Technical Knowledge

What are 8-bit, 10-bit, 12-bit, 4:4:4, 4:2:2 and 4:2:0

What are 8-bit, 10-bit, 12-bit, 4:4:4, 4:2:2 and 4:2:0

Jan 07 2020

When it comes to digital video production, we often see 8-bit, 10-bit or even 12-bit presenting the spec of image processing.  Sometimes you also find the numbers like 4:4:4, 4:2:2 and 4:2:0 on recording devices. What exactly do these numbers mean and how they affect the image quality and colors? We will answer all of your questions in this article. 

What is 8-bit, 10-bit and 12-bit color depth? 

Color depth is also known as bit-depth which refers to the number of bits used to define the color channels, red, green or blue, for each pixel. 

In most RGB systems, there are 256 shades per color channel. If you know binary system well enough, this number 256 should sound very familiar to you. The number, 256, is 2 raised to the 8th power or the 8-bit color depth. This means that each of the RGB channels has 256 shades so there are 256x256x256 or 16,777,216 colors in total in this 8-bit RGB system.

An 8-bit color system is capable of producing over 16 million colors. This may look humungous, but when it compared to 10 bit, this is actually nothing. In a 10-bit system, you can produce 1024 x 1024 x 1024 = 1,073,741,824 colors which is 64 times of the colors of the 8-bit. What is more shocking is that a 12-bit system is able to produce a whopping 4096 x 4096 x 4096 = 68,719,476,736 colors!  As a result, increasing the color depth will enable you to better represent your colors. 

What are chroma sampling and the numbers  4:4:4, 4:2:2 and 4:2:0?

We often see numbers 4:4:4, 4:2:2 and 4:2:0 on recording devices and these are known as chroma subsampling. Have you ever wondered how does chroma subsampling affects the colors of the image? And what exactly do these numbers  4:4:4, 4:2:2 and 4:2:0 mean? 

Before we dive into chroma subsampling, let’s first talk about the image pixel.  The image pixel is defined by luma and chrominance components. Without the chrominance components, the luma of each pixel produces a greyscale representation of the image.  Also, study indicates that human eyes are more sensitive to light or luminance than colors. 

YCbCr is a family of color spaces used as a part of the color image pipeline in video and digital photography systems.  Y refers to the brightness of the pixel and shares 1/3 the amount of signal.  The brightness signal is always retained without compressing.  Cb and Cr are the two chroma signals which share 2/3 amount of signal.  The chroma signals can be compressed for saving the amounts of data loading.

Let’s take 4:4:4 for the example.  The first 4 represents the number of pixels across we are subsampling.  The second 4 means 4 colors yield in the first row of chroma sampling, and the third 4, again, means 4 colors yield in the second row of chroma sampling.  Technically speaking, 4:4:4 represents each pixel has its’ own color value which includes all the chroma information, so it isn’t chroma subsampling.  Now let’s take a look for 4:2:2.  The second 2 means two chroma subsampling in the first row.  And the third 2 means two chroma subsampling in the second row, too.  Therefore, a 4:2:2 image only retains a half of the chroma samples that a 4:4:4 image does.  As to 4:2:0, it indicates two chroma subsampling in the first row, and no chroma subsampling in the second row, so the pixels in the second row copy the same chroma value of the first row.  As a result, a 4:2:0 image retains only a quarter of the color subsampling that a 4:4:4 image does.


Why is broadcast level video camera so powerful? 

Pixels are very tiny dots of color, so it is very hard to find out the noticeable visual difference whether the video is recorded in 4:4:4, 4:2:2 or 4:2:0.  However, 4:4:4 is able to record more color information than 4:2:2 and 4:2:0, thus the 4:4:4 chroma subsampling model still has advantages over 4:2:0 and 4:2:2 in terms of the color quality.

Most of the available DSLR and mirrorless cameras on the market use 4:2:0 chroma subsampling model to compress the video files.  Even though you can yield good image quality from the 4:2:0 video, you might still encounter problems when doing chromakeying or post-editing because of the low resolution for chroma information.  Compared to 4:4:4 images, it will be more difficult and time-consuming to achieve a clean chromakey result with 4:2:0 videos. This is why professional video producers still prefer working with 4:4:4 or 4:2:2 video, which contains more chroma information facilitating the post-edits, only the final video is compressed in 4:2:0 for saving the size of file.  This production procedure is like that a professional photographer always shoot photos with RAW files and then output the post-edits pictures in JPG format for the subsequent applications.

By knowing the theory of chroma subsampling, audience should by now know why only the professional broadcast level video equipment is able to present a very high-quality image, and why they are expensive than the consumer digital cameras and mobile phones.  Let’s take the example of Datavideo’s BC-100 Interchangeable Lens Video camera. The BC-100 is a broadcast level video camera designed for virtual studio. The camera is equipped with a 12-bit image processing sensor, capable of capturing a massive amount of color information and presenting the finest color differences.  The rich colors and sharp image quality are not only for visual enjoyment but also essential properties for attaining clean and pure objects from background by employing chromakeying.  With the advanced technique, you can easily chroma key the challenging objects such as a glass or hair, presenting the finest details combined with virtual background.  Additionally, the High Dynamic Range (HDR) technique allows video camera to record the details of the bright and dark parts of the image in high contrast conditions, making images more real as seen by the human eye.