Method and apparatus prepare an image for printing--and in some cases include an output stage, performing the actual printing--when the overall number of pixels is being scaled up or down from an original image or portion. Image-sharpening procedures are controlled automatically by a function of the number of pixels in the original image and the number in the output device. Any of a great variety of such functions may be used, but we prefer a step-function of a ratio of the two pixel counts--particularly a two-valued step-function in which one value is zero and the other serves as a sharpening factor in a sharpening algorithm. When the function yields its zero value, for efficiency the sharpening computations are preferably turned off. The nonzero value is automatically selected when the number of output pixels is more than twice the number of input pixels. The sharpening-factor value, ideally two, is multiplied by the deviation of each image element, e. g. pixel, from the average of its adjacent neighbors; this product is added to the level of the subject image element to form an adjusted level. Weighting or selection of neighbor values along particular axes allows for anisotropic resolution or anamorphic scaling.