is developed for reducing the
complexity of main criteria, and can be also applied to others
with SC. Let us simply overview CRMTO/SC explaining the
concept of CR. The scheme observes MTO penalty at regular
intervals, while it counts the number of SC at every bit
position. Then, it selects the codeword with minimum penalty.
If the penalty of two or more codewords is equal, one with the
largest SC is selected and transmitted.
C. Simulation Results
This section evaluates the performance of the proposed
criteria. Figure 3 shows the efficiency comparison among
the proposed and conventional criteria as the number of
redundancy bits changes.
Authorized licensed use limited to: Kees Immink. Downloaded on August 11, 2009 at 00:48 from IEEE Xplore. Restrictions apply.
IEEE Transactions on Consumer Electronics, Vol. 55, No. 2, MAY 2009
556
In Figure 3, the connected points have the same
redundancy (1-R), codeword length r/(1-R), and selection
set of size 2r. From Figure 3, we can notice the following.
1. The MRDS/SC criterion is not efficient for long-term
low frequency minimization like MRDS even if it
outperforms MRDS irrespective of the number of
redundancy bits. The MTO/SC criterion is efficient for
long-term low frequency suppression, and its efficiency
outperforms the MTO criterion. In addition, the efficiency
Cadidate codewords
Compute PRDS
Count SC
Select codewords with MPRDS
Recording (or transmitting)
the codeword
Only one ?
yes
no Find one with
the largest SC
Fig. 2. Codeword selection procedure for MPRDS/SC criterion.
(a) Criteria without SC
(b) Criteria with SC
Fig. 3. The efficiency of the proposed criteria at (1-R)=1/56.
difference between MTO and MPRDS is clearly
noticeable, but the efficiency of MTO/SC really
approaches to that of MPRDS/SC. The results imply that
MTO can select a lot of codewords with large DC-
contribution if we do not add SC criterion, and SC
definitely prohibits the random selection among
codewords with the same penalty.
2. The efficiency of MPRDS/SC outperforms that of
conventional criteria except for MSW, and the
efficiency difference between MPRDS and MPRDS/SC
is small. Thus, the difference of spectral performance
between them is not distinguished. The fact can be
identified through the power spectral density of Figure 4.
It infers that MPRDS itself can achieve better DC-
control unlike MTO criterion.
3. The ABSRDS criterion almost perfectly uses the chance
provided by broader selection sets. Thus, its efficiency
approaches to that of MSW. The fact implies that
ABSRDS well confine the upper bound of absolute RDS
like MSW. The efficiency difference between ABSRDS
(or MSW) and ABSRDS/SC (or MSW/SC) is much
smaller than that between MPRDS and MPRDS/SC.
Thus, the difference of spectral performance between
them is not noticeable. Here, we note that criteria with
SC clearly have the spectral gain even if its difference is
small.
Figure 4 illustrates the power spectral density (PSD) of
codewords generated by each criterion when the number
of redundancy bits is 6. Figure 4 has a horizontal axis fc
(dB) and a vertical axis H(fc) (dB), where dB is defined
by 10log fc and 10logH(fc), respectively, and fc is channel
sequence frequency. Here, the channel sequence denotes
the sequence encoded by GS algorithm. The results in
Figure 4 show that the proposed main criteria can
independently control DC-content in sequences, but the
conventional main criteria except for MSW need the
support of SC. The suggested main criteria clearly reduce
the possibility that two or more codewords in a selection
set have the same penalty value. As a result the
performance of the proposed main criteria is not
deteriorated compared to that of the proposed main
criteria with SC.
Figure 5 presents the sum variance of codewords selected
by GS scheme with the efficiency corresponding to each
point of Figure 3. The results indirectly supports why the
proposed criteria achieve larger efficiency, and MTO/SC
obtains the efficiency better than MTO, and the efficiency
difference between MTO/SC and MPRDS/SC is not
noticeable.
Table 1 shows the spectral performance of CR criteria
at regular intervals I = 5, 10, 15 and 20. From Table 1 and
Figure 4, we can identify that CR criteria have the
performance loss at critical frequency (H(10-4)) compared
to criteria without CR, but the loss is not noticeable. As a
result, the CR criteria achieve the reliable DC-suppression
at critical frequency for optical recording system. The fact
Authorized licensed use limited to: Kees Immink. Downloaded on August 11, 2009 at 00:48 from IEEE Xplore. Restrictions apply.
J. Lee and K. A. S. Immink: DC-free Multimode Code Design Using Novel Selection Criteria for Optical Recording Systems 557
infers the possibility that GS scheme with CR criteria can
be applied to optical recording system with less
complexity.
TABLE 1
The spectral performance of CR criteria at r=6 and H (10-4)
I CRMSW CRABSRDS CRMPRDS/SC CRMTO/SC
5 -33.42dB -34.17dB -33.53dB -33.45dB
10 -32.82dB -34.08dB -33.48dB -32.82dB
15 -32.75dB -33.99dB -32.62dB -32.75dB
20 -32.48dB -33.60dB -32.55dB -32.48dB
(a) Criteria without SC
(b) Criteria with SC
Fig. 4. The spectral performance of the proposed criteria when r=6.
(a) Criteria without SC
(b) Criteria with SC
Fig. 5. The sum variance of the proposed criteria at (1-R)=1/56.
IV. APPLICATIONS
This section introduces two applications exploiting the
proposed criteria applying to GS scheme with RLL
constraints and the 2/3 (1, 7) PP code.
A. GS scheme with Runlength-Limited (RLL) Code
The scheme can be implemented by inserting RLL code
encoder between the second scrambler and selection
criterion in Figure 1. As a result, the whole transmitted
sequence satisfies RLL constraints. In the scheme, the RLL
code is the fixed-length code because GS scheme
accomplishes the block based encoding and decoding. The
RLL code used in this work is the 2/3 (1,7) code, and its
DC-content is large.
Figure 6 presents the spectral performance of GS
scheme using conventional and proposed criteria. The
code rate of GS scheme is 330/504 (r=6), and the
redundancy is 1/56. Here, the code rate (R) of the scheme
is given by R=(m×r)/((m+1)×r×3/2). From simulation
Authorized licensed use limited to: Kees Immink. Downloaded on August 11, 2009 at 00:48 from IEEE Xplore. Restrictions apply.
IEEE Transactions on Consumer Electronics, Vol. 55, No. 2, MAY 2009
558
(a)
(b)
Fig. 6. Spectral performance of the GS scheme with RLL constraints
and R=330/504 using various criteria.
results, we can identify that the proposed criteria can
conduct better DC-control compared to conventional
criteria even if GS scheme satisfies given RLL constraints.
However, compared to GS scheme without RLL code, it
shows the performance loss even if it has the acceptable
performance at critical frequency for optical recording.
We can conclude that GS scheme satisfying RLL
constraints can clearly construct the DC-free code with
small rate loss even if it uses RLL code with large DC-
content, and thus it is a promising candidate for optical
recording.
B. 2/3 (1, 7) PP Code Using Look-Ahead Algorithm
Parity preserving (PP) means that the modulo-2 addition
of a source word (the “parity”) is always equal to that of
the corresponding channel word. The mapping rule of 2/3
(1, 7) PP code used in this paper is based on the second
Table of [7]. This paper uses look ahead (LA) algorithm for
better DC-control of (1, 7) PP code.
(a) n =0
(b) n=3
Fig. 7. Spectral performance of (1,7) PP code using n LA algorithm.
Encoding step of (1, 7) PP code using n LA algorithm is
processed as follows. The scheme primarily inserts n+1
DC-control bits in the bitstream of the n+1 source words at
regular interval. Then, it generates the two tree structure
consisting of 2n possible candidate codewords. Finally, the
2n+1 possible codewords is evaluated by selection criterion,
and the root codeword of the tree with the least DC-
contribution is transmitted. Here, the encoding method with
n = 0 is the normal encoding method of (1, 7) PP code.
Figure 7 represents the spectral performance of (1,7) PP
code using new and conventional criteria when n = 0 and n
= 3, respectively. Here, the redundancy is 1/56. From
results, we can identify that the spectral performance of (1,
7) PP code using LA algorithm is more reliable, and the
proposed criteria contribute to the phenomenon.
V. CONCLUSION
This paper has proposed two new criteria for evaluating
candidate codewords for the multimode coding scheme. Two
main criteria are MPRDS and ABSRDS. The MPRDS
Authorized licensed use limited to: Kees Immink. Downloaded on August 11, 2009 at 00:48 from IEEE Xplore. Restrictions apply.
J. Lee and K. A. S. Immink: DC-free Multimode Code Design Using Novel Selection Criteria for Optical Recording Systems 559
criterion is very efficient for minimizing long-term low
frequency content, and it has a very simple complexity, while
the performance of ABSRDS is less than that of MSW,
ABSRDS requires less complexity than MSW. The SC
criterion complements the random selection among codewords
with the same penalty. Thus, the criterion contributes to their
efficiency improvement by supporting the main criteria. CR
criteria can be realized by computing penalty at regular
intervals instead of every bit position. Certainly, CR criteria
reduce the computation time for investigating the quality of
codewords, and the performance difference between CR
criteria and criteria without CR is not noticeable.
We have discussed the performance of the proposed criteria
in terms of efficiency, sum variance and spectral performance.
From our simulation results, we have identified that the
proposed criteria achieve a higher efficiency than the
conventional criteria, leading to a smaller sum variance and
improved spectral performance. In two applications, DC-
control using the proposed criteria is compatible with MSW
even if they have much less complexity. Thus, we conclude
that the proposed criteria are promising candidates for the
reliable DC suppression method of multimode codes in the
next generation of optical recording.
REFERENCES
[1] K. A. S. Immink, Codes for Mass Data Storage Systems,
Shannon Foundation Publishers, 1990.
[2] K. A. S. Immink, “A survey of codes for optical disk
recording,” IEEE Journal on Selected Areas in
Communications, vol. 19, no. 4, pp. 756-764, April 2001.
[3] K. A. S. Immink and W. Y. H. Wilson, “A comparison of two
schemes for generating DC-free RLL sequences,” IEEE
International Symposium on Information Theory, pp. 352,
June 2000.
[4] K. A. S. Immink and L. Patrovics, “Performance assessment
of DC-free multimode codes,” IEEE Trans. on
Communications, vol. 45, no. 3, pp. 293-299, March 1997.
[5] K. A. S. Immink, “Weakly constrained codes,” IEE
Electronics Letters, vol. 33, pp. 1943-1944, no. 23, Nov.
1997.
[6] I. J. Fair, Q. Wang and V. K. Bhargava, “Polynomial for
guided scrambling line codes,” IEEE Journal on Selected
Areas in Communications, vol. 13, no.3, pp. 449-509, April
1995.
[7] J. A. H. Kahlman and K. A. S. Immink, “Device for
encoding/decoding N-bit source words into
corresponding M-bit channel words, and vice versa,”
US Patent 5,477,222, Dec. 1995.
[8] G. L. Pierobon, “Codes for zero spectral density at zero
frequency,” IEEE Trans. on Inform. Theory, vol. 30,
no. 2, pp. 435-439, March 1984.
[9] S. K. Ahn, S. W, Suh and K. A. S. Immink, “Method for
modulation digital data and apparatus therefor,”
US patent 6,603,411, Aug. 2002.
BIOGRAPHY
Jun Lee received his B.S. and M.S. degree from
Dongguk University, Seoul, Korea in 1998 and
2000, respectively. Since March 2000, he has been
a Ph.D. student in Dept. of Electronic Engineering
at Dongguk University. In 2003, he joined the
faculty of Samsung Advanced Institute of
Technology (SAIT), Suwon, Korea, and he is
currently working with LG Electronics. His
research interests are signal processing and coding
for storage systems and communication theory.
Kees Schouhamer Immink received his PhD
degree from the Eindhoven University of
Technology. He was with Philips Research
Labs in Eindhoven from 1968 till 1998. He
founded and became president of Turing
Machines Inc. in 1998. He is, since 1994, an
adjunct professor at the Institute for
Experimental Mathematics, Essen University,
Germany. Immink designed coding techniques
of virtually all consumer-type digital audio and
video recording products, such as Compact Disc, CD-ROM, CD-Video,
Digital Audio Tape recorder, Digital Compact Cassette system, DCC,
Digital Versatile Disc, DVD, Video Disc Recorder, and Blu-ray Disc. He
received widespread recognition for his many contributions to the
technologies of video, audio, and data recording. He received a
Knighthood in 2000, a personal ‘Emmy’ award in 2004, the 1996 IEEE
Masaru Ibuka Consumer Electronics Award, the 1998 IEEE Edison Medal,
1999 AES Gold Medal, and the 2004 SMPTE Progress Medal. He was
named a fellow of the IEEE, AES, and SMPTE, and was inducted into the
Consumer Electronics Hall of Fame, and elected into the Royal
Netherlands Academy of Sciences and the US National Academy of
Engineering. He served the profession as President of the Audio
Engineering Society inc., New York, in 2003.
Authorized licensed use limited to: Kees Immink. Downloaded on August 11, 2009 at 00:48 from IEEE Xplore. Restrictions apply.
Article
Full-text available
We investigate a new approach for designing spectral shaping block codes with a target spectrum, H_t(f), that has been specified at a plurality of frequencies. We analyze the probability density function of the spectral power density function of uncoded n-symbol bipolar code words. We present estimates of the redundancy and the spectrum of spectral shaping codes with specified target spectral densities H_t(f_i) at frequencies f_i. Constructions of low-redundancy codes with suppressed low-frequency content are presented that compare favorably with conventional dc-balanced codes currently used in data transmission and data storage devices with applications in consumer electronics.
Article
Guided scrambling is an example of multimode coding technique, which belongs to line code. This technique uses scrambling polynomials to form selection sets, by using proper selection criteria, the codeword with least contribution to DC (direct current) content will be chosen for transmitting. This paper designs this kind of GS (guided scrambling) line code which is suitable to be transmitted in long-distance UTP channel. According to the requirement of the real channel, this technique optimizes the design parameters, such as augmenting bits, scrambling polynomials, bit rate and selection criteria, to achieve the high bit rate of 64/67. The simulation experiment shows that this scheme is efficient in suppression of low frequency, and its PSD characteristics meet the requirement of transmission in long-distance UTP channel.
Article
DC-free run-length limited codes have been the cornerstone of all three generations of optical recording, CD, DVD and BD. Research into efficient coding methods is paramount for the upcoming fourth generation. A Multimode coding technique is an efficient coding method that has been reported in the literatures. Under multimode coding rules, a user word is translated into a plurality of possible candidate words, and among the candidate words the encoder selects the codeword with the least low-frequency spectral content. In our paper, we will present a new coding technique for constructing error correcting high-rate DC-free multimode code using a low-density parity-check (LDPC) code for future high-density optical recording. The scheme achieves reliable DC-suppression and obtains considerable bit error rate (BER) performance gain through the iterative decoding (or turbo equalization) with the concatenated detector.
Patent
Full-text available
A method for modulating digital data and apparatus therefor is capable of determining a digital data stream coded by Run Length Limited swiftly and precisely so as to record data in a recording medium. The digital data modulating method includes the steps of comparing a preset critical value and DSV (Digital Sum Value) of a certain digital data stream, computing the penalty of the digital data stream by multiplying the number of the time that the DSV of the digital data stream is larger than the critical value by a preset weight value of the critical value, comparing the penalty of the digital data and a penalty of another digital data stream and selecting a digital data stream with a smaller penalty as the digital data stream. Therefore, with the method for modulating digital data and apparatus therefor, the digital data stream which will be modulated among a number of digital data streams can be selected precisely and swiftly.
Article
Full-text available
A device for encoding a stream of data bits of a binary source signal (S) into a stream of data bits of a binary channel signal (C), wherein the bit stream of the source signal is divided into n-bit source words (x1, x2), which device includes a converting circuit (CM) adapted to convert the source words into corresponding m-bit channel words (y1, y2, y3). The converting circuit (CM) is further adapted to convert n-bit source words into corresponding m-bit words, such that the conversion for each n-bit source word is parity preserving (table I). The relations hold that m>n, p>0, and that p can vary. Preferably, m=n+1. Further, a decoding device is disclosed for decoding the channel signal obtained by means of the encoding device.
Book
Full-text available
Preface to the Second Edition About five years after the publication of the first edition, it was felt that an update of this text would be inescapable as so many relevant publications, including patents and survey papers, have been published. The author's principal aim in writing the second edition is to add the newly published coding methods, and discuss them in the context of the prior art. As a result about 150 new references, including many patents and patent applications, most of them younger than five years old, have been added to the former list of references. Fortunately, the US Patent Office now follows the European Patent Office in publishing a patent application after eighteen months of its first application, and this policy clearly adds to the rapid access to this important part of the technical literature. I am grateful to many readers who have helped me to correct (clerical) errors in the first edition and also to those who brought new and exciting material to my attention. I have tried to correct every error that I found or was brought to my attention by attentive readers, and seriously tried to avoid introducing new errors in the Second Edition. China is becoming a major player in the art of constructing, designing, and basic research of electronic storage systems. A Chinese translation of the first edition has been published early 2004. The author is indebted to prof. Xu, Tsinghua University, Beijing, for taking the initiative for this Chinese version, and also to Mr. Zhijun Lei, Tsinghua University, for undertaking the arduous task of translating this book from English to Chinese. Clearly, this translation makes it possible that a billion more people will now have access to it. Kees A. Schouhamer Immink Rotterdam, November 2004
Article
Full-text available
The author reports on the performance of a new class of constrained codes, called weakly constrained codes. These codes do not strictly guarantee the imposed channel constraints, but rather generate codewords that violate, with a given (small) probability, the prescribed constraint. Weakly constrained codes are specifically of interest when it is desirable that the code rate R=p/q is very high, requiring codewords of length q>100
Article
Full-text available
Codes were designed for optical disk recording system and future options were explored. The designed code was a combination of dc-free and runlength limited (DCRLL) codes. The design increased minimum feature size for replication and sufficient rejection of low-frequency components enabling a simple noise free tracking. Error-burst correcting Reed-Solomon codes were suggested for the resolution of read error. The features of DCRLL and runlength limited (RLL) sequences was presented and practical codes were devised to satisfy the given channel constraints. The mechanism of RLL codes supressed the components of the genarated sequences. The construction and performance of alternative Eight to fourteen modulation (EFM)-like codes was studied.
Article
Full-text available
We report on a class of high-rate de-free codes, called multimode codes, where each source word can be represented by a codeword taken from a selection set of codeword alternatives. Conventional multimode codes will be analyzed using a simple mathematical model. The criterion used to select the "best" codeword from the selection set available has a significant bearing on the performance. Various selection criteria are introduced and their effect on the performance of multimode codes will be examined.
Conference Paper
We will discuss the generation of DC-free runlength-limited (DCRLL) sequences. We propose to employ standard RLL codes, where DC-control is effectuated by multiplexing the source data or the encoded data with DC-control bits. The DC-control bits offer the degree of freedom required to shape to spectrum. It is shown that a new technique, called parity preserving assignment, will offer great benefits over other constructions
Article
Guided scrambling (GS) is a line coding technique developed specifically for high bit rate fiber optic transmission systems. This paper continues the development of GS coding by introducing more scrambling polynomials that can be used with this technique. First, a method is presented for constructing polynomials which can be used with balanced encoding. Then it is shown that these polynomials can be used as bases for large families of polynomials, an expansion whose usefulness is demonstrated with the presentation of power spectra for several code configurations. These results indicate that polynomials can be selected from these families to exercise control over average as well as worst-case statistics of the encoded bit stream, a property not available with any other single efficient, balanced line coding technique developed to date. Criteria for polynomial selection are given, and polynomials are recommended for several GS code configurations
Article
In pulse-amplitude modulation (PAM) digital transmission systems line encoding is used for shaping the spectrum of the encoded symbol sequence to suit the frequency characteristics of the transmission channel. In particular, it is often required that the encoded symbol sequence have a zero mean and spectral density vanishing at zero frequency. We show that the finite running digital sum condition is a necessary and sufficient condition for this to occur. The result holds in particular for alphabetic codes, which are the most widely used line codes.