content
Zero, foreword one, digital video two, digital video format three, H.265 / hevc encoding video format
Zero, preface
This series of blogs make a simple introduction to H.265 / HEVC video encoding, mainly introducing approximately coding frameworks and processes, do not involve the detailed derivation and proof of algorithms and formulas, suitable for just contact video coding and H.265 / HEVC's novice player. My reference book is Wan Shuai, Yang Fuzheng edited "New Generation Efficient Video Code H.265 / HEVC: Principle, Standard and Implementation". The electronic version of this book is stored in my Baidu network disk, and everyone can take themselves. Link: https://pan.baidu.com/s/1zhczlci1wnb3pir9ehmq4q Baidu network disk - "h.265" electronic version code: PRBB
This blog is the second blog of the series of blogs, mainly to explain some basic knowledge of some digital video, such as representations, formats, etc., and H.265 / HEVC standards supported video format.
Links from Series Blog: 1, H.265 / HEVC Brief Report 1 Lesson (Introduction) 2, H.265 / HEVC Brief Report 2 (Digital Video Format) 3, H.265 / HEVC Brief Report Third Lesson 3 ( Encoding structure)
First, digital video
The video consists of many consecutive images arranged in time, each full image is referred to as one frame. Due to visual suspension mechanism, when the continuous play image exceeds 24 frames per second, the human eye cannot distinguish the single static screen, and the image sequence looks smooth and continuous visual effect, which is called video. Digital video is a video recorded in digital form. The digital video consists of a portable digital image. Each image consists of N rows, each row M pixels, each pixel being represented by digitized numerical values. Digital video has time resolution, spatial resolution, color space, quantization depth and other parameters, and a combination of these parameters is called video format. The unit resolution is the number of frames per second, ie the frame rate. The higher the frame rate, the better the video, but after the frame rate reaches a certain value, the human eye will not be able to see the difference, all of which do not have to pursue a particularly high frame rate. Spatial resolution refers to the pixel rows of the image and the number of pixels per row, the higher the spatial resolution, the clearer the image, the more space resolution of common HD video is 1920 * 1080. Color space, or color space, is in the form of pixel color, and its purpose is to illustrate some standard way to color, a color pixel is usually described by three components, often used in video compression, RGB, YUV, YCBCR color space. The quantization level corresponding to the value of each component of the pixel is quantified, the larger depth, the more accurate the pixel value, such as the traditional image, each color space component, with 8 bits to represent the pixel value, there is 256 grayscale levels. .
Second, digital video format
In the process of digital video acquisition, two different acquisition methods can be used: the first is to sample the brightness and chroma signal by the same spatial sampling frequency, the second is the spatial sampling frequency of the chroma signal. The spatial sampling frequency used for the brightness signal is low, and this sampling method is called chromaticity sub-sampling. The reason why chromatic sub-sampling can be used because the human eye is low-sensitive to the chroma signal, which can remove some of the chromaticity information in the image without being perceived by the person.
In YCBCR color space, Y represents the difference between the brightness, the CB represents the difference between the blue portion of the RGB input signal and the signal brightness value, and the CR represents the difference between the red portion of the RGB input signal and the signal luminance value. There are four sampling formats in this color space, namely 4: 4: 4, 4: 2: 2, 4: 1: 1, 4: 2: 0, and their position is shown below. Among them, the red circle represents the brightness sample Y, and the blue small triangle represents the color difference sample CB and Cr. The principle of the position shows the principle, only the YCBCR sampling format of 4: 2: 0 here. In this sampling format, the chrominance component is sampled in the horizontal and vertical directions, i.e., each 4 Y sample corresponds to a CB sample and a Cr sample. 0 means that both color difference samples are alternately performed when transmitting, that is, one of two color difference samples is transmitted during each row. In this format, the resolution of the chromaticity component in the horizontal and vertical direction is 1/2 of the luminance component.
Third, H.265 / HEVC encoding video format
The video source of different formats transformed into a unified data format through the pre-processing module. The codec need only a small number of unified video data formats, and the decoded video is re-transformed into a resume video that is consistent with the video source format. . The principle block diagram of video communication is shown below: We said that the input video of the encoder is encoded video, and its format information is referred to as encoding video format. The encoded video in different formats uses a more unified data structure. The decoder output video is referred to as a decoded video, and its format information is referred to as a decoded video format. Encoding video adopts a unified multi-level data structure: (1) encoding video is an image sequence arranged in order. (2) The image is one or more rectangular sampling matrices, each matrix corresponding to the brightness or chromaticity component. (3) The element of the pixel correspondence matrix, the element value is the value of the pixel, and the row of elements represents the spatial position of the pixel. The encoded video format mainly includes encoding image format, the encoded image format mainly includes matrix number, spatial relationship, image spatial resolution, pixel quantization depth, and the like. These syntax elements belong to the sequence parameter set (SPS). H.265 / hevc only specifies four types of encoded image formats, each type corresponding to the number of matrices and corresponding chrominance components. 1, only contain sampling matrices Y 2, a brightness component plus two chroma components (YCBCR or YCGCO) 3, three chroma sampling matrices (RGB) 4, other unspecified monochrome or three excitation color space sampling matrix decoded video The format is not only encoded video format, as well as scan type, image type, color space, etc. The scan type means that the image is a progressive scan or an interlaced scan. The progressive scanning method is to simultaneously display all the screens per frame, the interlaced method is to be divided into two screens, divided into the top field and the bottom field. . The image type refers to the spatial structure of the pixel in the image. If the image is a row of frames or the top field of the interlaced scan. The sampling aspect is a ratio of the pixel width and height, and the spatial resolution of the inclined image can obtain a width ratio of the image display. A scanning refers to a portion of the pixels near the edge of the display input image in the display. The video format is the code of the video. Pixel value characteristics refer to the color in the pixel, which is commonly represented by brightness and chrominance.
Our other product:
