High-Rate Spatially Coupled LDPC Codes Based on Massey's Convolutional Self-Orthogonal Codes

Daniel J. Costello; David G. M. Mitchell; Jr.; Michael Lentmaier; Min Zhu

REVIEW 2 major objections 5 minor 1 cited by

High-Rate Spatially Coupled LDPC Codes Based on Massey's Convolutional Self-Orthogonal Codes

T0 review · 2 major / 5 minor · reviewed 2026-08-09 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Massey's self-orthogonal convolutional codes, stripped of their degree-one variable nodes, become high-rate spatially coupled LDPC codes with guaranteed girth at least six and free distance at least J+1.

desk verdict Useful high-rate SC-LDPC construction with guaranteed girth 6, but the empirical “at least as good” claim is rate-confounded. read the letter →

arxiv 2502.03774 v2 pith:VPO7QFTR submitted 2025-02-06 cs.IT math.IT

Daniel J. Costello , Jr. , Min Zhu , David G. M. Mitchell , Michael Lentmaier This is my paper · ORCID

classification cs.ITmath.IT MSC 94B1094B3594B60

keywords spatiallycoupledLDPCcodesconvolutionalself-orthogonalprotographliftingslidingwindowdecodinggirthfreedistancethresholdsaturationhigh-rate

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper argues that Massey's convolutional self-orthogonal codes (CSOCs), originally built for low-complexity threshold decoding, can be repurposed as high-rate spatially coupled LDPC (SC-LDPC) codes. The key move is to discard the systematic parity columns, which removes the degree-one variable nodes that hurt belief propagation. The resulting non-systematic protographs are regular, with fixed variable-node degree J and check-node degree (n−1)J, and they inherit two guarantees from the CSOC structure: girth at least six and free distance at least J+1. Lifting these protographs with permutation matrices preserves both properties, and systematic encoding remains possible through an equivalent recursive systematic convolutional generator. Numerical simulations and protograph EXIT analysis indicate that the lifted codes, decoded with BP sliding-window decoding, perform at least as well as classical edge-spreading SC-LDPC codes while offering favorable high-rate flexibility.

What carries the argument

The machinery is the non-systematic CSOC parity-check matrix H_ns(D) = [g^(1)(D), ..., g^(n−1)(D)], obtained by deleting the degree-one parity column from each time unit of the systematic CSOC parity-check matrix. This converts the irregular systematic CSOC graph into a regular (J,(n−1)J)-regular convolutional protograph while preserving the self-orthogonality that forbids 4-cycles. The proof of Theorem 1 relies on two facts: a 4-cycle in any lifted graph must correspond to a 4-cycle in the base protograph, and a codeword of weight less than J+1 would necessarily create a 4-cycle in the lifted matrix. The construction is completed by replacing each 1 in the protograph with a permutation matrix of size M, which increases the constraint length and improves performance, and then decoding with iterative belief-propagation sliding-window decoding (SWD) over a window of W constraint lengths.

What would settle it

Measure the BER or threshold gap of a CSOC-based and a classical edge-spreading code at the same effective rate, for example by increasing the termination length L of the CSOC code to match the classical code's rate at equal latency; if the CSOC code no longer outperforms at L=200, the equality claim fails. At L=1000 the PEXIT gaps already become essentially identical, so a careful simulation at matched effective rate would either confirm parity or expose a residual advantage.

Watch

Extended reading notes

Core claim

The central discovery is that the non-systematic form of a rate (n−1)/n CSOC parity-check matrix is a (J,(n−1)J)-regular convolutional protograph with no degree-one variable nodes, and every permutation-matrix lift of it is a high-rate SC-LDPC code with girth at least six and free distance at least J+1. This is Theorem 1: the proof shows that a 4-cycle in a lifted graph would project to a 4-cycle in the base protograph, which the orthogonality of the CSOC forbids; the free-distance lower bound then follows because any codeword of weight at most J would require two columns sharing two row positions, forcing a 4-cycle. The authors verify the construction on published CSOC tables, using rates from 1/2 to 49/50 and strengths J from 3 to 6, and report BER performance within 0.25–0.60 dB of finite-length bounds at BER $10^{-5}$. Threshold analysis with the PEXIT technique shows that, for large termination length L, the iterative decoding thresholds approach the MAP thresholds of the underlying LDPC block codes, matching the behavior of classical edge-spreading constructions.

Load-bearing premise

The numerical claim that CSOC-based codes perform at least as well as classical edge-spreading assumes the comparison is fair even though the CSOC codes have slightly lower termination rates; the paper's own threshold table shows the advantage shrinking when this rate loss is taken into account.

Editorial extensions

If this is right

The lifted non-systematic CSOC protographs form a class of (J,(n−1)J)-regular SC-LDPC codes with guaranteed girth at least six, eliminating the need to search liftings for 4-cycles.
The guaranteed free distance of at least J+1 contributes to improved error-floor behavior at high signal-to-noise ratios.
Systematic encoding is maintained through the equivalent recursive systematic convolutional generator matrix, so the non-systematic parity-check matrix does not complicate practical encoding.
The large existing catalog of CSOCs, with rates up to 49/50 and strengths J = 3 to 6, provides ready-made high-rate protographs without new table searches.
For sufficiently large lifting factor M and termination length L, iterative SWD thresholds approach the MAP thresholds of the corresponding LDPC block codes, matching the threshold-saturation behavior of classical edge-spreading SC-LDPC codes.
The same construction applied with larger J yields spatially coupled moderate-density parity-check (SC-MDPC) codes that can approach capacity with BP-based decoding.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A fair head-to-head at matched effective rate would likely remove most of the small BER advantage seen at L=200; the paper's own threshold table already shows gaps to capacity nearly equal at L=1000, suggesting that the apparent performance edge comes mainly from the CSOC codes' slightly lower termination rates.
The construction points to a broader recipe: any convolutional code whose parity-check matrix has orthogonal checks can be used as a protograph seed, so perfect difference-set techniques might generate lower-rate SC-LDPC codes with similar girth and distance guarantees.
Because CSOC-based designs naturally use large memory m with small lifting factor M, they offer an operating point complementary to classical edge-spreading (small m, large M); optimizing the m/M tradeoff for latency-constrained streaming would be a natural next step.
The paper's mention of circulant and time-invariant liftings suggests a practical extension: quasi-cyclic high-rate SC-LDPC codes with low implementation complexity that preserve the girth-6 guarantee, an avenue worth testing against random-lifting performance.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The new thing here is the non-systematic CSOC protograph: dropping the degree-1 parity columns yields a regular (J, (n-1)J) graph with no degree-1 VNs, and Theorem 1 shows that lifting with permutation matrices preserves girth at least 6 and free distance at least J+1. That is a clean, useful result, and the paper is honest about the tradeoff: the non-systematic modification lowers the rate, but a large catalog of high-rate CSOCs from Massey, Robinson-Bernstein, and Wu makes it easy to find good starting points. The structural part is solid and the proof is straightforward. Credit where due: the construction genuinely removes the need to search for 4-cycle-free liftings, which is a practical convenience for high-rate SC-LDPC design.

The soft spot is the performance comparison. Example 5 compares at equal latency but unequal termination rate: Code I has R_t=0.635 vs Code III's 0.662, and Code V has 0.650 vs Code VII's 0.663. The paper's own Table I shows that at L=200 the CSOC codes have larger gaps to capacity (0.470 and 0.510 dB) than the classical codes (0.152 and 0.348 dB). The authors concede in Section IV-A that the lower rates may explain some or all of the BER difference. At L=1000, where rates nearly equalize, the gaps are essentially identical, so the claimed advantage disappears. The abstract's claim (ii), “performance at least as good,” is simply not established by the evidence presented. This is a rate-confounded comparison, not a fatal flaw in the construction, but it needs to be reworded and backed by equal-rate simulations or a proper latency-rate tradeoff analysis.

Minor points: no simulation code or BER data is shipped, and the PEXIT scripts are absent, which makes the numerical results harder to verify. The self-citations to [7], [17], and [21] are standard and appropriate.

Bottom line: this is a legitimate high-rate SC-LDPC construction with formal guarantees that deserves referee time, but the empirical claims should be tempered. The paper is for coding theorists and practitioners who want a simple guaranteed-girth construction from existing tables, not for anyone looking for a new capacity result.

Recommendation: send to peer review with a request for revision, primarily to fix the performance claim and either add equal-rate comparisons or explicitly frame the result as a rate-loss-tolerant construction with comparable thresholds. The structural theorem is worth publishing.

Referee Report

2 major / 5 minor

Summary. This paper proposes a new class of high-rate spatially coupled LDPC (SC-LDPC) codes constructed from Massey's convolutional self-orthogonal codes (CSOCs). The authors treat the parity-check matrix of a rate-(n-1)/n systematic CSOC as a convolutional protograph, then modify it to a non-systematic form by removing the degree-1 parity columns. The resulting protograph is (J,(n-1)J)-regular and the authors prove, in Theorem 1, that any permutation-matrix lifting has girth at least 6 and free distance at least J+1. They also show that systematic encoding is still possible using an equivalent recursive systematic convolutional encoder. Numerical BER simulations and PEXIT threshold analyses are presented to compare the CSOC-based codes with classical edge-spreading protograph codes at rate 2/3.

Significance. The construction is attractive: it leverages published CSOC tables to obtain high-rate protograph-based SC-LDPC codes with guaranteed girth and free distance, avoiding the need for lifting searches. The structural results in Theorem 1 are plausible and standard. The main weakness is the empirical claim that the proposed codes "perform at least as well as" classical edge-spreading codes. The paper's own Table I and Section IV-A show that this performance comparison is confounded by unequal termination rates; at L=200 the CSOC-based codes have worse thresholds gaps to capacity, and at L=1000 the gaps are essentially equal. Thus the contribution reduces to a rate-loss-tolerant construction with comparable thresholds at large L, not a demonstrated performance gain. The structural and encoding contributions are still valuable and publishable, but the performance claims need to be corrected.

major comments (2)

[Abstract, Section IV-A, Section IV-B, Conclusion] The claim that CSOC-based SC-LDPC codes "perform at least as well as" classical edge-spreading protograph codes is not supported by the evidence. In Example 5, Code I has actual rate 0.635 versus 0.662 for Code III, and Code V has 0.650 versus 0.663 for Code VII (Table I, rows A). The PEXIT gaps to capacity at L=200 are 0.470 dB and 0.510 dB for the CSOC codes versus 0.152 dB and 0.348 dB for the classical codes, i.e., the CSOC codes are worse. At L=1000 the gaps are nearly equal (0.166 versus 0.161 and 0.349 versus 0.362). The authors themselves concede in Section IV-A that the lower rates "may explain some or all of the difference in the observed code performance." The observed 0.1 dB BER advantage in Figure 6 is therefore rate-confounded. The abstract and conclusion overstate the result; a rate-fair comparison (for example, using larger L for the CSOC codes so that the actual rates match) is needed, or the claim must be weakened to "comparable thresholds at large L."
[Section IV-A, Example 5] The numerical comparison in Example 5 fixes decoding latency while allowing the actual code rates to differ. Codes I and V use larger memories (m=19 and m=10) than Codes III and VII (m=3 and m=2), which lowers their terminated rates at L=200. This makes the comparison a latency-fair but rate-unfair one. The bullet in Section IV-A stating that the lifted CSOCs "perform about 0.1 dB better" is therefore not a valid demonstration of a performance advantage; the authors should either compare codes at the same actual rate (e.g., by increasing L for the CSOC codes until their terminated rates match the classical codes) or remove the causal interpretation. As written, the numerical results support only a rate-loss-tolerant design with comparable performance once rates are accounted for.

minor comments (5)

[Section IV-A, Example 4] The "finite length bound" comparisons (0.55 dB and 0.60 dB) do not specify which finite-length bound is used (e.g., normal approximation or sphere-packing) or the exact blocklength for each code. Please add this information so the gaps are reproducible.
[Figure 6] The curves in Figure 6 are not labeled with the code numbers (Code I through Code VIII) in the figure itself. Please add labels or a legend to make the figure self-contained.
[Section III, after Eq. (7)] The phrase "fixed CN degree (n-1)J" appears both before and after Eq. (7), making the paragraph read as if the sentence was accidentally duplicated. Please remove the redundant occurrence.
[Theorem 1 proof] The girth argument implicitly assumes that a 4-cycle in the lifted graph projects to a 4-cycle in the base protograph. This is standard for permutation-matrix liftings of simple protographs, but stating it explicitly would make the proof more rigorous.
[Section IV-A, bullet list] Minor wording: the final bullet uses "doesn't" informally, and "Code VI" is lowercase in one place. Please make the style consistent.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the girth and free-distance claims are proved from the CSOC orthogonality construction, and the numerical comparison is rate-confounded but not circular.

full rationale

The paper's structural claims are derived in-text rather than assumed via a fitting or self-citation chain. Theorem 1 proves that any lifted non-systematic CSOC protograph has girth at least six and free distance at least J+1. The base protograph's lack of 4-cycles follows directly from the CSOC orthogonality condition stated in Section II, and the lifting argument shows that a 4-cycle in the lift would project to a 4-cycle in the base protograph. The free-distance claim uses the fact that each column has weight J and that any shorter zero-sum column combination would force two columns to overlap in at least two positions, again creating a 4-cycle. These are genuine graph-theoretic arguments, not a restatement of the conclusion. The non-systematic matrix Hns is explicitly obtained by deleting the degree-one parity columns from the systematic H, and the systematic-encoding claim is a standard recursive-systematic transformation, not a fitted parameter. The numerical comparison in Example 5 and Table I does contain a real evidentiary weakness: the CSOC-based Codes I and V have lower termination rates than the classical Codes III and VII, and the authors concede in Section IV-A that this 'may explain some or all of the difference in the observed code performance.' Table I shows larger threshold gaps for the CSOC codes at L=200 and essentially equal gaps at L=1000, so the abstract's 'at least as good' performance claim is not fully established. This is a correctness/experimental-design concern, however, not circularity: the simulations and PEXIT thresholds are not used as inputs to the construction, and no predicted quantity is fitted from the data that is then reported as a prediction. The self-citations [7], [17], and [21] are prior published results on protograph lifting, girth preservation, and rate/latency tradeoffs; they are background support rather than the load-bearing basis for the central theorem, which is proved in the paper. Accordingly, no circular step meeting the quoted-reduction standard is present.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

No free parameters were fitted to data; the code parameters (m, J, M, L, W) are hand-selected design choices, not fitted values, and the central theorem is independent of their specific values. No new physical or mathematical entities are introduced. The construction leans on published CSOC tables and standard lifting and threshold-saturation results.

assumptions (4)

domain assumption Published CSOC generator polynomials from [6], [8], [10]-[12] satisfy self-orthogonality, have g0=1, weight J, and free distance J+1.
Section II and Examples 1, 2, and 4 rely on these tables without re-deriving them. A table error would change the degree profile and invalidate the guaranteed girth and distance claims.
domain assumption Permutation-matrix lifting of a simple protograph with no 4-cycles produces a lifted graph with no 4-cycles.
Theorem 1(i) uses this standard lifting property, citing [17], to transfer the girth-6 guarantee from the base non-systematic CSOC protograph to the lifted SC-LDPC code.
domain assumption The rational systematic generator matrices Gsys(D) with entries g^(i)(D)/g^(n-1)(D) are realizable convolutional encoders, because the denominators have nonzero constant term.
Section III claims systematic encoding is preserved by the recursive form without a full realizability or non-catastrophicity proof. The property is standard for RSC encoders with g0=1 denominators.
domain assumption PEXIT chart analysis and threshold saturation accurately describe the iterative decoding thresholds of these SC-LDPC ensembles for large lifting factors and termination lengths.
Section IV-B interprets the PEXIT thresholds as approaching MAP thresholds, relying on [2], [3], and [22]. The numerical conclusions depend on this standard theory.

how reviews work

0 comments

Cite this review

Pith. "Pith review of High-Rate Spatially Coupled LDPC Codes Based on Massey's Convolutional Self-Orthogonal Codes." pith.science (2026). https://pith.science/paper/VPO7QFTR

@misc{pith2026250203774,
  author       = {Pith},
  title        = {Pith review of: High-Rate Spatially Coupled LDPC Codes Based on Massey's Convolutional Self-Orthogonal Codes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VPO7QFTR}},
  note         = {Machine review of arXiv:2502.03774}
}

read the original abstract

In this paper, we study a new class of high-rate spatially coupled LDPC (SC-LDPC) codes based on the convolutional self-orthogonal codes (CSOCs) first introduced by Massey. The SC-LDPC codes are constructed by treating the irregular graph corresponding to the parity-check matrix of a systematic rate R = (n - 1)/n CSOC as a convolutional protograph. The protograph can then be lifted using permutation matrices to generate a high-rate SC-LDPC code whose strength depends on the lifting factor. The SC-LDPC codes constructed in this fashion can be decoded using iterative belief propagation (BP) based sliding window decoding (SWD). A non-systematic version of a CSOC parity-check matrix is then proposed by making a slight modification to the systematic construction. The non-systematic parity-check matrix corresponds to a regular protograph whose degree profile depends on the rate and error-correcting capability of the underlying CSOC. Even though the parity-check matrix is in non-systematic form, we show how systematic encoding can still be performed. We also show that the non-systematic convolutional protograph has a guaranteed girth and free distance and that these properties carry over to the lifted versions. Finally, numerical results are included demonstrating that CSOC-based SC-LDPC codes (i) achieve excellent performance at very high rates, (ii) have performance at least as good as that of SC-LDPC codes constructed from convolutional protographs commonly found in the literature, and (iii) have iterative decoding thresholds comparable to those of existing SC-LDPC code designs.

Figures

Figures reproduced from arXiv: 2502.03774 by the authors.

**Figure 2.** Block diagram of a SWD with window size W constraint lengths for an SC-LDPC code. where S (D) = g (1) (D) g (3) (D), g (2) (D) g (3) (D) = (1 + D6 + D11 + D13) (1 + D3 + D15 + D19) , (1 + D8 + D17 + D18) (1 + D3 + D15 + D19) , (15) non-systematic polynomial parity-check matrix Hns (D) = g (1) (D), g (2) (D), g (3) (D) = [1 + D6 + D11 + D13 , 1 + D8 + D17 + D18 , 1 + D3 + D15 + D19] (16) and corresponding non-… view at source ↗

**Figure 3.** Performance comparison of two CSOCs, one systematic [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: BER performance of two high-rate CSOC-based SC-LDPC [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]

**Figure 5.** Figure 5: (a) The multi-edge protograph representing the base [PITH_FULL_IMAGE:figures/full_fig_p014_5.png]

**Figure 6.** Figure 6: Performance of SC-LDPC codes obtained by lifting conv [PITH_FULL_IMAGE:figures/full_fig_p016_6.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

METTLE: Efficient Streaming Erasure Code with Peeling Decodability
cs.IT 2026-02 conditional novelty 7.0 of 10

METTLE achieves streaming erasure decoding with near-RaptorQ efficiency and LT-like peeling speed by turning Walzer's spatially-coupled IBLT into a time-coupled code with multi-edge-type and touch-less leading edge.

Reference graph

Works this paper leans on

23 extracted references · 22 canonical work pages · cited by 1 Pith paper

[7]

Spat ially coupled LDPC codes constructed from protographs,

D. G. M. Mitchell, M. Lentmaier, and D. J. Costello, “Spat ially coupled LDPC codes constructed from protographs,” IEEE Transactions on Information Theory , vol. 61, no. 9, pp. 4866-4889, Sept. 2015

work page 2015
[17]

Q uasi-cyclic LDPC codes based on pre-lifted protographs,

D. G. M. Mitchell, R. Smarandache, and D. J. Costello, “Q uasi-cyclic LDPC codes based on pre-lifted protographs,” IEEE Transactions on Information Theory , vol. 60, no. 10, pp. 5856-5874, Oct. 2014

work page 2014
[21]

On the block error rate performance of spatially co upled LDPC codes for streaming applications,

D. G. M. Mitchell, A. E. Pusane, M. Lentmaier, and D. J. Co stello, “On the block error rate performance of spatially co upled LDPC codes for streaming applications,” IEEE Information Theory W orkshop (ITW), Cambridge, UK, 2016, pp. 236-240

work page 2016
[1]

A. J. Felstrom, and K. Sh. Zigangirov, “Time-varying per iodic convolutional codes with low-density parity-check m atrix,“ IEEE Trans. Inf. Theory , vol. 45, no. 6, pp. 2181-2191, Sep. 1999

work page 1999
[2]

Iterative decoding threshold analysis for L DPC convolutional codes,

M. Lentmaier, A. Sridharan, D. J. Costello, Jr., and K. Sh . Zigangirov, “Iterative decoding threshold analysis for L DPC convolutional codes,” IEEE Trans. Inf. Theory , vol. 56, no. 10, pp. 5274-5289, Oct. 2010

work page 2010
[3]

Thresho ld saturation via spatial coupling: Why convolutional LDPC ensembles perform so well over the BEC,

S. Kudekar, T. J. Richardson, and R. L. Urbanke, “Thresho ld saturation via spatial coupling: Why convolutional LDPC ensembles perform so well over the BEC,” IEEE Trans. Inf. Theory , vol. 57, no. 2, pp. 803-834, Feb. 2011

work page 2011
[4]

Windowed decoding of proto graph-based LDPC convolutional codes over erasure channels,

A. R. Iyengar, M. Papaleo, P . H. Siegel, J. K. Wolf, A. V ane lli-Coralli, and G. E. Corazza, “Windowed decoding of proto graph-based LDPC convolutional codes over erasure channels,” IEEE Trans. Inf. Theory , vol. 58, no. 4, pp. 2303-2320, Apr. 2012

work page 2012
[5]

Staircase codes: FEC for 100 Gb/s OTN,

B. P . Smith, A. Farhood, A. Hunt, F. R. Kschischang and J. L odge, “Staircase codes: FEC for 100 Gb/s OTN,” J. Lightwave Technol. , vol. 30, no. 1, pp. 110-117, 2012

work page 2012

Show all 23 references

[6]

Threshold Decoding

J. L. Massey, “Threshold Decoding”, MIT Press, Cambridg e, MA, 1963

1963
[8]

A class of binary recurren t codes with limited error propagation,

J. Robinson and A. Bernstein, “A class of binary recurren t codes with limited error propagation,” IEEE Transactions on Information Theory, vol. 13, no. 1, pp. 106-113, January 1967

1967
[9]

Low-density parity-check (LDPC) codes cons tructed from protographs

J. Thorpe, “Low-density parity-check (LDPC) codes cons tructed from protographs”, IPN progress report , vol. 42, no. 154, pp. 42-154, 2003

2003
[10]

New convolutional codes - part I,

W. W. Wu, “New convolutional codes - part I,” IEEE Transactions on Communications , vol. COM-23, no. 9, pp. 942-955, Sept. 1975

1975
[11]

New convolutional codes - part II,

W. W. Wu, “New convolutional codes - part II,” IEEE Transactions on Communications , vol. COM-24, no. 1, pp. 19-33, Jan. 1976

1976
[12]

Some difference triangles for constructi ng self-orthogonal codes,

E. Klieber, “Some difference triangles for constructi ng self-orthogonal codes,” IEEE Transactions on Information Theory (Corresp.), vol. 16, no. 2, pp. 237-238, March 1970

1970
[13]

Search and dete rmination of convolutional self-doubly orthogonal codes f or iterative threshold decoding,

D. Haccoun, C. Cardinal, and F. Gagnon, “Search and dete rmination of convolutional self-doubly orthogonal codes f or iterative threshold decoding,” IEEE Transactions on Communications , vol. 53, no. 5, pp. 802-809, May 2005

2005
[14]

Moderate-density parity-check codes,

S. Ouzan and Y . Beery, “Moderate-density parity-check codes,” CoRR, vol. abs/0911.3262, 2009

2009 arXiv
[15]

Designing protograph-based quasi-cyc lic spatially coupled LDPC codes with large girth,

S. Mo, L. Chen, D. J. Costello, D. G. M. Mitchell, R. Smara ndache, and J. Qiu, “Designing protograph-based quasi-cyc lic spatially coupled LDPC codes with large girth,” IEEE Transactions on Communications , vol. 68, no. 9, pp. 5326-5337, Sept. 2020

2020
[16]

Finite- length construction of high performance spatially-couple d codes via optimized partitioning and lifting,

H. Esfahanizadeh, A. Hareedy, and L. Dolecek, “Finite- length construction of high performance spatially-couple d codes via optimized partitioning and lifting,” IEEE Transactions on Communications , vol. 67, no. 1, pp. 3-16, Jan. 2019

2019
[18]

Performance comparison of LDPC block and spatially cou pled codes over GF(q),

K. Huang, D. G. M. Mitchell, L. Wei, X. Ma, and D. J. Costel lo, “Performance comparison of LDPC block and spatially cou pled codes over GF(q),” IEEE Transactions on Communications , vol. 63, no. 3, pp. 592-604, March 2015

2015
[19]

Channel c oding rate in the ﬁnite blocklength regime,

Y . Polyanskiy, H. Vincent Poor, and S. V erdú, “Channel c oding rate in the ﬁnite blocklength regime,” IEEE Trans. Inf. Theory , vol. 56, no. 5, pp. 2307-2359, May 2010

2010
[20]

A low complexity PEG-like algorithm to construct quasi-c yclic LDPC codes,

A. Gomez-Fonseca, R. Smarandache, and D. G. M. Mitchell , “A low complexity PEG-like algorithm to construct quasi-c yclic LDPC codes,” Proc. International Symposium on Topics in Coding , pp. 1-5, Brest, France, Sept. 2023

2023
[22]

Protograph LDPC codes design bas ed on EXIT analysis,

G. Liva and M. Chiani, “Protograph LDPC codes design bas ed on EXIT analysis,” Proc. IEEE Global Telecommunications Conference , Washington, DC, USA, 2007, pp. 3250-3254

2007
[23]

Low Density Parity Check Codes,

R. G. Gallager, “Low Density Parity Check Codes,” Cambr idge, MA, MIT Press, 1963

1963

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.

[1] [7]

Spat ially coupled LDPC codes constructed from protographs,

D. G. M. Mitchell, M. Lentmaier, and D. J. Costello, “Spat ially coupled LDPC codes constructed from protographs,” IEEE Transactions on Information Theory , vol. 61, no. 9, pp. 4866-4889, Sept. 2015

work page 2015

[2] [17]

Q uasi-cyclic LDPC codes based on pre-lifted protographs,

D. G. M. Mitchell, R. Smarandache, and D. J. Costello, “Q uasi-cyclic LDPC codes based on pre-lifted protographs,” IEEE Transactions on Information Theory , vol. 60, no. 10, pp. 5856-5874, Oct. 2014

work page 2014

[3] [21]

On the block error rate performance of spatially co upled LDPC codes for streaming applications,

D. G. M. Mitchell, A. E. Pusane, M. Lentmaier, and D. J. Co stello, “On the block error rate performance of spatially co upled LDPC codes for streaming applications,” IEEE Information Theory W orkshop (ITW), Cambridge, UK, 2016, pp. 236-240

work page 2016

[4] [1]

A. J. Felstrom, and K. Sh. Zigangirov, “Time-varying per iodic convolutional codes with low-density parity-check m atrix,“ IEEE Trans. Inf. Theory , vol. 45, no. 6, pp. 2181-2191, Sep. 1999

work page 1999

[5] [2]

Iterative decoding threshold analysis for L DPC convolutional codes,

M. Lentmaier, A. Sridharan, D. J. Costello, Jr., and K. Sh . Zigangirov, “Iterative decoding threshold analysis for L DPC convolutional codes,” IEEE Trans. Inf. Theory , vol. 56, no. 10, pp. 5274-5289, Oct. 2010

work page 2010

[6] [3]

Thresho ld saturation via spatial coupling: Why convolutional LDPC ensembles perform so well over the BEC,

S. Kudekar, T. J. Richardson, and R. L. Urbanke, “Thresho ld saturation via spatial coupling: Why convolutional LDPC ensembles perform so well over the BEC,” IEEE Trans. Inf. Theory , vol. 57, no. 2, pp. 803-834, Feb. 2011

work page 2011

[7] [4]

Windowed decoding of proto graph-based LDPC convolutional codes over erasure channels,

A. R. Iyengar, M. Papaleo, P . H. Siegel, J. K. Wolf, A. V ane lli-Coralli, and G. E. Corazza, “Windowed decoding of proto graph-based LDPC convolutional codes over erasure channels,” IEEE Trans. Inf. Theory , vol. 58, no. 4, pp. 2303-2320, Apr. 2012

work page 2012

[8] [5]

Staircase codes: FEC for 100 Gb/s OTN,

B. P . Smith, A. Farhood, A. Hunt, F. R. Kschischang and J. L odge, “Staircase codes: FEC for 100 Gb/s OTN,” J. Lightwave Technol. , vol. 30, no. 1, pp. 110-117, 2012

work page 2012