Chroma is the colour part of a pixel, separated from brightness. In the YCbCr colour format that every modern video codec uses internally, each pixel is described by three numbers: Y is brightness (luma), Cb is "how blue is it compared to grey", Cr is "how red is it compared to grey". Together Cb and Cr reconstruct the full colour; alone they tell you nothing about brightness.
Why the split exists is the cornerstone of how video compression actually works. Human vision is dramatically more sensitive to brightness changes than to colour changes — you can pick out a tiny shadow but you struggle to distinguish two similar greens. By separating chroma from luma, codecs can store chroma at lower resolution (chroma-subsampling: 4:2:2, 4:2:0) without you noticing. That single decision typically halves the raw data of a video before any further compression even kicks in.
For product teams, chroma is an invisible mechanism behind a few visible decisions. Chroma subsampling (whether your pipeline uses 4:4:4, 4:2:2 or 4:2:0) directly affects bandwidth and quality on colour-heavy content. Chroma keying (greenscreen compositing) requires enough chroma resolution to work cleanly — 4:2:0 is usually unsuitable, 4:2:2 is the production minimum. Chroma artefacts like "blocking" in red areas or coloured edges around text are signs of either over-compression of chroma channels or wrong subsampling for the content type. None of these need direct chroma manipulation; they're consequences of higher-level decisions.
