Undersampling and oversampling

Oversampling and undersampling

When the photons hit the telescope aperture, diffraction of the light occurs and a star is imaged as a small diffraction disk.

KaiMartin, CC BY-SA 3.0, via Wikimedia Commons

In general, it is not intended to distribute the light of a star (or its diffraction disk) over too many pixels. Otherwise there would be too many gray levels and the star would look blurred, which would be intensified by bad seeing (air turbulence). Fainter stars no longer stand out from the background and are hardly recognizable on an image. In addition, with each additional pixel the noise component increases, worsening the signal-to-noise ratio (see section ‘Basics’ – ‘Exposure time and noise’) This whole effect is called oversampling.

The opposite phenomenon, undersampling, describes the case where a star is displayed on only one pixel. This is mainly the case with cameras with very large pixels, or with lenses with a very short focal length. The star looks square on the image and has no more brightness gradations.

The goal is to select a camera that matches the optical system, so that stars and objects are displayed as finely graduated as possible. The pixels must be matched to the image scale of the telescope, taking into account the air turbulence (seeing).