The Imatest program: comparing cameras with different amounts of sharpening

The Imatest program was developed to enable photographers and imaging system developers to conveniently measure the key image quality factors in cameras, lenses, and printers. For cameras and lenses these factors include sharpness (measured by MTF), noise, dynamic range, tonal response (OECF curve), color and exposure accuracy, lens distortion, light falloff (vignetting), and lateral chromatic aberration. For printers they include tonal response, Dmax (the maximum reproducible black density), color response, and color gamut. Although some measurements follow ISO standards, emphasis is on simple, convenient, and affordable measurements. We begin with an overview of Imatest, briefly describing each module. Then we focus on the issue of comparing cameras with different degrees of sharpening. Oversharpening, which is common in compact digital cameras, results "halos" near edges that make small prints look good but can be objectionable at large magnifications. It also results in exaggerated MTF measurements. Most digital single-lens reflex cameras (DSLRs) have little sharpening, putting them at a disadvantage in MTF comparisons. Imatest uses an algorithm called standardized sharpening that facilitates comparisons between cameras by adding or removing sharpening to make edge overshoot relatively consistent. The present algorithm adjusts the sharpening amount so that MTF at 0.3 times the Nyquist frequency is equal to MTF at low spatial frequencies. Determining the optimum sharpening radius R can be challenging because of the large variety of camera edge responses. We discuss considerations in selecting R and constraints on the sharpening amount that make it difficult to find a unique solution that fits all cameras-noisy compacts as well as low-noise digital SLRs.