COMPASS

English

한국어

English

한국어

Appearance

Dynamic Range Calculator ​

Calculate sensor dynamic range from full well capacity, read noise, dark current, and exposure settings. Compare single-exposure vs HDR modes across operating temperatures.

Dynamic Range Calculator

Calculate sensor dynamic range from FWC, read noise, dark current, and exposure time. Compare single vs HDR modes.

Single Exposure DR
66.0 dB
Single (stops)
11.0 EV
HDR DR
84.0 dB
HDR (stops)
14.0 EV
Noise Floor
5.03 e−
Dark Charge
0.3 e−
DR vs Temperature
-20°C-10°C0°C10°C20°C30°C40°C50°C60°C70°C405060708090100TemperatureDR (dB)SingleHDR

Model scope

Use this browser tool for intuition, relative trends, and design-space exploration. Its local simplified model is not a substitute for RCWA/FDTD sign-off, silicon calibration, or vendor process data.

Learn more

Noise, SNR, and Dynamic Range

Physics Notes

Dynamic Range and Saturation

Plain-English Intuition

Dynamic range is the gap between the brightest highlight a pixel can record before it saturates and the faintest detail it can pull out of the noise. Wide DR means you can see both bright clouds and dark shadows in the same shot. Lifting the ceiling (larger full-well) helps highlights; lowering the floor (less read noise, less dark current) helps shadows.

Dynamic range compares the largest usable signal to the smallest distinguishable signal, usually limited by full well and noise floor.

Assumptions

  • Dynamic range is computed from saturation capacity and minimum detectable signal under a compact noise model.
  • HDR extension is represented by simplified short/long exposure or knee behavior, not by a full sensor readout architecture.
  • Tone mapping and perceptual rendering are outside the physical DR estimate.

Outputs

  • Linear dynamic range, dB/stops conversion, noise floor, saturation limit, HDR extension proxy, and bottleneck diagnosis.
  • A comparison of whether FWC, read noise, dark noise, or HDR strategy is limiting usable range.

Validation Example

  • Increasing full well at fixed noise floor should increase DR by $20\log_{10}$ of the ratio.
  • Increasing read noise should reduce DR even when full well is unchanged.

Core Equations

Single-exposure DR
$$DR_{\text{dB}} = 20 \log_{10} \left( \frac{FWC}{\sigma_{\text{floor}}} \right)$$
  • \(DR\): Dynamic Range
  • \(FWC\): Full Well Capacity

The noise floor often includes read noise plus dark noise.

Saturation
$$S_{\text{clip}} = \min(S, FWC)$$
  • \(S\): Input photoelectron signal

No linear signal information remains above full well.

HDR exposure span
$$DR_{\text{HDR}} \approx 20 \log_{10} \left( \frac{S_{\text{max, long}}}{S_{\text{min, short}}} \right)$$
  • \(S_{\text{max, long}}\): Max signal in long frame
  • \(S_{\text{min, short}}\): Min signal in short frame

Multi-exposure HDR extends range but introduces motion and merge constraints.

Model Interpretation

  • Increasing FWC helps highlights; reducing read noise helps shadows.
  • Dark current raises the floor for long exposures and high temperatures.
  • Published DR values depend on measurement definitions and SNR thresholds.

Floor And Ceiling

  • The ceiling is set by full well, ADC range, or nonlinearity, whichever occurs first.
  • The floor is set by read noise, dark noise, quantization noise, and the chosen SNR threshold.
  • A DR number without a threshold definition is not directly comparable across cameras.

HDR Caveats

  • Multi-exposure HDR expands range but adds motion artifacts and merge nonlinearity.
  • Dual conversion gain changes the effective read-noise/full-well tradeoff across the response curve.
  • Highlight recovery is limited by what was not clipped in at least one exposure or gain state.

Known Missing Physics

  • This compact DR model does not include tone mapping, local contrast, flare, or black-sun artifacts.
  • It treats saturation as a hard clip, while real pixels can show blooming, compression, and color-channel clipping.
  • Scene dynamic range after ISP processing can differ from sensor-domain DR.

References