Various image quality characteristics are evaluated for digital imaging systems. Among them, the noise power spectrum (NPS) gauges granularity in the spatial frequency domain. It has been reported that digital NPS, measured using a direct-type FPD or photon-counting detector, is approximately flat with respect to frequency [1,2]. These findings contradict the established concept that digital NPS decreases in proportion to the square of the presampled modulation transfer function (MTF) [3,4].

 Several studies based on the NPS measurement results have been conducted to better understand the reason for the flatness of the NPS of the direct-type FPD [1,2,5,6]. However, no study has thoroughly investigated the phenomenon of NPS flattening. If the contradiction, regarding the NPS as a function of the spatial frequency, can be clarified, a detailed understanding of the noise factor becomes possible, which is expected to improve image quality. Furthermore, it useful for accurate derivation of the detective quantum efficiency (DQE) [7], which is a comprehensive measure of imaging system performance.

 Therefore, this study aims to clarify the factors affecting digital NPS using Monte Carlo simulation and experimental measurement, for more accurate image quality evaluation. In particular, we focused on sampling aperture and aliasing [8,9].