The Constant Battle: Signal vs. Noise
Every photograph of the night sky is a fight between two elements: signal vs. noise. The 'signal' is the good stuff — the faint, precious photons that have traveled millions of years from a nebula or galaxy to reach the camera's sensor. The 'noise' is everything
else that gets in the way. This includes random electronic interference from the camera's sensor (thermal noise), unwanted light from cities (light pollution), and even the inherent graininess of digital images. In a single, long exposure, this noise can easily overwhelm the faint signal, resulting in a grainy, washed-out image where the cosmic object is barely visible. The entire goal of modern astrophotography is to improve this 'signal-to-noise ratio' (SNR), making the celestial signal stronger and the random noise weaker.
Gathering the Light
This is where stacking begins. Instead of taking one extremely long photo, a photographer takes many shorter exposures of the same object, often hundreds of them. These individual pictures are called 'light frames'. Any single light frame might look noisy and unimpressive. But crucially, the signal — the stars, the galaxy — is in the same place in every frame, while the electronic noise is random and different in each shot. When these frames are combined using a stacking algorithm, the magic happens. The software precisely aligns the stars in every image. Then, it averages the pixel values. Because the signal is constant, it gets reinforced and becomes clearer. Because the noise is random, it gets averaged out and fades into the background. The result is a dramatic improvement in the signal-to-noise ratio. Stacking just four frames can double your SNR, while stacking 100 frames can improve it tenfold.
The Art of Calibration
However, simply stacking light frames isn't enough to create a flawless image. Professional and serious amateur photographers use 'calibration frames' to remove predictable, non-random noise from the camera and optical system. There are three main types. 'Dark frames' are taken with the lens cap on, using the same exposure time and temperature as the light frames. This captures the camera's thermal noise pattern, which can then be subtracted from the main images. 'Bias frames' are extremely short exposures also taken with the cap on, which record the electronic 'read noise' inherent in the sensor. Finally, 'flat frames' are photos of an evenly lit, uniform surface. These are used to correct for optical imperfections like dust spots on the sensor or vignetting, where the corners of the image are darker than the center. Using these calibration frames is not optional for high-quality results; they clean the canvas before the final painting.
The Digital Darkroom Does the Math
The stacking algorithm, found in software like DeepSkyStacker, Siril, or PixInsight, is the digital darkroom where all these elements come together. The process is computationally intensive. First, the software subtracts the master dark and bias frames from every light frame to remove sensor noise. Then, it divides out the flat frames to correct for optical flaws. Once each light frame is 'calibrated', the software analyzes the stars to perfectly align every image, even correcting for tiny shifts or rotations between shots. Finally, it integrates all the calibrated and aligned light frames. Using statistical methods like averaging or median clipping, it combines them into a single, master file with an incredibly high signal-to-noise ratio, ready for final editing.
