Basic Video Compression Techniques in Multimedia

Video Compression Techniques:

Out of different multimedia elements, the need for compression is greatest for video as the data volume for fullscreen full-motion (FSFM) video is very high. The frame size for NTSC video is 640 pixels by 480 pixels and if we use 24 bits of colour depth then each frame occupies 640 x 480 x 3 bytes = 900 KB. So each second of NTSC video comprising 30 frames occupies 900 x 30 x KB which is around 26 MB and each minute occupies 26×60 = 1.9 GB.

Thus a 600 MB CD would contain a maximum of 22 seconds of FSFM video. Now, imagine the storage space required for a 2-hour movie. So, the only way to achieve digital motion video on a PC is to reduce or compress the redundant data in video files.

Spatial Redundancy:

In Digital video, Redundancy occurs when the same information is transmitted more than once. Primarily in an area of an image frame where the same colour or intensity spans more than the one-pixel location that is called Spatial Redundancy.

Spatial Compression:

This redundancy is removed by compressing each image frame in isolation. This techniques are generally used that says spatial compression or intra-frame compression. This compression applies different lossless and lossy methods same as those applied for still images. Some of these methods are:

i. Truncation of least significant image data
ii. Run Length Encoding (RLE)
iii. Interpolative Techniques
iv. Predictive Technique – DPCM (Differential Pulse Code Modulation), ADPCM (Adaptive DPCM)
v. Transform Coding Technique – DCT (Discrete Cosine Transform)
vi. Statistical or Entropy Coding – Huffman Coding, LZW coding, Arithmetic coding.

Temporal Redundancy:

When a scene is stationary or only slightly moving, there is redundancy between frames of the motion sequence. The contents of consecutive frames in time are similar or they may be treated by a simple translation function. This kind of redundancy says Temporal Redundancy.

Temporal Compression

It is removed by storing only the differences of subsequent frames instead of compressing each frame independently. This technique says temporal compression or inter-frame compression.