With data rate as high as 165 mbps, the analog video sequence is surely not going to fulfill our digital video requirements in its native forms. A two hour Hollywood movie transport stream (TS) or a pro-res file would cost you as much as 60 GB of data if watched online in the same format. There is a need to reduce the media overheads in order to distribute the videos. Video compression is important but at the same time, there is a trade-off for quality given that some data is going to be lost.
We know that videos are made up of frames and the video compression techniques are also of two types:
- Frame level – Here the compression is done within the frame. Color nuances, resolution and small invisible parts of picture are compressed
- Video/Motion – Here the adjacent frames are checked and the information which is repetitive or unchanged between frames is optimized
To understand motion compression, let us take three frames from a video sequence as displayed in Image 1
The frame to the left is the first frame in the sequence followed by the frame in the middle and then the frame to the right. When displayed, the video sequence shows a man walking from right to left with a tree that stands still. This is shown in Image 2
In all of the compression techniques that you want to follow, there is going to be a key frame on which the motion compression is not applied while the later frames will carry data from the key frame and just add up changes that have happened between frames.
In the above example video, only the new parts of the video sequence are included together with information of the moving parts. So, the second and the third frame will have information of the walking man only while the compression algorithm will take information of the tree from the first frame while displaying the video. Hence the second and third frames will look like Image 3 and Image 4 respectively.
The basis of all of the video compression techniques is the understanding of how human eyes and brain work together to decipher movement in a sequence. However, there is a price to pay for any compression. There is degradation of the images and addition of so called ‘artifacts’. So, the compression techniques or transcoding has to be closely monitored to recognize the trade-off between video-quality and file size.