Pith. sign in

REVIEW 3 major objections 2 minor 1 cited by

The high-speed X-ray camera on AXIS: design and performance updates

T0 review · 3 major / 2 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read The AXIS High-Speed Camera's prototype CCDs and ASIC chain achieve the required fast readout while preserving X-ray spectral response, according to the paper's reported measurements.

desk verdict The submission is two papers glued together: an AXIS X-ray camera abstract with no body, and an unrelated MLLM paper—so the camera's central claim is uncheckable. read the letter →

arxiv 2508.14157 v1 pith:NZVQ4ZBA submitted 2025-08-19 astro-ph.IM

Eric D. Miller , Catherine E. Grant , Robert Goeke , Marshall W. Bautz , Christopher Leitz , Kevan Donlon , Steven W. Allen , Sven Herrmann
show 20 more authors
Abraham D. Falcone F. Elio Angile Tanmoy Chattopadhyay Michael Cooper Mallory A. Jensen Jill Juneau Beverly LaMarr Andrew Malonis R. Glenn Morris Peter Orel Abigail Y. Pan Steven Persyn Artem Poliszczuk Gregory Y. Prigozhin Ilya Prigozhin Andrew Ptak Christopher Reynolds Haley R. Stueber Keith Warner Daniel R. Wilkins
This is my paper · ORCID
classification astro-ph.IM PACS 95.55.Ka
keywords AXISmissionX-rayCCDhigh-speedreadoutlow-noiseASICspectroscopytransientalertsCCID100PhaseA
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

AXIS is a NASA Probe mission concept needing a camera that reads out large X-ray CCDs more than an order of magnitude faster than previous missions without sacrificing the spectral resolution that makes high-resolution X-ray spectroscopy possible. This paper updates the design of that High-Speed Camera and reports prototype measurements from MIT and Stanford test facilities. The claim is that the detector chain—Lincoln Laboratory CCID100 CCDs with multiple high-speed, low-noise amplifiers, Stanford's low-power ASIC readout, and digital front-end signal processing—achieves the required readout rate with excellent X-ray spectral response. If true, the camera would meet AXIS mission requirements and enable its planned studies of fast-changing high-energy sources. The first lot of CCID100 CCDs has just completed fabrication and is slated for X-ray performance testing, so the decisive end-to-end validation is still ahead.

What carries the argument

The central mechanism is parallel, low-power CCD readout: each CCID100 CCD uses a single-layer polysilicon gate structure for fast, low-power clock transfers and multiple high-speed, low-noise output amplifiers, so a large frame is read through many channels at once; a Stanford-designed ASIC replaces conventional discrete readout electronics, and the front-end captures digital video waveforms with advanced signal processing to keep noise low at high speed. Together these parts let the camera move charge and digitize events faster than heritage detectors while preserving the low readout noise needed for Fano-limited X-ray spectroscopy.

What would settle it

Run the flight-like CCID100 plus ASIC chain under X-ray illumination at the mission's target frame rate and measure equivalent noise charge and line width; if the readout noise per pixel is high enough to broaden a monoenergetic X-ray line beyond the mission's spectral-resolution requirement at that rate, the central claim is refuted.

Watch

Extended reading notes

Core claim

The paper's central claim is that the AXIS High-Speed Camera's detector chain—MIT Lincoln Laboratory CCID100 CCDs with multiple high-speed, low-noise output amplifiers and a single-layer polysilicon gate structure, read out by Stanford-designed low-power ASICs and digital waveform-capture front-end electronics—achieves the mission's required readout rate while preserving excellent X-ray spectral response. Prototype devices have been built and tested at MIT and Stanford test facilities; the measurements are reported as evidence that the camera will meet AXIS mission requirements. The first lot of CCID100 CCDs has completed fabrication and is about to begin X-ray performance testing, which wil

Load-bearing premise

That prototype CCDs and ASICs measured in MIT and Stanford labs will behave the same when integrated into the flight camera at the mission's required readout rate.

Editorial extensions

If this is right

  • AXIS will be able to read out its large-format CCDs more than ten times faster than heritage X-ray cameras while keeping the spectral response needed for high-resolution X-ray imaging.
  • The High-Speed Camera design is on track to meet mission requirements at the end of Phase A, supporting the AXIS Probe mission's science case.
  • Back-end electronics can identify candidate X-ray events at high speed and issue transient alerts, enabling rapid follow-up of changing sources.
  • Fast readout plus low noise means the detector can record bright, rapidly varying high-energy sources without the level of pile-up that limited prior missions.
  • The completed CCID100 prototype lot, once X-ray tested, will provide the first end-to-end validation of the full detector chain.

Reading between the lines

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

  • If the prototype results transfer to the flight camera, AXIS should be able to do sub-second time-resolved spectroscopy of bright transients—gamma-ray bursts, stellar flares, accreting black holes—without pile-up, a capability no current X-ray CCD mission has.
  • The readout architecture is not AXIS-specific: the combination of many on-chip amplifiers and an ASIC per channel could be adopted by other X-ray and UV instruments that need low noise at high frame rates, if the ASIC's power and noise scale as measured.
  • Because this submission reports integrated measurements without showing the underlying noise spectra and line-width data, the public case for meeting requirements will stand or fall on the upcoming CCID100 X-ray tests; those tests should be published with noise-versus-rate curves.
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.

Referee Report

3 major / 2 minor

Summary. The submission under review consists of an abstract that describes design and performance updates for the AXIS High-Speed Camera, an X-ray detector system for the AXIS Probe mission. The abstract's central claim is that prototype measurements 'achieve excellent spectral response at the required readout rate, demonstrating that we will meet mission requirements.' However, the full text of the manuscript is arXiv:2508.14160v2, an unrelated paper on the RynnEC embodied video multimodal large language model. The body contains no X-ray detector measurements, no noise or spectral-resolution data, no readout-rate analysis, no test descriptions, and no comparison against AXIS requirements. The central empirical claim of the abstract is therefore entirely unsupported by any content in the manuscript.

Significance. If the abstract's claim were substantiated, the result would be important: a large-format CCD camera reading out more than an order of magnitude faster than heritage X-ray instruments while retaining Fano-limited spectral resolution would be a key enabling technology for the AXIS Probe science goals. The claimed combination of speed and spectral performance is nontrivial and of clear interest to the high-energy astrophysics community. However, as submitted, the manuscript provides no evidence whatsoever for this claim. The body text is a different paper, so the claimed prototype measurements, test facilities, and performance comparisons are entirely absent. The significance of the possible result cannot be assessed from the submitted document.

major comments (3)
  1. [Abstract / Full text] The abstract's final paragraph makes a load-bearing empirical claim: 'These measurements achieve excellent spectral response at the required readout rate, demonstrating that we will meet mission requirements.' No supporting data appear anywhere in the full text. Instead, the full text is an unrelated MLLM paper (RynnEC, arXiv:2508.14160v2) containing no X-ray detector measurements, no readout noise figures, no spectral FWHM, no operating temperature, no event-rate analysis, and no comparison to AXIS requirements. The central claim cannot be checked or reproduced.
  2. [Abstract, second sentence of camera description] The abstract states that a first lot of prototype CCID100 CCDs 'has completed fabrication and will soon begin X-ray performance testing,' yet later states that current measurements 'achieve excellent spectral response at the required readout rate.' If testing has not yet begun, the latter assertion is unsupported; if testing has begun, the earlier sentence is misleading. Either way, the manuscript does not report the measurements themselves, so the requirements demonstration is missing.
  3. [Full text (RynnEC paper)] The manuscript body is a self-contained paper on embodied video MLLMs, with its own abstract, figures, tables, references, and appendices. It has no connection to the AXIS High-Speed Camera described in the submitted abstract. This is not a minor copy-and-paste typo; it is a structural absence of the actual paper promised by the title and abstract. The referee cannot evaluate the design updates, prototype performance, or mission-readiness claims because none of that content is present.
minor comments (2)
  1. [General] The abstract contains no figures, tables, or section references. If the intended camera paper were to be submitted, it would need quantitative presentation of measured noise versus readout rate, spectral resolution at relevant X-ray lines, and a requirements traceability table.
  2. [Full text metadata] The full text lists authors and affiliations for RynnEC, none of whom are the MIT/Stanford/Penn State/SwRI collaboration named in the AXIS abstract. The arXiv identifiers and dates also differ. This mismatch should be resolved by the authors before any resubmission.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity detected: the abstract makes an empirical claim against external mission requirements, and the body text (an unrelated MLLM paper) contains no derivation chain that could reduce to its own inputs.

full rationale

The submission pairs an abstract for the AXIS High-Speed Camera with a full text that is actually the RynnEC paper on embodied video MLLMs. The abstract's load-bearing sentence, 'These measurements achieve excellent spectral response at the required readout rate, demonstrating that we will meet mission requirements,' is a report of measurements against an external benchmark (AXIS mission requirements). It is not derived from fitted parameters, nor is it justified by self-citation, nor does it rename a known result. The full text contains no X-ray detector data, noise figures, spectral resolution measurements, or test descriptions at all, so the central empirical claim cannot be verified from the supplied text. However, absence of supporting data is a completeness/verifiability problem, not a circularity problem. No equation or argument in the manuscript reduces to its own input, no fitted quantity is relabeled as a prediction, and no load-bearing uniqueness or ansatz is imported through self-citation. Under the instruction that circularity may be claimed only when a specific reduction can be quoted and exhibited, none exists here. The appropriate finding is no significant circularity, score 0.

Assumptions & free parameters 0 free parameters · 2 assumptions · 0 invented entities

On abstract-only evidence, with the full text belonging to a different paper, no free parameters are visible and no invented entities are proposed; the camera is engineered hardware, not a postulated new entity. The central claims rest on two domain assumptions about prototype-to-flight transfer and heritage CCD physics. A complete ledger requires the actual AXIS paper body, including its measured noise curves and test conditions.

assumptions (2)
  • domain assumption Performance measured on prototype CCID100 CCDs and Stanford ASICs in lab test facilities represents the performance of the flight camera.
    The abstract's final claim, that the mission will meet requirements, is an extrapolation from prototype lab measurements, and the submission contains no flight-like test evidence.
  • domain assumption Heritage X-ray CCD readout physics (noise, charge transfer, spectral response models) transfers to the new single-layer polysilicon gate structures.
    The abstract asserts 'maintaining excellent spectral performance' at higher speed, which presupposes no unknown loss mechanism introduced by the new gate structure or high-speed clocking.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The high-speed X-ray camera on AXIS: design and performance updates." pith.science (2026). https://pith.science/paper/NZVQ4ZBA

@misc{pith2026250814157,
  author       = {Pith},
  title        = {Pith review of: The high-speed X-ray camera on AXIS: design and performance updates},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NZVQ4ZBA}},
  note         = {Machine review of arXiv:2508.14157}
}
read the original abstract

AXIS, a Probe mission concept now in a Phase A study, will provide transformative studies of high-energy astrophysical phenomena thanks to its high-resolution X-ray spectral imaging. These capabilities are enabled by improvements to the mirror design that greatly increase the X-ray throughput per unit mass; and to the detector system, which operates more than an order of magnitude faster than heritage instruments while maintaining excellent spectral performance. We present updates to the design of the AXIS High-Speed Camera, a collaborative effort by MIT, Stanford University, the Pennsylvania State University, and the Southwest Research Institute. The camera employs large-format MIT Lincoln Laboratory CCDs that feature multiple high-speed, low-noise output amplifiers and an advanced single-layer polysilicon gate structure for fast, low-power clock transfers. A first lot of prototype CCID100 CCDs has completed fabrication and will soon begin X-ray performance testing. The CCDs are paired with high-speed, low-noise ASIC readout chips designed by Stanford to provide better performance than conventional discrete solutions at a fraction of the power consumption and footprint. Complementary Front-End Electronics employ state-of-the-art digital video waveform capture and advanced signal processing to further deliver low noise at high speed. The Back-End Electronics provide high-speed identification of candidate X-ray events and transient monitoring that relays fast alerts of changing sources to the community. We highlight updates to our parallel X-ray performance test facilities at MIT and Stanford, and review the current performance of the CCD and ASIC technology from testing of prototype devices. These measurements achieve excellent spectral response at the required readout rate, demonstrating that we will meet mission requirements and enable AXIS to achieve world-class science.

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

  1. A Systematic Study of Deep Learning Models and xAI Methods for Region-of-Interest Detection in MRI Scans

    eess.IV 2025-08 unverdicted novelty 2.0 of 10

    A systematic comparison of deep learning models and xAI methods for ROI detection in knee MRI, based on the abstract.

Reference graph

Works this paper leans on

18 extracted references · 9 canonical work pages · cited by 1 Pith paper

  1. [27]

    Llava-onevision: Easy visual task transfer.arXivpreprintarXiv:2408.03326, 2024

    Bo Li, Yuanhan Zhang, Dong Guo, Renrui Zhang, Feng Li, Hao Zhang, Kaichen Zhang, Peiyuan Zhang, Yanwei Li, Ziwei Liu, et al. Llava-onevision: Easy visual task transfer.arXivpreprintarXiv:2408.03326, 2024

    arXiv 2024

  2. [28]

    Mvbench: A comprehensive multi-modal video understanding benchmark

    Kunchang Li, Yali Wang, Yinan He, Yizhuo Li, Yi Wang, Yi Liu, Zun Wang, Jilan Xu, Guo Chen, Ping Luo, et al. Mvbench: A comprehensive multi-modal video understanding benchmark. InProceedingsoftheIEEE/CVF ConferenceonComputerVisionandPatternRecognition, pages 22195–22206, 2024

    work page 2024

  3. [29]

    Manipllm: Embodied multimodal large language model for object-centric robotic manipulation

    Xiaoqi Li, Mingxu Zhang, Yiran Geng, Haoran Geng, Yuxing Long, Yan Shen, Renrui Zhang, Jiaming Liu, and Hao Dong. Manipllm: Embodied multimodal large language model for object-centric robotic manipulation. In ProceedingsoftheIEEE/CVFConferenceonComputerVisionandPatternRecognition, pages 18061–18070, 2024

    work page 2024

  4. [30]

    Industryeqa: Pushing the frontiers of embodied question answering in industrial scenarios.arXivpreprintarXiv:2505.20640, 2025

    Yifan Li, Yuhang Chen, Anh Dao, Lichi Li, Zhongyi Cai, Zhen Tan, Tianlong Chen, and Yu Kong. Industryeqa: Pushing the frontiers of embodied question answering in industrial scenarios.arXivpreprintarXiv:2505.20640, 2025

    arXiv 2025

  5. [31]

    Sti-bench: Are mllms ready for precise spatial-temporal world understanding?arXivpreprintarXiv:2503.23765, 2025

    Yun Li, Yiming Zhang, Tao Lin, XiangRui Liu, Wenxiao Cai, Zheng Liu, and Bo Zhao. Sti-bench: Are mllms ready for precise spatial-temporal world understanding?arXivpreprintarXiv:2503.23765, 2025

    arXiv 2025

  6. [32]

    Describe anything: Detailed localized image and video captioning.arXivpreprint arXiv:2504.16072, 2025

    Long Lian, Yifan Ding, Yunhao Ge, Sifei Liu, Hanzi Mao, Boyi Li, Marco Pavone, Ming-Yu Liu, Trevor Darrell, Adam Yala, et al. Describe anything: Detailed localized image and video captioning.arXivpreprint arXiv:2504.16072, 2025

    arXiv 2025

  7. [33]

    Video-llava: Learning united visual representation by alignment before projection.arXivpreprintarXiv:2311.10122, 2023

    Bin Lin, Yang Ye, Bin Zhu, Jiaxi Cui, Munan Ning, Peng Jin, and Li Yuan. Video-llava: Learning united visual representation by alignment before projection.arXivpreprintarXiv:2311.10122, 2023

    arXiv 2023

  8. [34]

    Draw-and-understand: Leveraging visual prompts to enable mllms to comprehend what you want.arXivpreprintarXiv:2403.20271, 2024

    Weifeng Lin, Xinyu Wei, Ruichuan An, Peng Gao, Bocheng Zou, Yulin Luo, Siyuan Huang, Shanghang Zhang, and Hongsheng Li. Draw-and-understand: Leveraging visual prompts to enable mllms to comprehend what you want.arXivpreprintarXiv:2403.20271, 2024

    arXiv 2024

Show all 18 references
  1. [35]

    Perceive anything: Recognize, explain, caption, and segment anything in images and videos, 2025

    Weifeng Lin, Xinyu Wei, Ruichuan An, Tianhe Ren, Tingwei Chen, Renrui Zhang, Ziyu Guo, Wentao Zhang, Lei Zhang, and Hongsheng Li. Perceive anything: Recognize, explain, caption, and segment anything in images and videos, 2025

    2025

  2. [36]

    Visual instruction tuning.Advancesinneural informationprocessingsystems, 36:34892–34916, 2023

    Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. Visual instruction tuning.Advancesinneural informationprocessingsystems, 36:34892–34916, 2023

    2023

  3. [37]

    Visual embodied brain: Let multimodal large language models see, think, and control in spaces.arXivpreprintarXiv:2506.00123, 2025

    Gen Luo, Ganlin Yang, Ziyang Gong, Guanzhou Chen, Haonan Duan, Erfei Cui, Ronglei Tong, Zhi Hou, Tianyi Zhang, Zhe Chen, et al. Visual embodied brain: Let multimodal large language models see, think, and control in spaces.arXivpreprintarXiv:2506.00123, 2025

    2025 arXiv

  4. [38]

    Mmscan: A multi-modal 3d scene dataset with hierarchical grounded language annotations.AdvancesinNeuralInformationProcessingSystems, 37:50898–50924, 2024

    Ruiyuan Lyu, Jingli Lin, Tai Wang, Xiaohan Mao, Yilun Chen, Runsen Xu, Haifeng Huang, Chenming Zhu, Dahua Lin, and Jiangmiao Pang. Mmscan: A multi-modal 3d scene dataset with hierarchical grounded language annotations.AdvancesinNeuralInformationProcessingSystems, 37:50898–50924, 2024

    2024

  5. [39]

    Video-chatgpt: Towards detailed video understanding via large vision and language models.arXivpreprintarXiv:2306.05424, 2023

    Muhammad Maaz, Hanoona Rasheed, Salman Khan, and Fahad Shahbaz Khan. Video-chatgpt: Towards detailed video understanding via large vision and language models.arXivpreprintarXiv:2306.05424, 2023

    2023 arXiv

  6. [40]

    Videogpt+: Integrating image and video encoders for enhanced video understanding.arXivpreprintarXiv:2406.09418, 2024

    Muhammad Maaz, Hanoona Rasheed, Salman Khan, and Fahad Khan. Videogpt+: Integrating image and video encoders for enhanced video understanding.arXivpreprintarXiv:2406.09418, 2024

    2024 arXiv

  7. [41]

    Openeqa: Embodied question answering in the era of foundation models

    Arjun Majumdar, Anurag Ajay, Xiaohan Zhang, Pranav Putta, Sriram Yenamandra, Mikael Henaff, Sneha Silwal, Paul Mcvay, Oleksandr Maksymets, Sergio Arnaud, et al. Openeqa: Embodied question answering in the era of foundation models. InProceedingsoftheIEEE/CVFconferenceoncomputer...

    2024

  8. [42]

    Generation and comprehension of unambiguous object descriptions

    Junhua Mao, Jonathan Huang, Alexander Toshev, Oana Camburu, Alan L Yuille, and Kevin Murphy. Generation and comprehension of unambiguous object descriptions. InProceedingsoftheIEEEconferenceoncomputervision andpatternrecognition, pages 11–20, 2016

    2016

  9. [43]

    Videoglamm: A large multimodal model for pixel-level visual grounding in videos

    Shehan Munasinghe, Hanan Gani, Wenqi Zhu, Jiale Cao, Eric Xing, Fahad Shahbaz Khan, and Salman Khan. Videoglamm: A large multimodal model for pixel-level visual grounding in videos. InProceedingsoftheComputer VisionandPatternRecognitionConference, pages 19036–19046, 2025

    2025

  10. [44]

    kitchen" or

    Riku Murai, Eric Dexheimer, and Andrew J Davison. Mast3r-slam: Real-time dense slam with 3d reconstruction priors. InProceedingsoftheComputerVisionandPatternRecognitionConference, pages 16695–16705, 2025. Appendix A Implementation Details for Data Pipeline A.1 Instance Segment...

    2025

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.