Acquisition Time Gating

• Acquisition time gating is a temporal resource-allocation technique that restricts detector activation to specific intervals, improving dynamic range and signal fidelity.
• It employs methods like external optical gating, phase-selective sorting, retrospective self-gating, and adaptive gate placement to optimize limited measurement budgets.
• This approach enhances performance in various applications such as photon-counting OTDR, SPAD-based LiDAR, MRI, and RF or X-ray measurement systems by mitigating saturation and reducing interference.

Acquisition time gating denotes a family of measurement strategies in which temporal support is made an explicit design variable: a detector, receiver, or reconstruction pipeline is activated only during selected time intervals, or continuously acquired data are retrospectively assigned to temporally meaningful windows. Across the literature, the term spans external optical gating before a photon counter, bias-current gating of superconducting or avalanche detectors, symbol-synchronous receiver activation, on-pulse/off-pulse differencing in interferometry, paired pumped/reference counters in synchrotron experiments, retrospective windows in radar and RF scattering, self-gating from MRI calibration data, and adaptive gate placement driven by posterior inference or Fisher information (Li et al., 2019, Huang et al., 2021, Rosenzweig et al., 2018, Sumaya-Martinez et al., 1 Jan 2026).

1. Conceptual scope

A useful way to organize the literature is to distinguish whether the gate acts on the physical detector, on the incoming signal before detection, on the assignment of already acquired samples to temporal classes, or on an adaptive policy that changes during acquisition.

Mode	Operational meaning	Representative examples
Physical or signal-path gating	Detector sensitivity or incident flux is restricted to a chosen window	PC-OTDR with an MZI optical switch; gated SNSPD biasing; TPC gating grid (Li et al., 2019, Hummel et al., 2022, Tangwancharoen et al., 2016)
Differential or phase-selective gating	Samples are sorted into on/off, pumped/unpumped, or pulse-phase bins	MSP on-off imaging; EIGER2 Counter A/Counter B (Roy et al., 2013, Naumenko et al., 2023)
Retrospective or self-gating	Data are acquired continuously and sorted afterward using timing or motion surrogates	SSA-FARY in radial MRI; UWB and bistatic RCS post-processing windows (Rosenzweig et al., 2018, Gharamohammadi et al., 2019, Azizi et al., 2021)
Adaptive gating	Gate position or width is updated from prior detections or optimized from a statistical criterion	Thompson-sampled SPAD LiDAR; FI-driven FLIM (Po et al., 2021, Sumaya-Martinez et al., 1 Jan 2026)

This range of usage matters because “gating” is not synonymous with a hardware shutter. In some systems it is a real-time bias or optical window; in others it is a post-acquisition decision rule or a retrospective motion binning procedure. The common structure is temporal selectivity under a constrained acquisition budget.

2. Detector-side and signal-path implementations

In photon-counting OTDR, acquisition time gating is implemented as an external optical gate rather than internal detector gating. A Mach-Zehnder interferometer intensity modulator is inserted in the return path so that the single-photon detector sees only one selected return-time segment at a time. The paper’s short-fiber model gives a dynamic-range improvement

$$\Delta D_R =10\log\!\left(\frac{\gamma N_b}{\gamma m+N_b-m}\right),$$

which approaches

$$\Delta D_R \approx 10\log\!\left(\frac{N_b}{m}\right)$$

for large extinction ratio. In a 70 m system operated in the 850 nm wavelength band with 50 ns external gates, 14 sequentially delayed 5 m segments, and the same 420 s total measurement time as the ungated reference, the reconstructed trace reached 30.0 dB dynamic range instead of 19.0 dB, and a 0.37 dB micro-bend loss event became visible. The same work explicitly distinguishes this from internal detector gating: the external optical gate reduces the actual optical flux incident on the detector outside the desired interval, whereas internal gated mode was shown not to remove saturation and produced false rising-edge spikes (Li et al., 2019).

A closely related detector-side realization appears in superconducting nanowire single-photon detectors, where the gate is created by pulsing the detector bias current. Cryogenic control circuitry mounted at 2.3 K and coupled to an SNSPD at 0.8 K achieved a minimum off-to-on rise time of $2.4\pm0.1\,\mathrm{ns}$, a minimum total gate length of $5.0\,\mathrm{ns}$, and electrically tunable windows up to $500\,\mathrm{ns}$ at $1.0\,\mathrm{MHz}$. The current-loading time constant was modeled as

$\tau_{\mathrm{rm} = \frac{ \mathrm{L}\left( \mathrm{R_{s}\left(2\mathrm{R_{L1}+R_{L2}\right)+\mathrm{R_{p}\left(R_{L1}+R_{L2}\right)} \right)} {\mathrm{R_p R_s}\left(2\mathrm{R_{L1}+R_{L2}\right)},$

making detector inductance and cryogenic bias impedance the dominant speed constraints. Gated operation was used both for dynamic-range extension by a factor of $11.2\pm0.1$ and for temporal filtering of a bright pulse by $31\pm2\,\mathrm{dB}$ in a pump-probe emulation (Hummel et al., 2022).

In gaseous tracking detectors, the same concept appears as a literal electrostatic shutter. The TPC gating grid is open when all wires are at the average bias $V_a$, and closed when neighboring wires alternate between $$\Delta D_R \approx 10\log\!\left(\frac{N_b}{m}\right)$$0 and $$\Delta D_R \approx 10\log\!\left(\frac{N_b}{m}\right)$$1. The acquisition window is therefore the interval during which drift electrons are allowed into the amplification region. The reported driver opened the installed SπRIT TPC grid in $$\Delta D_R \approx 10\log\!\left(\frac{N_b}{m}\right)$$2 and restored 99% of the original charges within $$\Delta D_R \approx 10\log\!\left(\frac{N_b}{m}\right)$$3; the dead region under the grid followed $$\Delta D_R \approx 10\log\!\left(\frac{N_b}{m}\right)$$4 (Tangwancharoen et al., 2016).

In optical wireless communication, a symbol-synchronous SPAD gate creates a repeated per-symbol acquisition window. The detector is ON for $$\Delta D_R \approx 10\log\!\left(\frac{N_b}{m}\right)$$5 and OFF for $$\Delta D_R \approx 10\log\!\left(\frac{N_b}{m}\right)$$6, with the gate-ON time optimized for BER rather than for raw count rate. The crucial regime distinction is whether $$\Delta D_R \approx 10\log\!\left(\frac{N_b}{m}\right)$$7 exceeds the dead time $$\Delta D_R \approx 10\log\!\left(\frac{N_b}{m}\right)$$8: if $$\Delta D_R \approx 10\log\!\left(\frac{N_b}{m}\right)$$9, a detection in one symbol cannot block the next symbol’s gate-ON interval; if $2.4\pm0.1\,\mathrm{ns}$0, residual intersymbol interference remains. The paper shows that an optimal $2.4\pm0.1\,\mathrm{ns}$1 exists because signal collection, dead-time carryover, and background rejection compete (Huang et al., 2021).

3. Differential, phase-selective, and correlation-based gating

In radio interferometry of millisecond pulsars, acquisition time gating is not an amplitude shutter but a phase-selective folding operation performed after coherent dedispersion and before imaging. The correlator first dedisperses antenna voltages, then folds high-time-resolution visibilities with a topocentric rotational model that can include $2.4\pm0.1\,\mathrm{ns}$2, $2.4\pm0.1\,\mathrm{ns}$3, and $2.4\pm0.1\,\mathrm{ns}$4, and finally forms on-pulse minus off-pulse gated visibilities. This was necessary because, for example, at $2.4\pm0.1\,\mathrm{ns}$5 across $2.4\pm0.1\,\mathrm{ns}$6 bandwidth and $2.4\pm0.1\,\mathrm{ns}$7, the dispersion delay is about $2.4\pm0.1\,\mathrm{ns}$8, much longer than a typical MSP period. The gated correlator localized five newly discovered Fermi MSPs to approximately $2.4\pm0.1\,\mathrm{ns}$9, enabled an approximately $5.0\,\mathrm{ns}$0 reduction in GMRT follow-up timing telescope time, and used 10 to 15 phase gates across the sample (Roy et al., 2013).

A different differential construction appears in time-resolved X-ray diffraction with the EIGER2 hybrid photon counting detector. In double-gating mode, two interleaved counters acquire alternating pump-synchronized and unpumped reference signals from the isolated single bunch in the Elettra hybrid filling mode. The detector was triggered at $5.0\,\mathrm{ns}$1, the laser at $5.0\,\mathrm{ns}$2, so every second selected X-ray pulse coincided with laser excitation. Counter A recorded the pumped image and Counter B the unpumped reference shifted by a half period in the storage-ring time frame. This architecture yielded an effective temporal resolution of $5.0\,\mathrm{ns}$3 FWHM and strongly suppressed source drift through combinations such as $5.0\,\mathrm{ns}$4 and the normalized differential signal $5.0\,\mathrm{ns}$5 (Naumenko et al., 2023).

All-optical sampling via four-wave mixing provides another form of acquisition gate. In telecom nanolaser characterization, a synchronized $5.0\,\mathrm{ns}$6 gate pulse of approximately $5.0\,\mathrm{ns}$7 was mixed with the nanolaser emission in $5.0\,\mathrm{ns}$8 of dispersion-shifted fiber, and the idler at $5.0\,\mathrm{ns}$9 was counted as a function of delay. The effective temporal resolution was about $500\,\mathrm{ns}$0–$500\,\mathrm{ns}$1. Because the counter averaged over one million pulses, the measured trace reflected both intrinsic pulse shape and pulse-to-pulse timing jitter; Langevin modeling showed timing jitter near threshold dropping from about $500\,\mathrm{ns}$2 at zero power to about $500\,\mathrm{ns}$3 above threshold (Monti et al., 2 Feb 2026).

4. Retrospective, post-acquisition, and self-gating

In several systems, “gating” is applied after the raw data have already been acquired. For shallow-buried-object UWB imaging, the main problem is that the clutter-removal window cannot be chosen accurately when target depth is unknown. The proposed solution first uses Wiener filtering with the average trace as clutter reference, then uses the Average Similarity Function to pre-detect likely target positions, and only then sets the time-gating boundary at the midpoint between the first clutter peak and the second peak. In the reported experiment, this combined Wiener plus adaptive time-gating pipeline detected all buried objects, whereas one-step time gating nearly suppressed one Mine D target and missed the rock at $500\,\mathrm{ns}$4 (Gharamohammadi et al., 2019).

A closely related RF-scattering workflow appears in bistatic RCS measurement. Complex $500\,\mathrm{ns}$5 sweeps are transformed to time domain, a high-order Kaiser window is centered on the desired target return around 27–29 ns, and the result is transformed back to frequency domain for comparison between the DUT and an equal-size ground-plane reference. Unwanted responses around 6–10 ns were associated with sidelobe-mediated or environmental returns, and the time gate suppressed background contributions by more than 80 dB. At 18 GHz and 32 GHz, the difference between simulation and measurement was reported as less than 1.5 dB at $500\,\mathrm{ns}$6 (Azizi et al., 2021).

MRI self-gating generalizes the idea further: no explicit acquisition window is imposed during the scan. Instead, the data are acquired continuously under free breathing, and motion surrogates are extracted afterward from central k-space auto-calibration data. SSA-FARY performs PCA/SVD on a zero-padded time-delay embedded Block-Hankel matrix, yielding EOF quadrature pairs that encode respiratory and cardiac motion phase. The method recommends a bandwidth $500\,\mathrm{ns}$7, found a window of about 3 s robust in practice, and used 25 cardiac bins and 9 respiratory bins for retrospective sorting prior to PICS reconstruction. In this usage, gating means retrospective binning of continuous acquisition rather than intermittent detector activation (Rosenzweig et al., 2018).

5. Adaptive and information-theoretic gating

The adaptive-gating literature turns temporal support into a sequential decision variable. In SPAD-based 3D imaging under strong ambient light, the detector records only the first photon after activation, so pile-up makes early background photons especially harmful. The proposed adaptive policy maintains a posterior over depth, samples a hypothesis via Thompson sampling, and places the next gate at that sampled depth. With adaptive exposure, acquisition terminates when the expected posterior error falls below a prescribed threshold. Outdoors under direct sunlight, this produced about $500\,\mathrm{ns}$8 RMSE reduction at fixed acquisition time and about $500\,\mathrm{ns}$9 exposure-time reduction relative to free-running mode; adding a flatness prior yielded a 70% decrease in exposure time and a 60% decrease in RMSE, while a monocular depth prior gave 45% lower total acquisition time (Po et al., 2021).

In FLIM, programmable time gating is cast directly as an optimal experimental design problem. For gate $1.0\,\mathrm{MHz}$0 spanning $1.0\,\mathrm{MHz}$1, the expected count is

$1.0\,\mathrm{MHz}$2

and the Poisson Fisher information is computed over the gate set. Nuisance robustness is handled by the Schur-complement effective matrix

$1.0\,\mathrm{MHz}$3

with D-optimal utility

$1.0\,\mathrm{MHz}$4

The practical workflow is a short scout acquisition, estimation of $1.0\,\mathrm{MHz}$5, evaluation of a hardware-feasible candidate library, and adaptive selection of gate start times and widths. The typical programmable-gating settings reported were a 0–12.5 ns time window, 4–16 gates, a 1–2 frame scout budget, and optimized use of the remaining frames (Sumaya-Martinez et al., 1 Jan 2026).

This adaptive perspective also clarifies an implicit form of gating in dead-time-limited ranging. In high-flux TCSPC LiDAR, the detection-time distribution converges to the stationary distribution of a Markov chain rather than to the arrival-time distribution, and the paper notes that when $1.0\,\mathrm{MHz}$6 is only slightly smaller than $1.0\,\mathrm{MHz}$7, the dead time can act as a signal-triggered gate. That interpretation suggests that gating need not always be explicit hardware control; it can also emerge from detector dynamics and be exploited statistically (Rapp et al., 2018).

6. Acquisition-time consequences, tradeoffs, and terminology

A recurrent misconception is that gating necessarily shortens measurement duration. The PC-OTDR study is explicit that the gated comparison used the same 420 s total measurement time as the ungated reference; gating improved dynamic range by redistributing the detector’s finite count budget across serially measured segments rather than by increasing the detector’s maximum usable counts (Li et al., 2019). A related misconception is that all gating is equivalent. External optical gating in PC-OTDR is not equivalent to internal SPD gating, post-acquisition windows in UWB and bistatic RCS are not front-end hardware shutters, and MRI self-gating extracts motion surrogates from continuously acquired data rather than interrupting acquisition (Li et al., 2019, Gharamohammadi et al., 2019, Rosenzweig et al., 2018).

The dominant tradeoffs are application-specific but structurally similar. Narrower windows or more segments typically improve selectivity, but they also raise power requirements, magnify pile-up or threshold sensitivity, or increase the number of sequential acquisitions. In symbol-gated SPAD communications, the optimal $1.0\,\mathrm{MHz}$8 balances signal collection, dead-time recovery, ISI suppression, and background rejection (Huang et al., 2021). In MRI self-gating, finer retrospective binning improves motion resolution but leaves each bin more undersampled and increases dependence on compressed-sensing regularization (Rosenzweig et al., 2018). In adaptive SPAD depth sensing, shorter effective exposure is obtained not by faster hardware but by making each cycle more informative (Po et al., 2021).

The term also appears in broader acquisition-time analyses where the “gate” is the observation slot of a burst or scan. In turbo-coded offset-QPSK burst acquisition for geosynchronous satellite channels, the receiver uses up-sampling, matched filtering, and a feedforward approach so that clock recovery acquisition time is equal to the preamble length (Vasudevan, 2015). In free-space optical acquisition with photon-counting arrays, one full scan lasts approximately $1.0\,\mathrm{MHz}$9, and the acquisition-time upper bound is written as

$\tau_{\mathrm{rm} = \frac{ \mathrm{L}\left( \mathrm{R_{s}\left(2\mathrm{R_{L1}+R_{L2}\right)+\mathrm{R_{p}\left(R_{L1}+R_{L2}\right)} \right)} {\mathrm{R_p R_s}\left(2\mathrm{R_{L1}+R_{L2}\right)},$0

linking one-shot thresholded counting performance to repeated full scans. That study concludes that an array of smaller detectors gives better acquisition performance in terms of acquisition time than one large detector of similar dimensions (Bashir et al., 2019).

Taken together, the literature suggests that acquisition time gating is best understood as a general temporal resource-allocation principle. It can mean restricting when a detector is sensitive, which samples are compared, which phase bins accumulate data, which continuous measurements are retrospectively retained, or which gate pattern maximizes information under a fixed photon or dwell budget. The shared function is temporal selectivity in service of sensitivity, robustness, or acquisition efficiency rather than time-windowing as an end in itself.