Home » Why is 4K video or picture clearer than 2K on 1080p screen?

Why is 4K video or picture clearer than 2K on 1080p screen?

  • by

Before we talk about chroma sampling, let’s make clear two things, RGB and YUV

RGB color is easy to understand, the three channels represent red, green and blue respectively; The 8-bit binary number of each channel can express 256 colors, and the 24 bit binary number of three channels can express 16777216 colors in total. Usually, we use six hexadecimal numbers to express RGB. For example, the following blue is ᦇ 1453ad. You can use this color to draw a Greek flag. After all, they don’t really stipulate what kind of blue the flag must use.

YUV comes from an engineering problem, that is, how to make color TV signal compatible with black and white TV. The advantage of this is to ensure that the users who buy color TV can watch color TV programs normally, and at the same time, the black-and-white TV users can watch black-and-white pictures normally when they receive color signals. Extracting brightness information from color signal involves a problem of conversion.

The conversion formula used by NTSC is as follows:

Please note that these three formulas actually imply that YCbCr is not exactly the same as YUV sometimes, but we don’t study them first and think they are similar for the time being.

In fact, by fine-tuning the above coefficients, different conversion formulas can be obtained, which are suitable for different systems. Of course, we can use vector and matrix to express, for example, HDTV broadcasting in North America:

Let’s take out the YUV three channels separately to see what they look like.

YUV comes from an engineering problem, that is, how to make color TV signal compatible with black and white TV. The advantage of this is to ensure that the users who buy color TV can watch color TV programs normally, and at the same time, the black-and-white TV users can watch black-and-white pictures normally when they receive color signals. Extracting brightness information from color signal involves a problem of conversion.

The conversion formula used by NTSC is as follows:

Please note that these three formulas actually imply that YCbCr is not exactly the same as YUV sometimes, but we don’t study them first and think they are similar for the time being.

In fact, by fine-tuning the above coefficients, different conversion formulas can be obtained, which are suitable for different systems. Of course, we can use vector and matrix to express, for example, HDTV broadcasting in North America:

Let’s take out the YUV three channels separately to see what they look like. After converting RGB to YUV, the amount of information will not change, but there is an obvious advantage that the color information and brightness information are separated. According to the research, we can compress the color signal without significantly reducing the quality of the picture.

This figure is difficult to understand, so I will just give a brief introduction, which is only for qualitative understanding.

J: A: B, where j is usually taken as 4, representing 4 columns with 2 pixels in each column, so the figure above shows 8 pixels;

a. The first 2 in 4:2:2 means that there are two samples in the first horizontal row;

b. Represents the number of chroma information samples in the second row relative to the first row. The second 2 in 4:2:2 means that the pixels in the second row have two more samples than those in the first row.

Or more simply, 4:4:4 means that each pixel has chroma information; 4: 2:2 means that two horizontal pixels share chroma information; 4: 2:0 represents that four pixels of 2×2 share chroma information.

Because it can save a lot of bandwidth (storage capacity), many video programs are actually encoded with 4:2:2 or even 4:2:0. For example, Blu ray disc uses 4:2:0 chroma sampling. For a 1080p video, its chroma component is only 540p. However, you won’t feel any problem. In fact, its equivalent data per pixel has changed from 24 bits to 12 bits (both chroma channels are a quarter of the previous one), which compresses half of the space without causing a huge loss of image quality, If we don’t continue to use H.264 and other compression algorithms, the bit rate per second is still as high as 12 bits by 1920 by 1080 by 24fps, which is about 570mbps. A 120 minute movie will need about 500GB of space.

When you use 1920×1080 screen to watch 3840×2160 video, the input video data stream includes a 2160p brightness signal and two 1080p chroma signals. Because your screen resolution can’t keep up with the video input resolution, it’s supersampling. What game players like to say is that you turn on double full screen anti aliasing (FSAA’s algorithm is to render with twice or even four times the screen resolution, and then zoom back to the original resolution for output; The resolution is a multiplication relation. If you double the full screen anti aliasing, you have to render four times as many pixels as the original resolution, which is a great pressure on the GPU). You also get the complete 1080p absolutely uncompressed chroma information. In fact, with the increase of pixel density per inch of screen, such as doubling from 1920 * 1080 to 3840 * 2160, the image quality loss of chroma sampling will become less obvious.

Leave a Reply

Your email address will not be published. Required fields are marked *