Photon Counting Statistics

Poisson statistics, SNR, dead time corrections, and count distributions.

Model: Exact

SNR per Interval √μ

10.0 (20.0 dB)

Relative Error 1/√μ

10.00%

Dead Time Loss

4.8%

recorded 0.952 of 1.000 Mcps

Corrected-Rate SNR in 1 s

975.9

√N, N = 9.524e+5 recorded

P(n) = μn·e−μ / n! [Poisson]; SNR = √μ (shot-noise limit)

m = R / (1 + R·τ) [non-paralyzable dead time]; corrected R̂ = m / (1 − m·τ)

Recorded counts are sub-Poissonian: Var N / N = 1/(1 + R·τ)² = 0.9070. The correction amplifies that smaller spread back, so the corrected rate has relative error 1/√N of the recorded counts.

Count Distribution

SNR & Error vs Counts

Model and references

Exact. The equations are exact within the stated assumptions, or are a standard's tabulated values. The results are as good as the inputs. About the tiers

Exact Poisson counting statistics for an ideal detector, with a non-paralyzable dead time. The corrected-rate SNR √N of the recorded counts holds for long gates, where the recorded counts have the renewal-process variance N/(1 + Rτ)².

  1. Knoll G. F. (2010). Radiation Detection and Measurement, 4th ed. Wiley. Ch. 4, section VII: non-paralyzable m = n/(1 + nτ) and paralyzable m = n·e^(−nτ) dead-time models.
  2. Müller J. W. (1973). Dead-time problems. Nucl. Instrum. Methods 112, 47–57. Dead-time models and the statistics of the recorded counts. doi:10.1016/0029-554X(73)90773-8
Report a problem

A wrong value, a formula, a reference, or a page that misbehaves. The report includes this page and its current inputs.