Luma is the brightness component of a pixel in YCbCr — usually written with a small prime mark as Y′ to distinguish it from true linear luminance. While colour is split into Cb (blue difference) and Cr (red difference), luma alone tells you how bright each point in the picture should be. If you display only the luma of a video, you get a high-quality black-and-white version of the picture.
Why luma matters: it's where almost all of the perceptual information in video lives. Human vision is much more sensitive to brightness than to colour — that's the whole reason video codecs split the signal this way in the first place. A picture's structure, edges, faces, motion, depth cues are virtually all encoded in luma, while chroma just adds the colour overlay. Practically, this lets codecs keep luma at full resolution while halving or quartering chroma resolution (chroma-subsampling) with no visible loss.
For product teams, luma is mostly invisible plumbing but worth knowing for two reasons. First, quality issues that show up "in the brightness channel" — banding, compression blocking, noise — affect luma directly, and luma at full resolution means those problems are unmaskable. So encoder choices that hurt luma quality are immediately visible, while choices that hurt chroma quality often aren't. Second, modern codec features like perceptual quantization spend most of their attention on luma — protecting faces, edges and detail areas — because that's where viewers actually look.
