{"id":"992f5061-c462-4b46-bf59-c11dd64a3671","arxiv_id":"1909.00729","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"low","formal_verification":"none","parameter_count":2,"one_line_summary":"Laboratory and on-sky characterization shows the Andor Marana sCMOS camera has low blemished-pixel fraction, good linearity up to half dynamic range, and 1% photometric precision, making it suitable for sky surveys.","lead":"This paper tests a new scientific CMOS camera (Andor Marana) in the lab and on the sky, measuring its dark current, noise, linearity, and photometric stability. It finds the camera is stable and accurate to about 1% in on-sky measurements, making it a strong candidate for wide-field sky surveys.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Linearity claim is self-normalized to the 0-1000 ADU fit in Figure 6, so smoothing nonlinearity at low signal is untested; the on-sky 1% precision does not independently verify linearity.","rationale":"The reader identified the self-normalized linearity measurement as the weakest assumption; I agree. The Figure 6 caption explicitly states that the interval below 1000 ADU is used to define the linear slope, so the ratio is identically 1 there. This is a genuine circular step, not merely an omitted error bar. The concern is load-bearing because the abstract and conclusion advertise 'nearly perfect linearity' as a key result supporting the camera's suitability for sky surveys. The on-sky 1% precision, while encouraging, is a differential measurement of constant stars and does not constrain linearity. That said, the central 'promising detector' conclusion is not fatally undermined: the sensor shows low read noise, no horizontal correlation, a small blemish fraction, and no sub-pixel systematics above 0.5%, all of which are independent positives. The linearity issue can be remedied by an absolute calibration, so a conditional verdict is appropriate. I would therefore keep the reader's CONDITIONAL verdict and recommend running the proposed absolute linearity test before relying on the linearity claim.","tokens_in":9414,"tokens_out":10775,"duration_ms":126613,"concrete_test":"Repeat the linearity measurement using the integrating-sphere monitor photodiode (or a two-lamp superposition method) as an independent intensity reference, instead of defining the expected signal from a self-fit over 0-1000 ADU. Fit signal versus reference flux over the full range and report residuals below 1000 ADU; if the residual nonlinearity exceeds about 1%, the 'nearly perfect linearity' conclusion is not supported. As a second check, plot the PTC-derived gain versus signal below 1500 ADU and test the slope for a statistically significant trend.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's headline lab result is that Marana linearity is 'nearly perfect up to approximately half of saturation level'. However, Figure 6 defines the 'expected' signal as the best-fit straight line over the 0-1000 ADU interval, so the Measured/Expected ratio is forced to 1 in that interval by construction. Any smooth nonlinearity below 1000 ADU is therefore invisible in that plot, making the 'nearly perfect linearity' claim partially tautological. The photon-transfer gain curve in Figure 4 is an independent check: a constant local gain below the 1500 ADU transition would support linearity. But the text does not quantitatively report the flatness of the gain below 1500 ADU, only that it matches the manufacturer's value; a mild low-signal nonlinearity could hide in the pixel-to-pixel scatter. This is load-bearing because linearity is one of the two properties highlighted in the abstract and conclusion as making the camera promising for sky surveys, and because photometric calibration over a wide dynamic range is essential for transient and survey science. The on-sky 1% photometric precision is a differential measurement of roughly constant stars and does not test linearity, so it does not rescue the linearity claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","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%.","tokens_in":9652,"tokens_out":5404,"duration_ms":58689,"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":[{"comment":"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.","section":"§3.2, Fig. 6"},{"comment":"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.","section":"§3.2, Figs. 4 and 6"},{"comment":"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.","section":"§4, Fig. 13"}],"minor_comments":[{"comment":"The abstract contains grammatical errors: 'a typical CCD detectors' and 'All these makes them promising' should be corrected to standard English.","section":"Abstract"},{"comment":"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.","section":"§3.3 and Fig. 6 caption"},{"comment":"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.","section":"Fig. 13"},{"comment":"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.","section":"After Conclusions"},{"comment":"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.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The camera used in the study was provided by the manufacturer, Andor, but the conflict-of-interest statement declares no potential conflicts. I recommend asking the authors to clarify the nature of the manufacturer's support and whether any conditions were attached; this does not change my technical assessment."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the useful part: this is the first published characterization of the Andor Marana (GSense400BSI) sensor, and it is a solid, honest piece of detector work. The dark current maps, blemish fraction (0.02% vs 0.66% for Neo), lack of column-level noise correlation, and the amplifier transition around 1500 ADU are all useful data points. The on-sky 1% photometric precision with critically sampled PSFs is a genuinely encouraging result, and the authors are upfront about the limitations, including the Anti-Glow over-compensation and the need for careful flat-fielding across the transition.\n\nThe main soft spot is the linearity claim. Figure 6 defines 'expected' signal as the best-fit line over 0-1000 ADU, so the Measured/Expected ratio is forced to 1 in that interval. Any smooth nonlinearity below 1000 ADU is invisible. The paper says linearity is 'nearly perfect up to approximately half of saturation,' but that is relative to the sensor's own low-signal slope. The PTC gain curve is an independent check, but the text only says the gain matches the manufacturer's value; it does not quantify flatness below the transition. So the stress-test note lands. It's not fatal — the gain transition is clearly visible and the linearity deviations above half-scale are real — but the abstract's emphasis on linearity is slightly overstretched.\n\nOther soft spots: several figures lack per-point error bars, and no raw data or scripts are provided, so others cannot verify the reductions or compare easily. The on-sky 1% is from a single setup and doesn't test linearity. These are proportionate, not fatal.\n\nWhat holds up: the methods are standard (PTC, dark current regression, on-sky calibration with a color term), the comparison to Neo is useful, and the conclusions mostly follow from the data. The paper is not circular overall; gain, dark current, blemish fraction, and on-sky precision are independently measured.\n\nWho it's for: anyone selecting or using sCMOS cameras for survey or time-domain work. The first data on this sensor is worth having, even with the caveats.\n\nRecommendation: send it to peer review. A good referee can ask for a better linearity plot (e.g., using the gain curve or an external reference) and error bars, and the paper will be stronger for it. It deserves referee time, not a desk reject.","headline":"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.","tokens_in":10182,"tokens_out":2400,"would_cite":true,"duration_ms":23372,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"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.","keywords":["scientific CMOS","sCMOS","Andor Marana","detector characterization","photometric precision","linearity","dark current","random telegraph signal"],"falsifier":"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.","tokens_in":9210,"feed_emoji":"🔭","tokens_out":6884,"duration_ms":55968,"temperature":0.7,"pith_summary":"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.","feed_headline":"sCMOS camera hits 1 percent photometry for sky surveys","feed_subtitle":"Lab and on-sky tests of Andor Marana find high stability, near-linear response, and few cosmetic defects.","key_machinery":"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.","core_discovery":"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.","pith_inferences":["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."],"forward_implications":["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."],"supporting_citations":[{"why":"Supplies the photon transfer technique used to derive gain and locate the amplifier transition.","marker":"Janesick (2007)"},{"why":"Manufacturer specifications for the Marana camera that the laboratory measurements are compared against.","marker":"(Andor, 2019)"},{"why":"Characterization of the earlier Andor Neo sCMOS that serves as the baseline for noise, linearity, and stability comparisons.","marker":"(Karpov, Beskin, et al., 2019)"},{"why":"Documents the horizontal striping in Andor Neo data that the Marana is shown to be free of.","marker":"(Schildknecht et al., 2013)"},{"why":"Provides the astrometric calibration used to align and reduce the on-sky frames.","marker":"(Lang, Hogg, Mierle, Blanton, & Roweis, 2010)"},{"why":"Supplies the synthetic photometric catalog used to model the color term and zero point of the unfiltered photometry.","marker":"(Pickles & Depagne, 2010)"},{"why":"Supplies the source extraction and photometry routines used to measure stars in the on-sky frames.","marker":"(Barbary, 2018)"},{"why":"Underlies the SExtractor algorithms on which the on-sky source measurement is based.","marker":"(Bertin & Arnouts, 1996)"}],"fun_headline_variants":["Marana sCMOS hits 1% photometry, 30x fewer defects","sCMOS Marana: 1% sky photometry with stable, quiet pixels","Andor Marana sCMOS: high stability for high-cadence surveys","Marana sCMOS: near-linear response boosts survey accuracy"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"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.","fun_headline_variants_meta":{"raw":{"variants":["Marana sCMOS hits 1% photometry, 30x fewer defects","sCMOS Marana: 1% sky photometry with stable, quiet pixels","Andor Marana sCMOS: high stability for high-cadence surveys","Marana sCMOS: near-linear response boosts survey accuracy"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000313,"raw_usage":{"total_tokens":1729,"prompt_tokens":845,"completion_tokens":884,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":461,"completion_tokens_details":{"reasoning_tokens":799}},"tokens_in":461,"tokens_out":884,"duration_ms":7343,"temperature":1.0,"reasoning_tokens":799,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T05:37:38.753097+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"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.","supporting_citations":[{"cited_title":"APACrefauthors \\ 2007 , Photon Transfer DN Photon Transfer DN","cited_arxiv_id":null,"evidence_quote":"Supplies the photon transfer technique used to derive gain and locate the amplifier transition."},{"cited_title":"APACrefauthors \\ 2019 , Marana sCMOS Specifications","cited_arxiv_id":null,"evidence_quote":"Manufacturer specifications for the Marana camera that the laboratory measurements are compared against."},{"cited_title":", Hinze , A","cited_arxiv_id":null,"evidence_quote":"Documents the horizontal striping in Andor Neo data that the Marana is shown to be free of."},{"cited_title":"APACrefauthors \\ 2018 Nov , SEP: Source Extraction and Photometry","cited_arxiv_id":null,"evidence_quote":"Supplies the source extraction and photometry routines used to measure stars in the on-sky frames."}],"review_version":1}