Pith. sign in

REVIEW 2 major objections 1 minor 47 references

Symbol Rate-Code Rate Trade-offs for IM/DD 200G/400G per Lane LPO Transceivers

T0 review · 2 major / 1 minor · reviewed 2026-06-26 · grok-4.3

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

Pith's one-line read Reducing FEC code rate below KP4 allows higher-order PAM to use lower symbol rates for 200G/400G IM/DD throughput

desk verdict The paper applies standard BSC capacity to show that lower code rates let PAM-6/8 drop symbol rate versus PAM-4 for 200G/400G LPO targets, but the model choice is the main limit on how far to trust the numbers. read the letter →

arxiv 2606.23390 v1 pith:AD47CW2P submitted 2026-06-22 eess.SP

José Núñez-Kasaneva , Yunus Can Gültekin , Stefanos Dris , Paraskevas Bakopoulos , Nikos Argyris , Gabriele Liga This is my paper · ORCID
classification eess.SP
keywords IM/DDPAM-4PAM-6PAM-8FECsymbolratecodeLPO
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 investigates symbol rate and code rate trade-offs in bandwidth-limited IM/DD systems for LPO transceivers using PAM-4, PAM-6 and PAM-8. It employs the capacity of the binary symmetric channel under hard-decision decoding to assess achievable information rates for 200G and 400G per lane targets. The analysis demonstrates that decreasing the code rate below the KP4 baseline enables higher-order PAM formats to meet the throughput at substantially lower symbol rates than PAM-4. A sympathetic reader would care because this could relax the bandwidth requirements on transceivers and associated electronics.

What carries the argument

The symbol rate-code rate trade-off analysis based on binary symmetric channel capacity for different PAM orders in IM/DD systems.

What would settle it

An experiment measuring the bit error rates or throughputs achieved by PAM-6 and PAM-8 at the lower symbol rates predicted by the model and finding they do not reach the 200G or 400G targets under realistic IM/DD conditions.

Watch

Extended reading notes

Core claim

Reducing the FEC code rate below the KP4 baseline allows higher-order PAM formats to operate at substantially lower symbol rates than PAM-4 while meeting the throughput requirement, as shown by using the binary symmetric channel capacity as an achievable information rate metric under hard-decision decoding for bandwidth-limited IM/DD channels.

Load-bearing premise

The binary symmetric channel capacity under hard-decision decoding serves as a reliable achievable information rate proxy for the bandwidth-limited IM/DD channels considered.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 1 minor

Summary. The manuscript analyzes symbol rate versus FEC code rate trade-offs for bandwidth-limited IM/DD LPO transceivers targeting 200 Gbps and 400 Gbps per lane. Using PAM-4, PAM-6, and PAM-8 formats, it adopts the binary symmetric channel capacity under hard-decision decoding as the achievable information rate metric and concludes that code rates below the KP4 baseline enable higher-order PAM formats to meet the throughput target at substantially lower symbol rates than PAM-4.

Significance. If the BSC proxy accurately reflects post-equalization performance, the trade-off curves would offer a practical guideline for choosing modulation order and code rate in high-speed optical transceiver design. The approach is computationally lightweight and directly ties information-theoretic rates to the 200G/400G throughput constraint, which is a useful first-order design tool.

major comments (2)
  1. [Abstract] Abstract: the central claim that lower code rates permit lower symbol rates with PAM-6/8 rests entirely on the BSC capacity 1−h(p) serving as a faithful achievable-rate proxy. The manuscript supplies no derivation details, error analysis, or comparison against full IM/DD channel models that incorporate signal-dependent noise and residual ISI.
  2. [Metric definition and results sections] Metric definition and results sections: the assumption that the effective channel after equalization is a memoryless BSC with uniform crossover probability is load-bearing for the reported symbol-rate reductions. IM/DD links with higher-order PAM typically exhibit signal-dependent noise, non-uniform bit-error rates across levels, and memory from bandwidth limitation, which can invalidate the BSC model and alter the predicted trade-offs.
minor comments (1)
  1. [Abstract] The abstract would be clearer if it stated the exact KP4 code rate and the range of code rates examined.

Simulated Author's Rebuttal

2 responses · 0 unresolved

Thank you for the opportunity to respond to the referee's comments. We clarify the role of the BSC capacity metric in our analysis and address the concerns regarding its applicability to IM/DD channels.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim that lower code rates permit lower symbol rates with PAM-6/8 rests entirely on the BSC capacity 1−h(p) serving as a faithful achievable-rate proxy. The manuscript supplies no derivation details, error analysis, or comparison against full IM/DD channel models that incorporate signal-dependent noise and residual ISI.

    Authors: The BSC capacity is employed as a standard information-theoretic metric for the achievable rate under hard-decision decoding in a memoryless binary channel, which serves as a first-order approximation for post-equalization performance in bandwidth-limited IM/DD systems. The derivation of the capacity 1 - h(p) is well-established in information theory literature, and we can include a brief reference to it in the revised manuscript. While a full comparison to more detailed channel models would be valuable, the present work intentionally adopts this lightweight proxy to directly link symbol rate and code rate to the throughput targets, as highlighted in the referee's significance assessment. We will revise the abstract and metric section to explicitly state the assumptions and limitations of this approach. revision: partial

  2. Referee: [Metric definition and results sections] Metric definition and results sections: the assumption that the effective channel after equalization is a memoryless BSC with uniform crossover probability is load-bearing for the reported symbol-rate reductions. IM/DD links with higher-order PAM typically exhibit signal-dependent noise, non-uniform bit-error rates across levels, and memory from bandwidth limitation, which can invalidate the BSC model and alter the predicted trade-offs.

    Authors: We agree that the BSC model is an approximation and that real IM/DD systems exhibit the mentioned impairments. However, the manuscript positions the BSC capacity as a proxy metric rather than an exact model, allowing for the exploration of symbol rate-code rate trade-offs in a computationally efficient manner. The uniform crossover probability is assumed post-equalization to simplify the analysis. To address this, we will expand the discussion in the metric definition section to acknowledge these limitations and note that the reported trade-offs are indicative under the BSC assumption. This does not invalidate the utility of the approach as a design guideline. revision: yes

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity; standard BSC capacity metric applied independently of paper inputs

full rationale

The paper selects the binary symmetric channel capacity 1−h(p) under hard-decision decoding as its achievable-rate proxy, a result taken directly from classical information theory and not fitted, redefined, or derived from any of the paper's own simulation outputs, parameters, or prior self-citations. The symbol-rate versus code-rate trade-off analysis for PAM-4/6/8 is performed by comparing this external benchmark against the 200G/400G throughput targets; no equation reduces the claimed performance advantage to a fitted input or to a self-referential definition. The derivation therefore remains self-contained against external benchmarks.

Assumptions & free parameters 0 free parameters · 1 assumptions · 0 invented entities

Only abstract available; ledger populated from stated metric choice.

assumptions (1)
  • domain assumption BSC capacity under HDD is an appropriate achievable rate for the IM/DD channels of interest
    Explicitly adopted as performance metric in abstract.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Symbol Rate-Code Rate Trade-offs for IM/DD 200G/400G per Lane LPO Transceivers." pith.science (2026). https://pith.science/paper/AD47CW2P

@misc{pith2026260623390,
  author       = {Pith},
  title        = {Pith review of: Symbol Rate-Code Rate Trade-offs for IM/DD 200G/400G per Lane LPO Transceivers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AD47CW2P}},
  note         = {Machine review of arXiv:2606.23390}
}
read the original abstract

We analyze the symbol rate-code rate trade-off in bandwidth-limited IM/DD systems targeting LPO transceivers, using PAM-4/6/8 as candidate modulation formats. We use the capacity of the binary symmetric channel as an achievable information rate under hard-decision decoding, serving as a performance metric for 200G and 400G per-lane throughput targets. Our results show that reducing the FEC code rate below the KP4 baseline allows higher-order PAM formats to operate at substantially lower symbol rates than PAM-4 while meeting the throughput requirement.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 17 canonical work pages

  1. [1]

    2506.01761 , archivePrefix=

    Felipe Villenas and Kaiquan Wu and Yunus Can Gültekin and Jamal Riani and Alex Alvarado , year=. 2506.01761 , archivePrefix=

    work page arXiv

  2. [2]

    Chorchos , title =

    L. Chorchos , title =. Metro and Data Center Optical Networks and Short-Reach Links II , editor =. 2019 , doi =

    2019

  3. [3]

    2506.19684 , archivePrefix=

    Felipe Villenas and Kaiquan Wu and Yunus Can Gültekin and Jamal Riani and Alex Alvarado , year=. 2506.19684 , archivePrefix=

    work page arXiv

  4. [4]

    Farhood, Arash and others , title =

  5. [5]

    OFC , journal =

    Charles St-Arnault and Santiago Bernal and Derek Kita and Ross Dickson and Mariam Yehia Abdelaziz and Aleksandar Nikic and Benton Qiu and Benjamin Krueger and Fabio Pittal\`. OFC , journal =. 2025 , url =. doi:10.1364/OFC.2025.Th4B.1 , abstract =

    work page doi:10.1364/ofc.2025.th4b.1 2025

  6. [6]

    Modulation

    Che, Di and Chen, Xi , year =. Modulation. JLT , volume =. doi:10.1109/JLT.2023.3311716 , urldate =

    work page doi:10.1109/jlt.2023.3311716 2023

  7. [7]

    Xie, Fei and Huang, Xiaoqian and Bai, Chenglin and Xu, Hengying and Qiao, Yaojun , year =. High-

  8. [8]

    Electronic Equalization of Transmission Impairments , year =

    Henning B\". Electronic Equalization of Transmission Impairments , year =. Optical Fiber Communications Conference , journal =

Show all 47 references
  1. [9]

    2024 , volume=

    Uchiyama, Asami and Okuda, Shinya and Tsuji, Toshiya and Hokama, Yohei and Shirao, Mizuki and Abe, Kenichi and Yamatoya, Takeshi and Yamauchi, Yasuhiro , booktitle=. 2024 , volume=

    2024

  2. [10]

    2020 , volume=

    Pang, Xiaodan and Ozolins, Oskars and Lin, Rui and Zhang, Lu and Udalcovs, Aleksejs and Xue, Lei and Schatz, Richard and Westergren, Urban and Xiao, Shilin and Hu, Weisheng and Jacobsen, Gunnar and Popov, Sergei and Chen, Jiajia , journal=. 2020 , volume=

    2020

  3. [11]

    2021 , volume=

    Hossain, Md Sabbir-Bin and Wei, Jinlong and Pittalà, Fabio and Stojanović, Nebojša and Calabrò, Stefano and Rahman, Talha and Wettlin, Tom and Xie, Changsong and Kuschnerov, Maxim and Pachnicke, Stephan , booktitle=. 2021 , volume=

    2021

  4. [12]

    Journal of Lightwave Technology , volume =

    Zou, Dongdong and Li, Fan and Li, Zibin and Wang, Wei and Sui, Qi and Cao, Zizheng and Li, Zhaohui , year =. Journal of Lightwave Technology , volume =. doi:10.1109/JLT.2020.2973902 , urldate =

    2020 doi

  5. [13]

    Digital Communications , author =

  6. [14]

    and Lee, Edward A

    Barry, John R. and Lee, Edward A. and Messerschmitt, David G. , date =. Digital. doi:10.1007/978-1-4615-0227-2 , url =

    doi

  7. [15]

    , date =

    Kuo, Sen-Maw and Lee, Bob H. , date =. Real-Time Digital Signal Processing: Implementations, Applications, and Experiments with the

  8. [16]

    Agrawal, Govind P , year=

  9. [17]

    and Tsiklauri, M

    Dikhaminjia, N. and Tsiklauri, M. and Kiguradze, Z. and He, J. and Drewniak, J. and Chada, A. and Mutnury, B. , booktitle=. Combined Optimization of FFE and DFE Equalizations , year=

  10. [18]

    Alvarado, Alex and Fehenberger, Tobias and Chen, Bin and Willems, Frans M. J. , journal=. Achievable Information Rates for Fiber Optics: Applications and Computations , year=

  11. [19]

    Higher-Order Modulation vs Faster-Than-Nyquist PAM-4 for Datacenter IM-DD Optics: An AIR Comparison Under Practical Bandwidth Limits , year=

    Che, Di and Chen, Xi , journal=. Higher-Order Modulation vs Faster-Than-Nyquist PAM-4 for Datacenter IM-DD Optics: An AIR Comparison Under Practical Bandwidth Limits , year=

  12. [20]

    Beyond 1

    Zhou, Xiang and Urata, Ryohei and Liu, Hong , year = 2020, month = jan, journal =. Beyond 1. doi:10.1109/JLT.2019.2956779 , urldate =

    2020 doi

  13. [21]

    doi:10.1109/JLT.2019.2926218 , urldate =

    Hu, Qian and Chagnon, Mathieu and Schuh, Karsten and Buchali, Fred and Bulow, Henning , year = 2019, month = oct, journal =. doi:10.1109/JLT.2019.2926218 , urldate =

    2019 doi

  14. [22]

    Wei, J L and Prodaniuc, C and Zhang, L and Stojanovi. 225

  15. [23]

    and Patra, Lenin and Lutkemeyer, Christian and Castrill

    Nagarajan, Radhakrishnan and Martino, Agustin and Morero, Damian A. and Patra, Lenin and Lutkemeyer, Christian and Castrill. Recent. Journal of Lightwave Technology , volume =. doi:10.1109/JLT.2024.3399032 , urldate =

    2024 doi

  16. [24]

    Zhou, Honghang and Li, Yan and Liu, Yuyang and Yue, Lei and Gao, Chao and Li, Wei and Qiu, Jifang and Guo, Hongxiang and Hong, Xiaobin and Zuo, Yong and Wu, Jian , year = 2019, month = jun, journal =. Recent. doi:10.3390/app9112342 , urldate =

    2019 doi

  17. [25]

    Berikaa, Essam and Alam, Md Samiul and Bernal, Santiago and. Next-. Journal of Lightwave Technology , volume =. doi:10.1109/JLT.2023.3307504 , urldate =

    2023 doi

  18. [26]

    Yamamoto, Shuto and Taniguchi, Hiroki and Nakamura, Masanori and Yamazaki, Etsushi , year = 2025, langid =. 200-

    2025

  19. [27]

    doi:10.1007/978-3-030-16250-4 , urldate =

    Springer. doi:10.1007/978-3-030-16250-4 , urldate =

    doi

  20. [28]

    Information

    Liga, Gabriele and Alvarado, Alex and Agrell, Erik and Bayvel, Polina , year = 2017, month = jan, journal =. Information. doi:10.1109/JLT.2016.2603419 , urldate =

    2017 doi

  21. [29]

    Optics Express , volume =

    On Achievable Rates for Long-Haul Fiber-Optic Communications , author =. Optics Express , volume =. doi:10.1364/oe.23.009183 , urldate =

    doi

  22. [30]

    and Leuthold, Juerg , year = 2024, pages =

    Kulmer, Laurenz and Blatter, Tobias and Kohli, Manuel and Horst, Yannik and Koepfli, Stefan M. and Leuthold, Juerg , year = 2024, pages =. Single. Optical. doi:10.1364/OFC.2024.W4H.5 , urldate =

    2024 doi

  23. [31]

    Impact of

    Zhou, Zichuan and Odedeyi, Temitope and Kelly, Brian and O'Carroll, John and Phelan, Richard and Darwazeh, Izzat and Liu, Zhixin , year = 2020, month = apr, journal =. Impact of. doi:10.1109/JPHOT.2020.2966617 , urldate =

    2020 doi

  24. [32]

    Experimental

    Cao, Juntao and Wei, Haiqiang and Guo, Changjian and Lu, Chao and Lau, Alan Pak Tao and Zhong, Kangping , year = 2025, langid =. Experimental

    2025

  25. [33]

    2025 , volume=

    Tang, Renjie and Wang, Kanan and Xiang, Shuyi and Su, Yu and Cao, Chenyao and He, Yukun and Gui, Xiaoyan , booktitle=. 2025 , volume=

    2025

  26. [34]

    Gallager , title=

    Robert G. Gallager , title=

  27. [35]

    and Windpassinger, C

    Fischer, R.F.H. and Windpassinger, C. and Lampe, A. and Huber, J.B. , booktitle=. Tomlinson-Harashima precoding in space-time transmission for low-rate backward channel , year=

  28. [36]

    1.6-Tbps low-power linear-drive high-density optical interface for machine learning/artificial intelligence , volume =

    Guilhem de Valicourt and Pete Pupalaikis and Randy Giles and Marco Lamponi and Lukas Elsinger and Shawn Liu and Brett Sawyer and Jon Proesel and Son Thai Le and Eugene Ho and Karen Liu and Zhiqui Zhu and Steve Corteselli and Laurent Alloine and Chris Daunt and Mark Ferriss and...

    2025 doi

  29. [37]

    Proceedings of DesignCon 2024 , year =

    Feng Luo and Luoyi Fang and Chuansheng Cheng and Lei Guo and Dayong Shen and Haohao Kang and Anbing Sun and Zhan Su and Xiuguo Jiang , title =. Proceedings of DesignCon 2024 , year =

    2024

  30. [38]

    Achievable Information Rates for Coded Modulation With Hard Decision Decoding for Coherent Fiber-Optic Systems , year=

    Sheikh, Alireza and Amat, Alexandre Graell i and Liva, Gianluigi , journal=. Achievable Information Rates for Coded Modulation With Hard Decision Decoding for Coherent Fiber-Optic Systems , year=

  31. [39]

    Che and X

    D. Che and X. Chen, ``Modulation Format and Digital Signal Processing for IM-DD Optics at Post-200G Era ,'' JLT, vol. 42, no. 2, pp. 588--605, Jan. 2024

    2024

  32. [40]

    St-Arnault, S

    C. St-Arnault, S. Bernal, D. Kita, R. Dickson, M. Y. Abdelaziz, A. Nikic, B. Qiu, B. Krueger, F. Pittal\` a , C. Reimer, B. Beggs, N. Ben-Hamida, and D. V. Plant, `` Net 3.2 Tbps 225 Gbaud PAM4 O-Band IM/DD 2 km Transmission Using FR8 and DR8 with a CMOS 3 nm SerDes and TFLN M...

    2025

  33. [41]

    X. Pang, O. Ozolins, R. Lin, L. Zhang, A. Udalcovs, L. Xue, R. Schatz, U. Westergren, S. Xiao, W. Hu, G. Jacobsen, S. Popov, and J. Chen, `` 200 Gbps/Lane IM/DD Technologies for Short Reach Optical Interconnects ,'' JLT, vol. 38, no. 2, pp. 492--503, 2020

    2020

  34. [42]

    G. P. Agrawal, Fiber-Optic Communication Systems . 1em plus 0.5em minus 0.4em John Wiley & Sons, 2012

    2012

  35. [43]

    Farhood et al., ``Concatenated FEC baseline proposal for 200 Gb/s per lane IM-DD optical PMD ,'' IEEE P802.3dj Ethernet Task Force , Feb

    A. Farhood et al., ``Concatenated FEC baseline proposal for 200 Gb/s per lane IM-DD optical PMD ,'' IEEE P802.3dj Ethernet Task Force , Feb. 2023

    2023

  36. [44]

    Uchiyama, S

    A. Uchiyama, S. Okuda, T. Tsuji, Y. Hokama, M. Shirao, K. Abe, T. Yamatoya, and Y. Yamauchi, `` Demonstration of 155 Gbaud PAM4 and PAM6 EML with Narrow High-Mesa EA Modulator for 400 Gbps per Lane Transmission ,'' in OFC, 2024, pp. 1--3

    2024

  37. [45]

    M. S.-B. Hossain, J. Wei, F. Pittalà, N. Stojanović, S. Calabrò, T. Rahman, T. Wettlin, C. Xie, M. Kuschnerov, and S. Pachnicke, `` Single-Lane 402 Gb/s PAM-8 IM/DD Transmission Based on a Single DAC and an O-Band Commercial EML ,'' in OECC, 2021, pp. 1--3

    2021

  38. [46]

    F. Luo, L. Fang, C. Cheng, L. Guo, D. Shen, H. Kang, A. Sun, Z. Su, and X. Jiang, `` 800G Linear Direct Drive Network System Design & Implementation ,'' in Proceedings of DesignCon 2024. 1em plus 0.5em minus 0.4em Santa Clara, CA, USA: DesignCon, 2024

    2024

  39. [47]

    Sheikh, A

    A. Sheikh, A. G. i. Amat, and G. Liva, ``Achievable information rates for coded modulation with hard decision decoding for coherent fiber-optic systems,'' JLT, vol. 35, no. 23, pp. 5069--5078, 2017

    2017

Pith tools

Reviewed June 26, 2026 · model on record in the stance chip above.