Chroma Subsampling

Compressed Digital video is used, from recording/shotting to broadcasting/streming. The codec’s jungle is a real source of science and can generate difficult workflow. But sometimes they can be destructive and there will be data loss we wouldn’t pick up. One of the method of data loss is the chromatic sub-sampling. Let’s see how it works. 

Il est connu que la couleur des images numériques est composé de 3 couches : R, V et B
It is known that the color of digital images is composed of 3 layers . Red, Green and Blue

Par une simple conversion on passe en Y(Luminance), U et V (Chrominances)
A simple conversion goes from RVB in Y (luminance), U (Chrominance 1) and V (Chrominance 2). 4:4:4 – all the layers have the image resolution

With 4 :4 :4 there is no sub-sampling… it happens!


4:2:2 - la résolution horizontale des couches de chrominance est divisée par 2
4 :2 :2 Sub-sampling – the horizontal resolution of the chrominance layers is divided by 2

With 4:2:2 we reduced to 1/3 of the quantity of information to store. Naked eye cannot see/notice it. 


4:2:0 - la résolution des couches de chrominance est divisée par 4
4:2:0 Sub-sampling – the resolution of the chrominance layers is divided by 4

With 4:2:0 we reduced to 1/2 of the quantity of information to store. Naked eye cannot see it. 


Little recep to understand what is going on with the first 8 pixels of an image. 
Cette page vous a plu ?