To confirm the validity of the MTF calculated from the edge image obtained by the constructed simulation, the sinc function and the MTF were compared as shown in figure 6.

 It was possible to examine what factors cause the quantum mottle blurring in the detector, and the value for NPS_digital(u), by conducting the simulation with clear conditions concerning the quantum mottle and unsharpness component, without other noise factors. As a result, it was found that even if the aperture changes, NPS_digital(u) is a constant value, which corresponds to NPS_{quantum mottle}(u) on the detector surface (Figure 7). This result suggested that NPS_digital(u) is not affected by the aperture. Furthermore, the effect of aliasing was also confirmed by calculation, using the theoretical value as shown in figure 8. The theoretical NPS value above the Nyquist frequency was folded back to the low spatial frequency region and added. This was compared with NPS_{quantum mottle}(u). The NPS value accounting for aliasing is lower than NPS_{quantum mottle}(u) in the high spatial frequency region, and the shape also greatly differs. In contrast, the simulated NPS_digital(u) corresponds well with the values and shape of NPS_{quantum mottle}(u) across all spatial frequencies as shown in figure 7.

 In addition, the digital NPS was measured using an indirect-type FPD (figure 9), and the effects of the MTF components were verified. When the analog component was separated from the measured presampled MTF(u) (figure 10), and the effect of MTF_analog(u) was corrected using the correction factor as shown in figure 11 obtained by Eq.1, the NPS value became flat with respect to frequency and showed a roughly constant value as shown in figure 12. On the other hand, when the measured NPS was corrected with the presampled MTF(u), including the unsharpness component due to the aperture, the NPS increased with increasing spatial frequency. The difference between the quantum mottle and the NPS also increased in the high spatial frequency region (figure 12). Thus, from the experimental measurements, it was estimated that the effect of the  aperture has not been included in the NPS. The corrected NPS value was about 2×10^-6 mm² higher than NPS_{quantum mottle}(u) (figure 13). This is considered to be an accurate representation of the electrical and structural noise in the detector and is a reasonable value from the results of previous studies [6,7].

 Here, we consider sampling with a reasonable aperture size for incident x-ray photons, uniformly exposed. Assuming that a set of random numerical values ​​in the aperture were measured the analog signal which is white noise becomes digitized random noise (also white noise). If there are no factors affecting the frequency domain other than the aperture, and given that this random signal is not correlated, when the autocorrelation function is calculated it becomes a delta function. Additionally, the Fourier transform of the delta function becomes a constant (with respect to frequency), so the power spectrum of the random signal also becomes a constant. Therefore, when a detector without the unsharpness factor of an analog component (that is, there is no correlation in information between adjacent pixels) as in our simulation, is uniformly exposed to radiation, the obtained power spectrum is constant with respect to frequency without depending on the aperture size. Furthermore, if the number of incident photons per unit area is the same, NPS(u=0) is constant regardless of the aperture size. Therefore, we believe that our simulation results demonstrate high reliability. On the other hand, in a detector with unsharpness due to the analog component, such as an indirect-type FPD, a correlation occurs between adjacent pixels. This leads to a change in frequency dependency according to the correlation that occurs in NPS_digital(u).