Astrophotography Quiz: Darks, Flats, Stacking & Tracking

A dark frame is shot with the lens capped at the same exposure time, ISO, and temperature as the real frames. It records only the sensor's own thermal signal and hot pixels, which can then be subtracted from the light frames. Vignetting and dust are handled by flat frames instead.

A flat frame is a picture of an evenly lit, featureless surface taken with the same optics, focus, and filters as the main images. Any darkening it shows comes from vignetting or dust shadows, so dividing the images by it evens them out. Color casts and sharpness are not what flats correct.

A bias frame uses the shortest possible exposure with the lens capped, so it records only the fixed offset and pattern that the sensor's readout electronics add to every image. Subtracting it removes that baseline from the other frames. It has nothing to do with white balance or motion.

Dithering moves the camera or telescope a few pixels between exposures. The stars then land on different pixels each time while hot pixels and fixed sensor patterns stay put, so stacking software can reject those patterns as outliers. Without dithering, fixed pattern noise stacks up into streaks or a grid.

An imprecise polar alignment means the mount's rotation axis does not match Earth's rotation axis, causing stars to drift out of position over time. The mount's mechanical quality determines whether it can hold an accurate track between guiding corrections.

Atmospheric turbulence causes air cells of varying temperature and density to move across the line of sight, bending light rays irregularly. Stars appear to shimmer and blur, and fine planetary detail is smeared, limiting the resolution that any optical system can achieve.

Autofocus usually struggles on point-like stars at night. A Bahtinov mask turns a bright star into a spike pattern whose middle spike centers only at perfect focus, and magnified live view lets the star be shrunk to its smallest point by hand. Stopping down to f/22 only hides focus error while adding diffraction.

If the angular size of an Airy disk is smaller than a single pixel, the sensor is under-sampling the optical image. Stars appear as single bright pixels with squared edges rather than smooth round profiles, and fine structure in the diffraction pattern is lost.

Random noise averages toward zero when frames are combined, while the consistent signal accumulates. Combining N frames improves the signal-to-noise ratio by the square root of N: doubling frames improves SNR by about 41 percent, not 100 percent.

Doubling the number of stacked frames improves SNR by the square root of 2, which is approximately 1.41. Quadrupling frames doubles SNR. The returns diminish progressively, so very large stacks are mainly useful when individual exposures are limited by sky glow or tracking.

Light pollution floods the sky with scattered light, raising the background brightness of every frame. This compresses the dynamic range available between faint celestial objects and the noise floor, reducing the contrast of dim targets relative to the background.

Narrowband filters pass only a tiny slice of the spectrum around emission lines, such as hydrogen alpha or oxygen III, the light that nebulae emit. Streetlight glow and most skyglow fall outside those slices and are blocked, so faint nebulae can be imaged even from light polluted places.

Trailing in an otherwise tracked shot indicates that the mount's rotation axis is not aligned with the celestial pole, causing the stars to drift across the field between guiding corrections. Errors in the guiding algorithm or poor mount mechanics can cause similar drift.

Dark frames must be taken at the same sensor temperature, ISO, and exposure duration as the light frames to accurately model the thermal noise pattern. Flat frames must match the optical path including lens, aperture, and filters used during capture.

Quality astrophotography results from applying each step in sequence: calibration frames remove sensor artifacts, accurate tracking minimizes trailing, stacking reduces noise, and careful processing extracts detail from the stacked result without introducing artifacts.