Suggested Number of Calibration Files

Since bias, arc, and trace frames are all used in a way that consolidates much of the pixel data into a much smaller amount of information (e.g. a polynomial fit to the arc lines in a given order can convert 60,000 pixels into a handful of polynomial parameters), a single frame of each type is generally sufficient. Dark frames are not often used in data reduction and again, a single frame is deemed sufficient.

Flat fields, however, are generally applied on a one-to-one basis; each pixel in a science image is divided by the corresponding pixel in the flat field. In order to minimize the statistical noise introduced by the flat fields, normally at least 5 flat field images are averaged or medianed together.

Focus frames are usually not used for calibration purposes, so the UI will only return focus frames for a science observation if they are found on the same observation date.