The difference between image---frame and field in multimedia development

---------------------------AUTHOR:PKF

----------------------------------------time:2015-08-20

------------------------------------------------------------qq:1327706646

First, what is the field?

Each TV frame is created by scanning the screen two times, and the second scanned line fills the gap left by the first scan. Each scan is called a farm. So 25 frames per second of the TV screen is actually 50 games/second (if NTSC is a & 60-because I am Chinese, so I use the value of PAL).

To get the frame with the most detail, you need to merge the information in two fields. So far, these ideas are simple, but it's not that easy to take into account the dynamic picture. Because the camera will scan the two fields sequentially, all the files in the image are moved to a different location than the first field in the second field. This will help the TV screen move more smoothly, but it is also a problem in editing.

Second, computer and television

When a computer plays a video on a monitor, it displays only a complete set of frames without using the TV technique of the interlaced field. Therefore, the video format and MPEG-1 designed for the computer monitor do not use the field. Video editing software designed for these formats will not be able to handle the farm correctly, so it is not suitable for producing video for television. If your movie will be played on TV, be sure to use the editing program that handles the farm.

Iii. Overview

The original video frame (the most original video data) is scanned in different ways according to the coding needs to produce two video frames: continuous or interlaced video frames, interlaced video frames including the top and bottom fields, continuous (and then) scanned video frames and interlaced video frames have different characteristics and coding characteristics, resulting in so-called frame coding and field coding. In general, the frame encoding of the line frame, interlaced frames can be selected between the frame encoding and field coding.

In frame coding, referring to frame image, using frame motion compensation, two fields are combined coding, and in field coding, the reference is the fields image, two fields are coded separately, and the field motion compensation is used.

Second, video sequence frame, field encoding method

1. Fixed frame encoding (full frame)----All frames of a video sequence are always frame encoded.

2. Fixed field code (full-court)

The frames in the video sequence are divided into two fields independently encoded. Encoding Rules:

1. I-frames can be encoded into two I-field or an I-field and a P-field, i.e. II, IP.

2. P-Frames can be encoded into two P-fields or a P-field and a B-field, i.e. PP, PB.

3. B-Frame can be encoded into two B-field, namely BB.

3. Image-level frame, field-adaptive coding (PAFF)

The video sequence can be encoded into a single frame or two fields, and the adaptive selection principle is based on the RD of each frame using that encoding

Value.

4. Macro block-level frame, field-adaptive (Mbaff)

In order to improve the coding efficiency, the macro block-level frame field Adaptive is adopted. The macro block level uses the macro block pair (MBP) as the basic Coding Unit (2)

The Mbaff solution adopted by H.:

In the image of the larger movement than the field code, the movement is smaller than the location of the use of frame coding, encoding sequence examples are as follows:

The difference of Paff and Mbaff in video coding collection
There are two fields per frame of interlaced images, so there are three ways to encode interlaced images:

1. Combine two fields into one frame to encode

2, two fields are coded separately

3. Merge two fields into one frame, but at the macro block level, divide a frame macro block into two field macro blocks for encoding.

The above two coding methods are called image adaptive frame/field coding (Paff), and the third is called macro block adaptive frame/field coding (Mbaff).

For a moving image, because there is a large scanning interval between the fields, the spatial correlation between two adjacent rows in the frame is smaller than the progressive scan image, so coding the two fields separately will save the code stream. For a non-moving image, there is a large spatial correlation between two adjacent rows, and it is more efficient to combine two fields into one frame encoding. Therefore, the Paff encoding method can be adapted to the whole image encoding, choosing frame coding or field coding.

When the image has both motion and non-moving regions, the disadvantage of Paff appears, Paff adaptive granularity is too coarse, can not achieve more sophisticated coding requirements, so mbaff came into being. Mbaff the selection of frame encoding or field encoding is based on macro block, Mbaff two fields into one frame to encode, but each frame macro block (16x16) is divided into a field macro block pair (8*16), for each frame macro block, compare the frame encoding and field encoding generated by the size of the stream, using the most efficient way to save the code stream.

Video sequence is generally divided into interlaced scanning and progressive scanning two ways, interlaced scanning frame is divided into the upper and lower structure, and video display technology is usually shown in the field as the basic unit for compatibility of interlaced and progressive scanning two ways.

There is a progressive_sequence tag in the video sequence extension to distinguish the structure sequence of the video frame.
When progressive_sequence=1, indicates that the following video is a progressive frame graph
Progressive_sequence=0, indicating that the following video can be a field graph, or a frame graph, which can be progressive or interlaced.

There are 3 fields in each image extension that are related to the display time of an image.
Top_field_first for each
Repeat_first_field
Progressive_frame

The frame here consists of 2 fields, such as 25 fps, which is actually 50 games per second.
if (progressive_sequence==1)
{
The 2 fields of the output frame are progressive.
The number of output times of the reconstructed frames is =top_field_first+repeat_first_field+1, and the maximum number of repeat outputs is 3
}
Else
{
if (progressive_frame==1)
{
The 2 fields of the output frame are progressive.
if (repeat_first_field=1)
{
Output 3 fields, one of which is repeat output
}
Else
{
Output 2 field
}
}
Else
{
2 fields of the output frame are interlaced
}
}

http://wenku.baidu.com/link?url=g8sxpd16m22rkkjus0toucdlxox2uw1lvcq_- Qlhhc4ivhie6gruniwkf8ojm7nyk9op9dxpgplaqge1jkd9yezon9vapqxzjimj8g9iprw

Http://blog.sina.com.cn/s/blog_498164df01010ew9.html

Http://www.cnblogs.com/yinxiangpei/articles/2824561.html

http://blog.csdn.net/maopig/article/details/6862077

Http://www.znczz.com/thread-208977-1-1.html

http://blog.csdn.net/maopig/article/details/6863253 Multicast

The difference between image---frame and field in multimedia development