Chroma Subsampling

This reduces the amount of information and therefore the storage space or bandwidth needed to record or transmit a television signal.  Because our eyes are less sensitive to colour differences than to luminance (black & white) differences, the loss of colour information caused by chroma subsampling goes largely un-noticed.

The sensors on a television camera interpret a scene as 3 colours: Red, Green & Blue.  If these are converted directly to three full quality channels we would call this 4:4:4 sampling. The number ‘4’ denotes that full information is stored for very pixel.

This RGB signal is converted to luminance (the black & white version of the image) known as the Y signal and 2 colour difference signals: R-Y and B-Y. These Y, R-Y, B-Y signals hold all the information necessary to display the image.

Each pixel is digitally sampled to derive its Y, R-Y and B-Y value and convert it into digital data. This can be done in the ratio of 4:4:4 which means that the luminance Y, and colour difference signals R-Y, B-Y are all sampled equally. This is "lossless" sampling as all the colour information is retained in the digital signal and is found in high end formats such as HDCAM.

Chroma Subsampling  can be applied to reduce the amount of data and therefore the size of the signal and consequent file size needed. Sampling at a ratio of 4:2:2 retains all of the luminance Y information by sampling each and every pixel for its Y value (this is most important to render an accurate image), but only half of the colour information by sampling every other pixel for the colour difference value. 4:2:2 is the "Gold Standard" for broadcast television and is classed as "virtually lossless"

Other ratios are 4:1:1 & 4:2:0 (1 in 4 pixels sampled for colour),  4:0:0 & 3:1:1