Matte

A matte functions as a visibility map for an image. White areas represent fully visible regions, black areas represent fully hidden regions, and gray values represent varying degrees of transparency. Mattes can exist as separate grayscale images or be embedded as alpha channels. The concept originates from traditional film compositing, where physical mattes were placed in front of camera lenses to control which parts of the frame were exposed.\n\nIn visual effects production, artists generate mattes for every element that needs to be composited. A scene with a CGI dinosaur walking through a real city requires separate mattes for the dinosaur body, its shadow, reflection in windows, and any dust or debris it kicks up. Each matte controls how its corresponding element blends with the live-action footage.\n\nMatte quality is measured by edge accuracy and the handling of semi-transparent regions. A high-quality portrait matte captures individual hair strands at partial opacity rather than cutting them off with hard edges. It also handles ears with their slight translucency, and correctly represents the transition between opaque clothing and the background. Low-quality mattes produce visible artifacts when the subject is placed on a new background.\n\nMagic Eraser generates high-quality mattes during background removal. The AI produces per-pixel opacity values that accurately represent hair strands, transparent fabrics, and soft shadow transitions. These mattes enable clean compositing when placing subjects on new backgrounds, whether solid colors for product listings or custom images for creative projects. The precision of the AI-generated matte eliminates the common post-compositing cleanup work that editors typically face, such as manually painting out halo artifacts, correcting edge color contamination, and refining areas where semi-transparent elements were incorrectly classified as either fully opaque or fully transparent.