Resampling

Pictures often appear "fuzzy" when scaled to a different size. The fuzziness of an image is controlled by two processes: resampling and filtering.

The filtering processes attempt to smooth (blur) the transformed pixels. When an image is being resampled to a different resolution, there's a need to interpolate the old pixelmap to a new one. This requires some form of image interpolation and a number of interpolation methods can be used. The simpler and faster method, is to use nearest-neighbour function, known as the quick filter. Provided the resolution change is not too great, this is normally perfectly adequate. However, for large upsampling operations, the pixels can appear as large blocks. Equally under large downsampling operations, a phenomenon known as anti-aliasing can occur that makes the image appear fuzzy. In such cases, a more accurate interpolation routine can be used but this requires more time for the image to be processed. The most commonly used filters are nearest neighbor, bilinear and B-spline, in order of accuracy. The full list of filters implemented in PhoA follows:

Nearest neighbor;
Linear (bilinear);
B-spline (bicubic);
Lanczos;
Mitchell.