Controlling the aliasing by zero-padding in the digital calculation of the scalar diffraction.

The well-sampling conditions for the digital calculations of the scalar diffraction, including the Huygens convolution method (HCM), the angular spectrum method (ASM), and the Fresnel diffraction integral (FDI), were discussed. We found the aliasing always occurs unless proper zero-padding--that is, to pad zero-value pixels around the initial field--is applied prior to the simulation of the diffraction. From the aspect of well-sampling, the ASM is applicable to a short propagation distance, while the HCM is applicable to a long propagation distance. However, we found that the free-space point spread function in the HCM is low-pass filtered when the propagation distance is long. As a result, it is recommended to always use the ASM in conjunction with sufficient zero-padding for the digital calculation of the diffraction field. The FDI can be directly applied to a long-distance propagation without the necessity of the zero-padding, provided only the intensity is of interest. If the phase of the diffraction field is important, the zero-padding is necessary and the propagation distance is severely restricted.