Green screen is a production technique in which subjects are filmed or photographed against a uniformly coloured backdrop ( typically a vivid, saturated green ) which is then removed in post-production and replaced with any desired background image or video. The process of digitally isolating and removing the coloured background is called chroma keying or keying, because it works by targeting a specific hue, or chroma value, rather than a luminance range. The resulting composite places the subject seamlessly into an environment that would be impractical, costly, or impossible to film in directly.
The technique works by identifying all pixels in the frame that fall within a defined range of the key colour and making them transparent. Software tools such as DaVinci Resolve, Adobe Premiere Pro, and After Effects offer dedicated keying effects that analyse the colour and refine the matte ( the mask that separates subject from background ) using controls for spill suppression, edge softening, and foreground colour correction. Spill refers to the green light that reflects off the backdrop and contaminates the edges of the subject, appearing as a faint green fringe that must be corrected to achieve a convincing composite.
Green is the preferred backdrop colour in most modern productions because digital camera sensors are more sensitive to green wavelengths, yielding a cleaner signal with less noise and more information for the keying algorithm to work with. Blue screen is used in contexts where the subject is wearing green clothing, or in certain film productions where the blue channel of analogue film offered the cleanest separation. The choice of colour is always dictated by the palette of the subject being filmed.
In AI video and image workflows, green screen remains highly relevant. AI-generated backgrounds can be composited behind live-action subjects filmed on green screen, allowing creators to place real people into fantastical, generated environments. Conversely, AI background replacement tools have emerged that can automatically segment subjects from footage without requiring a physical green screen, using learned semantic understanding of the scene to produce a clean matte: though these tools perform most reliably when the subject is well separated from a simple background.
