Smallest unit of a digital image — determines resolution and file size. Only matters when discussing compression artifacts or sensor limitations.
In digital image production, you ultimately always work with pixels—but it's only when resolution, compression, or visible errors come into play that you become aware of them. A pixel is the smallest addressable unit of a raster image, a tiny color value at a defined position. In 4K DCP (4096 × 2160), you have nearly 8.8 million pixels to manage per frame—at 24fps, this means massive amounts of data that must be stored, transported, and played back.
In practice, you primarily notice pixel problems in three situations: compression artifacts arise when you use overly aggressive codecs (H.265, ProRes) and details are quantized away—then you see blocky structures or color flickering in dark areas. Aliasing appears as a jagged effect on diagonal lines or fine patterns because the pixel resolution is too low to render smooth curves. And banding occurs when color gradients are encoded with too little bit depth—then you see visible bands instead of smooth transitions.
On set or in the edit suite, the pixel density of a display is relevant: a 55" monitor at 4K viewed from 3 meters appears pixelated because your eye can still resolve the individual dots. That's why colorists sit closer or use smaller, high-density reference monitors. For VFX work in DaVinci or Nuke, compositors zoom to a 100% view to work pixel-perfect—especially for keying or fine masking.
A practical tip: If color noise becomes visible in dark scenes, it's often due to 8-bit color depth (256 values per channel). With 10-bit or 12-bit, you have more intermediate steps and smoother gradients. Log formats (Alexa LogC, REDlogfilm) use pixel information more efficiently—more dynamic range with less bit expenditure. For the final export, pixel dimension compatibility is crucial: Full HD requires true downsampling, not just scaling, otherwise you'll see aliasing.