What is 4 4 + 4

4: 4: 0, 4: 2: 0 and Co .: Color subsampling explained

The basics of color subsampling

The amount of data for videos must remain manageable in today's home cinema world so that a video can also fit on a Blu-ray or through the “narrow channels” of the Internet. This is where the codec engineers benefit from the physiology of the human eye. Because we react more sensitively to differences in brightness (luminance) than to colors (chrominance).

A black and white photograph looks razor sharp or washed out to us. We perceive the color information that is “overlaid” superficially - whether it coincides exactly with the edges or protrudes into other areas is of secondary importance.

Color subsampling means that brightness and color information are encoded separately from one another. In simple terms, this means that color information is superimposed on a grayscale image. Assigning a color to each brightness value is the most memory-intensive. Savings are possible if one color information serves two or four brightness values.

It's still quite abstract, but easy to explain. The relationship between the brightness information and the colors is indicated in the format A: B: C.

Incidentally, chroma subsampling does not describe the bit depth in which the colors are present. The individual color spaces are responsible for this.

The cryptic number combinations explained

What is behind terms like 4: 4: 4, 4: 2: 2 or 4: 2: 0? They indicate how lightness and color information is weighted over two rows and usually four columns. To do this, we split the digits into the letters already mentioned: A: B: C.

A. is usually numbered with 4 - a low power of two. Four pixels in the horizontal form a series of shades of gray, which we call the chroma series for the sake of simplicity.

Digit B. indicates how much color information is available for the named horizontal chroma series. A 4 means that each pixel is assigned a color value in addition to the brightness information. A 2 means that half as much color information as brightness information is fed in.

For digit C. let's create a second, lower chroma row. Why it is like that? It's just in the notation. This number C indicates how many colors are available to the respective chroma pixels in the second row. If C is specified as zero, the codec fills the bottom row with the color information from the top row.

That's how much good color subsampling saves

Combined, the video qualities can be read off using the A: B: C scheme. How much bandwidth do the different color subsampling save? You can calculate that yourself.

To do this, first add A, B and C. Divide the result by 12 (the highest possible quality) and multiply it by the value 100. The result in percent indicates the memory requirement.

The playout in 4: 4: 4 is the highest quality that is used in professional film production. The same notation is used in the RGB color space. Since this does not separate chrominance and luminance from one another, there is no color subsampling in this.

When playing films on BluRay or in the stream, 4: 2: 0 is the standardized compromise between image quality and data volume. A good half is saved, but the picture still looks sharp and colorful.

Modern televisions, BluRay players and monitors have a 4: 4: 4 scaling - this means that an AI routine enriches the material with additional color information after an analysis and plays it out. That works reasonably well. However, these routines also depend on the material at hand. So if you are a movie buff and want to watch films in the best possible quality, UHD BluRays with 10-bit color information (HDR) and 4: 2: 0 are the ideal entry point. A real 4: 4: 4 format, on the other hand, is not to be expected in the near future.

Comment now!

Don't miss a post

Select topics - fresh every day at 5 p.m. in your mailbox

View sample newsletter

For your feed reader

EURONICS on social networks

similar posts