Video files contain a LOT of data, so engineers are always looking for ways to compress the data, or reduce it, so that it becomes much more manageable without badly affecting image quality.
Chroma subsampling is one of the ways they do it. Video is captured using three channels – a luminance channel and two chroma channels. If this sounds like the Lab mode in image-editors, then you've got it. In video terms these are the Y (luminance) and Cb, Cr channels.
Human vision is much more sensitive to changes in resolution than colour, so as long as the Y channel stays at its full resolution, you can reduce the resolution of the Cb and Cr channels without badly affecting the image quality.
This is usually expressed as a ratio:
4:1:1 – low but often adequate quality. The Cb and Cr channels have on-quarter the data of the luminance channel.
4:2:2 – higher quality, and widely used. The Cb and Cr channels have half the resolution of the luminance channel.
4:4:4 – top quality with no chroma subsampling. However, the small increase in image quality may not be worth the large increase in data/file sizes.
There are other variations but this quickly gets complicated. It's one of those things where it's useful to know what it is when you read about it but, unless you get heavily into video, you probably don't need to do anything about it.
