As was noted, in the early days, x-ray film was one of the primary detection methods for x-rays. Geiger-Müller detectors were also used. These days, most but not all x-ray detectors count individual photons as they hit the detector. There is often some mechanism to filter out events that occur with too little energy (likely noise) or too much energy (cosmic rays) so at least some spurious signals can be removed. Some detectors, such as CCD imaging devices, measure the deposited energy, and do not measure individual photons. It is important to know the difference. In the case of photon counting, one measures discrete events, but one wants to extrapolate that to a rate that photons are arriving on average. The more photons that are counted, the more accurate that rate is known. The uncertainty in this rate is known as Poisson noise or shot noise. and follow rules that are commonly called counting statistics. For larger numbers of counts, say 20, the uncertainty in the rate will be the square root of the number of counts. While this becomes inaccurate for smaller numbers, this is usually ignored. With non-photon counting detectors, the uncertainties in the measurement must be determined from the detector properties. This is not always simple, but is important as the optimum results from fitting are only obtained when observations are weighted properly with respect to their experimental uncertainties.