Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Evaluation of scientific CMOS sensors for sky survey applications

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read The Andor Marana scientific CMOS camera delivers about one-percent photometric precision on consecutive sky frames, making it a strong candidate for time-domain surveys that need fast read-out.

desk verdict Useful first characterization of the Andor Marana sCMOS, with standard methods and a real caveat: the headline linearity claim is self-normalized to the sensor's low-signal slope, so smooth low-signal nonlinearity is untested. read the letter →

arxiv 1909.00729 v1 pith:VDVIZ52D submitted 2019-09-02 astro-ph.IM

S.Karpov , A.Bajat , A.Christov , M.Prouza This is my paper · ORCID
classification astro-ph.IM
keywords scientificCMOSsAndorMaranadetectorcharacterizationphotometricprecisionlinearitydarkcurrentrandomtelegraphsignal
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

This paper aims to establish that a modern scientific CMOS camera—the Andor Marana, built around a back-illuminated sensor—has the stability, linearity, and low noise needed for wide-field sky surveys that demand fast read-out. It reports laboratory measurements of dark current, gain, linearity, and pixel temporal stability, plus an on-sky demonstration with a wide-field lens. The central finding is that the camera reaches about one-percent photometric precision on consecutive frames even with critically sampled stellar profiles, while showing almost no sub-pixel or cosmetic-image systematics. A reader should care because CCDs are limited in read-out speed; if this sensor performs as claimed, time-domain surveys could observe at tens of frames per second without giving up photometric accuracy.

What carries the argument

The load-bearing object is the back-illuminated GSense400BSI sCMOS chip inside the Andor Marana camera, read out in parallel through column-level amplifiers and dual 11-bit ADCs that reconstruct 16-bit pixel values. The argument is carried by the photon transfer curve—the dependence of per-pixel temporal variance on mean signal—which the paper uses to measure the gain curve across the full dynamic range. That curve locates the 1500-ADU amplifier transition, the single feature that dominates the sensor's calibration behavior: it changes the effective gain, alters the spatial structure of flat fields, and produces excess pixel noise when a pixel's signal straddles the two amplifiers. The paper also uses per-pixel dark-frame statistics to characterize random telegraph signal pixels and the on-board blemish and anti-glow corrections.

What would settle it

Illuminate the sensor with a calibrated lamp monitored by an independent photodiode, compare pixel values with the predicted signal from zero to 60,000 ADU, and check whether deviations from a single straight line exceed two percent below 30,000 ADU; in parallel, observe a constant star through one night and see whether the differential light curve goes below one percent after photon noise and flat-field errors are removed.

Watch

Extended reading notes

Core claim

The paper's discovery is that the Marana sCMOS camera combines near-perfect signal linearity up to roughly half its 65,535-ADU dynamic range with stable, spatially uncorrelated pixel noise, and that it does so while cutting the fraction of blemish-masked pixels by about thirty times relative to the earlier Andor Neo. The main calibration feature is a transition near 1500 ADU between low-gain and high-gain column amplifiers, where the effective gain drops by nearly a factor of two and flat-field spatial structure changes, yet the linearity discontinuity is typically below two percent. On the sky, unfiltered observations of field stars yield light-curve scatter of about one percent, with no systematic effects above 0.5 percent as a function of sub-pixel position. The paper concludes that the camera is a very promising detector for sky surveys, especially those requiring high temporal resolution.

Load-bearing premise

The linearity claim presumes the sensor is already linear in the 0–1000 ADU reference interval used to define the expected signal, so any nonlinearity that is also present at those low levels is invisible in the reported ratio curve.

Editorial extensions

If this is right

  • A survey built around this sensor can time-sample the sky at up to 48 frames per second while keeping photometric scatter near one percent, directly benefiting fast-transient detection, meteor observations, and satellite tracking.
  • Calibration pipelines must treat the region around 1500 ADU as a special zone, because flat-field structure and effective gain differ above and below the amplifier transition.
  • The near absence of hot columns, bleed trails, and blemish-masked pixels means a larger usable fraction of the sensor area than with typical CCD frames.
  • The lack of horizontal striping in Marana data removes a noise correlation that had to be modeled in earlier Andor sCMOS cameras.

Reading between the lines

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

  • The on-sky demonstration is a single setup with no color filters; a direct extension would be to test whether the one-percent precision survives standard broadband filters and a full range of airmasses over many nights.
  • Because the linearity curve is normalized to the 0–1000 ADU interval, absolute linearity at low signal levels is untested; comparing the sensor against an externally monitored calibrated lamp would settle whether the 'nearly perfect' claim holds from zero signal.
  • The amplifier-transition behavior may depend on temperature and read-out mode; a systematic study varying those parameters would show whether the 1500-ADU boundary is stable enough for long-term survey operations.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper presents a laboratory and on-sky characterization of the Andor Marana scientific CMOS camera for sky-survey use. Lab tests measure dark current and its Anti-Glow correction, the photon transfer curve and effective gain, linearity, pixel noise statistics, blemish fraction, and spatial noise correlations, comparing the camera with the earlier Andor Neo sCMOS. An on-sky sequence with a wide-field lens is used to estimate photometric precision for critically sampled stellar profiles. The main claims are that the Marana has nearly perfect linearity up to roughly half of saturation, a much smaller blemish fraction than the Neo, low spatially uncorrelated read noise, and on-sky photometric precision easily reaching 1%.

Significance. If the claims hold, the paper provides useful quantitative evidence that a modern back-illuminated sCMOS camera can be competitive with CCDs for wide-field time-domain surveys, where high frame rates are needed. The comparison to an existing sCMOS device, the use of standard laboratory methods (dark current regression, photon transfer curve, spatial autocorrelation), and the on-sky validation under realistic critically sampled conditions are genuine strengths. The paper also references public software used for acquisition and analysis, aiding reproducibility. The weakness is that the headline linearity statement is anchored to a fit of the same data, so the most load-bearing laboratory result needs to be re-derived against an independent reference before the conclusion can be accepted at face value.

major comments (3)
  1. [§3.2, Fig. 6] The linearity curve is normalized by a linear slope fitted to the same data over the 0-1000 ADU interval, so the Measured/Expected ratio is constrained to be approximately 1 in that interval by construction. Any smooth nonlinearity below 1000 ADU is therefore invisible, and the stated conclusion that linearity is 'nearly perfect up to approximately half of saturation level' is not supported as an absolute claim. Please re-derive the ratio against an independent reference (for example the photodiode monitor used in the setup), or, at minimum, quantify the constancy of the PTC gain in Fig. 4 over the full 0-1500 ADU range and report its rms scatter; a flat gain curve with a stated tolerance would provide an independent check of linearity at low signal.
  2. [§3.2, Figs. 4 and 6] No per-point uncertainties or error bars are shown for the gain or linearity curves. The claim that the amplifier-transition jump has 'amplitude typically less than a couple of percents' cannot be assessed without knowing the pixel-to-pixel scatter and the uncertainty of the fitted low-signal slope. Please report the dispersion of the Measured/Expected ratio and of the PTC gain in each intensity bin, together with the number of pixels or frames used for each point.
  3. [§4, Fig. 13] The on-sky 1% photometric precision claim needs a stated unit and an error budget. The vertical axis of Fig. 13 is labeled 'Lightcurve RMS' without units; if these are magnitudes, a reference line at 0.01 mag should be drawn to support the 1% statement. In addition, the zero-point model includes a third-order spatial polynomial that can absorb some frame-to-frame variations, so the contribution of this polynomial to the reported scatter should be quantified.
minor comments (5)
  1. [Abstract] The abstract contains grammatical errors: 'a typical CCD detectors' and 'All these makes them promising' should be corrected to standard English.
  2. [§3.3 and Fig. 6 caption] There are typos in 'every pixel pixel values' and in the Figure 6 caption 'with inverval below 1000 ADU'. These should be fixed in revision.
  3. [Fig. 13] The y-axis label 'Lightcurve RMS' should specify that the units are magnitudes (or flux), and the figure would be easier to interpret with a horizontal line marking the 1% level.
  4. [After Conclusions] The paper would benefit from a data availability statement. Since no raw frames or derived data tables are provided, making at least the linearity and PTC data available would significantly strengthen verifiability.
  5. [References] Some references are missing complete DOIs or page ranges (for example Fowler et al. and several conference contributions). Please ensure the reference list conforms to the journal's style.

Circularity Check

1 steps flagged · score 4.0 of 10

Linearity claim is normalized to a slope fitted to the same data below 1000 ADU, so near-perfect low-signal linearity is partly self-defined; other results are independent.

  1. self definitional [Section 3.2, Figure 6 caption; conclusion in Section 5]
    "Linearity curve for a random set of pixels. The curve represent the ratio of an actually measured signal to the one expected for a linear signal scaling with exposure time, with inverval below 1000 ADU used to define a linear slope."

    The 'expected' signal used as the denominator is not an independent photometric reference; it is a linear slope fitted to the same pixels over 0-1000 ADU. Therefore the Measured/Expected ratio is forced to be centered on 1 over exactly the interval that anchors the paper's claim of 'nearly perfect linearity.' Any smooth nonlinearity below 1000 ADU is absorbed into the fitted slope and does not appear in the ratio, so the low-signal part of the linearity claim is tautological rather than demonstrated. The high-signal deviations above the amplifier transition are genuinely measured, which limits the circularity, but the headline linearity statement in the Conclusion relies on this self-normalized curve.

full rationale

The paper's other laboratory results are self-contained and externally checkable: dark current is regressed against exposure time, the photon-transfer gain is derived from temporal variance versus mean and compared with the manufacturer value, blemish fraction is counted directly, and the on-sky photometric precision is a differential measurement reduced with standard astrometry and photometry packages. None of these reduce to the paper's own inputs by construction, and the self-citations to prior Mini-MegaTORTORA work are used only as a comparison to older Andor Neo data, not to justify the central Marana claims. The one partially circular element is the linearity curve: the caption states that the interval below 1000 ADU was used to define the linear slope against which measured signals are ratioed. The resulting ratio is therefore normalized to a fit of the same data, making the near-perfect low-signal linearity a self-definitional result rather than an absolute measurement. Because linearity is one of the two properties highlighted in the abstract and conclusion as making the camera promising, this normalization is load-bearing for that specific claim. The on-sky 1% precision does not rescue the linearity claim, since it is relative photometry of roughly constant stars and does not test response linearity over dynamic range. Score 4 reflects partial circularity confined to the linearity normalization while the rest of the evaluation retains independent content.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new entities. The central characterization relies on standard detector physics assumptions and on a self-normalized linearity definition. The free parameters are calibration constants, not physical parameters of the sensor.

free parameters (2)
  • Zero-point color term coefficient = 0.49
    Fitted to on-sky data as Instr = V - 0.49*(J-K); used for photometric calibration, not a property of the sensor.
  • Linearity reference slope = slope fitted over 0-1000 ADU interval
    Used to define 'expected' signal in the linearity curve; any low-signal nonlinearity is hidden by this normalization.
assumptions (4)
  • domain assumption Temporal variance of a pixel is dominated by Poisson shot noise plus read noise, so the slope of variance vs. mean gives the gain.
    Invoked in Section 3.2 to compute the photon transfer curve and gain.
  • domain assumption The sensor is perfectly linear in the interval below 1000 ADU.
    Used in Figure 6 to define expected signal; this is the least supported premise because it is the baseline for the linearity claim.
  • domain assumption The anti-glow correction subtracts a precomputed map proportional to exposure time.
    Inferred in Section 3.1 from the linear slope of over-compensated pixels; used to interpret dark current maps.
  • domain assumption A third-order spatial polynomial plus fixed color term models all spatial photometric systematics in the on-sky data.
    Used in Section 4; if unmodeled systematics remain, the 1% precision estimate would be optimistic.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Evaluation of scientific CMOS sensors for sky survey applications." pith.science (2026). https://pith.science/paper/VDVIZ52D

@misc{pith2026190900729,
  author       = {Pith},
  title        = {Pith review of: Evaluation of scientific CMOS sensors for sky survey applications},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VDVIZ52D}},
  note         = {Machine review of arXiv:1909.00729}
}
read the original abstract

Scientific CMOS image sensors are a modern alternative for a typical CCD detectors, as they offer both low read-out noise, large sensitive area, and high frame rates. All these makes them promising devices for a modern wide-field sky surveys. However, the peculiarities of CMOS technology have to be properly taken into account when analyzing the data. In order to characterize these, we performed an extensive laboratory testing of Andor Marana sCMOS camera. Here we report its results, especially on the temporal stability and linearity, and compare it to the previous versions of Andor sCMOS cameras. We also present the results of an on-sky testing of this sensor connected to a wide-field lens, and discuss its applications for an astronomical sky surveys.

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. phoptic -- a Python package for reducing astronomical images

    astro-ph.IM 2026-07 conditional novelty 6.0 of 10

    phoptic provides a single open-source Python pipeline that reduces images from OPTICAM, ULTRACAM, HiPERCAM, and MEXMAN into photometric light curves.

Reference graph

Works this paper leans on

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

  1. [1]

    APACrefauthors \ 2014 , Interpolative Blemish Corrections on sCMOS - How to switch ON/OFF

    neo_blemish APACrefauthors Andor, O I C. APACrefauthors \ 2014 , Interpolative Blemish Corrections on sCMOS - How to switch ON/OFF. Interpolative Blemish Corrections on sCMOS - How to switch ON/OFF. , available online at https://andor.oxinst.com/learning/view/article/interpolative-blemish-corrections-on-scmos-how-to-switch-on-off

    work page 2014

  2. [2]

    APACrefauthors \ 2019 , Marana sCMOS Specifications

    andor_marana APACrefauthors Andor, O I C. APACrefauthors \ 2019 , Marana sCMOS Specifications. Marana sCMOS Specifications. , published online at https://andor.oxinst.com/assets/uploads/products/andor/documents/Marana-sCMOS-Specifications.pdf

    work page 2019

  3. [3]

    APACrefauthors \ 2018 Nov , SEP: Source Extraction and Photometry

    SEP APACrefauthors Barbary , K. APACrefauthors \ 2018 Nov , SEP: Source Extraction and Photometry . SEP: Source Extraction and Photometry

    work page 2018

  4. [4]

    \ Arnouts , S

    sextractor APACrefauthors Bertin , E. \ Arnouts , S. APACrefauthors \ 1996 06 , 117 393-404 . APACrefDOI doi:10.1051/aas:1996164 APACrefDOI

    doi:10.1051/aas:1996164 1996

  5. [5]

    , Liu , C

    spie_neo APACrefauthors Fowler , B. , Liu , C. , Mims , S. \ et al. APACrefauthors \ 2010 Jan , A 5.5Mpixel 100 frames/sec wide dynamic range low noise CMOS image sensor for scientific applications A 5.5Mpixel 100 frames/sec wide dynamic range low noise CMOS image sensor for scientific applications . Sensors, Cameras, and Systems for Industrial/Scientific...

    work page doi:10.1117/12.846975 2010

  6. [6]

    APACrefauthors \ 2007 , Photon Transfer DN Photon Transfer DN

    ptbook APACrefauthors Janesick , J R. APACrefauthors \ 2007 , Photon Transfer DN Photon Transfer DN

    work page 2007

  7. [7]

    , Ebr , J

    fram_cta APACrefauthors Jane c ek , P. , Ebr , J. , Bla z ek , J. , Prouza , M. , Ma s ek , M. \ Eli \'a s ek , J. APACrefauthors \ 2017 01 , FRAM for the Cherenkov Telescope Array: an update FRAM for the Cherenkov Telescope Array: an update . Atmospheric Monitoring for High Energy Astroparticle Detectors (AtmoHEAD) 2016. Olomouc, Czech Republic, Edited b...

    work page arXiv 2017

  8. [8]

    APACrefauthors \ 2018 , FAST data acquisition software

    fast APACrefauthors Karpov, S. APACrefauthors \ 2018 , FAST data acquisition software. FAST data acquisition software. , published online at https://github.com/karpov-sv/fast . GitHub

    work page 2018

Show all 23 references
  1. [9]

    , Beskin , G

    karpov_2019 APACrefauthors Karpov , S. , Beskin , G. , Biryukov , A. \ et al. APACrefauthors \ 2019 Apr , Observations of Transient Events with Mini-MegaTORTORA Wide-Field Monitoring System with Sub-Second Temporal Resolution Observations of Transient Events with Mini-MegaTORT...

    2019

  2. [10]

    , Beskin , G

    karpov_2010 APACrefauthors Karpov , S. , Beskin , G. , Bondar , S. , Guarnieri , A. , Bartolini , C. , Greco , G. \ Piccioni , A. APACrefauthors \ 2010 , Advances in Astronomy 2010 784141 . APACrefDOI doi:10.1155/2010/784141 APACrefDOI

    2010 doi

  3. [11]

    \ Christov, A

    ccdlab APACrefauthors Karpov, S. \ Christov, A. APACrefauthors \ 2018 , CCDLab laboratory control software. CCDLab laboratory control software. , published online at https://github.com/karpov-sv/ccdlab . GitHub

    2018

  4. [12]

    , Katkova , E

    karpov_2016 APACrefauthors Karpov , S. , Katkova , E. , Beskin , G. \ et al. APACrefauthors \ 2016 Dec , Massive photometry of low-altitude artificial satellites on Mini-Mega-TORTORA Massive photometry of low-altitude artificial satellites on Mini-Mega-TORTORA . Revista Mexica...

    2016

  5. [13]

    , Orekhova , N

    karpov_meteors_2019 APACrefauthors Karpov , S. , Orekhova , N. , Beskin , G. \ et al. APACrefauthors \ 2019 Apr , Two-station Meteor Observations with Mini-MegaTORTORA and FAVOR Wide-Field Monitoring Systems Two-station Meteor Observations with Mini-MegaTORTORA and FAVOR Wide-...

    2019

  6. [14]

    , Jel \' nek , M

    rts2 APACrefauthors Kub \'a nek , P. , Jel \' nek , M. , Nekola , M. \ et al. APACrefauthors \ 2004 09 , RTS2 - Remote Telescope System, 2 ^ nd Version RTS2 - Remote Telescope System, 2 ^ nd Version . E. Fenimore \ M. Galassi \ ( ), Gamma-Ray Bursts: 30 Years of Discovery Gamm...

    2004 doi

  7. [15]

    , Hogg , D W

    astrometry.net APACrefauthors Lang , D. , Hogg , D W. , Mierle , K. , Blanton , M. \ Roweis , S. APACrefauthors \ 2010 May , 139 5 1782-1800 . APACrefDOI doi:10.1088/0004-6256/139/5/1782 APACrefDOI

    2010 doi

  8. [16]

    , Liu , Y

    spie_400 APACrefauthors Ma , C. , Liu , Y. , Li , J. , Zhou , Q. , Chang , Y. \ Wang , X. APACrefauthors \ 2015 Mar , A 4MP high-dynamic-range, low-noise CMOS image sensor A 4MP high-dynamic-range, low-noise CMOS image sensor . Image Sensors and Imaging Systems 2015 Image Sens...

    2015 doi

  9. [17]

    \ Depagne , \'E

    pickles APACrefauthors Pickles , A. \ Depagne , \'E . APACrefauthors \ 2010 12 , 122 1437 . APACrefDOI doi:10.1086/657947 APACrefDOI

    2010 doi

  10. [18]

    , Jel \' nek , M

    fram APACrefauthors Prouza , M. , Jel \' nek , M. , Kub \'a nek , P. , Ebr , J. , Tr \'a vn \' c ek , P. \ S m \' da , R. APACrefauthors \ 2010 , Advances in Astronomy 2010 849382 . APACrefDOI doi:10.1155/2010/849382 APACrefDOI

    2010 doi

  11. [19]

    , Mao , Y N

    qui_2013 APACrefauthors Qiu , P. , Mao , Y N. , Lu , X M. , Xiang , E. \ Jiang , X J. APACrefauthors \ 2013 May , Research in Astronomy and Astrophysics 13 5 615-628 . APACrefDOI doi:10.1088/1674-4527/13/5/012 APACrefDOI

    2013 doi

  12. [20]

    , Hinze , A

    schildknecht_2013 APACrefauthors Schildknecht , T. , Hinze , A. , Schlatter , P. , Silha , J. , Peltonen , J. , Santti , T. \ Flohrer , T. APACrefauthors \ 2013 Aug , Improved Space Object Observation Techniques Using CMOS Detectors Improved Space Object Observation Techniques...

    2013

  13. [21]

    , Fowler , B

    spie_scmos APACrefauthors Vu , P. , Fowler , B. , Liu , C. , Balicki , J. , Mims , S. , Do , H. \ Laxson , D. APACrefauthors \ 2008 Jul , Design of prototype scientific CMOS image sensors Design of prototype scientific CMOS image sensors . High Energy, Optical, and Infrared De...

    2008 doi

  14. [22]

    , Rao , P R

    noise_rts APACrefauthors Wang , X. , Rao , P R. , Mierop , A. \ Theuwissen , A J P. APACrefauthors \ 2006 Dec , Random Telegraph Signal in CMOS Image Sensor Pixels Random Telegraph Signal in CMOS Image Sensor Pixels . 2006 International Electron Devices Meeting 2006 Internatio...

    2006

  15. [23]

    write newline

    " write newline " cite write " FUNCTION editor.postfix editor num.names #1 > "( )" "( )" if FUNCTION editor.trans.postfix editor num.names #1 > "( )" "( )" if FUNCTION trans.postfix translator num.names #1 > "( )" "( )" if FUNCTION authors.editors.reflist.apa5 'field := 'dot :...

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.