VIDEO COMPRESSION MPEG BY ttchen 2003413 Outline VIDEO

VIDEO COMPRESSION & MPEG BY ttchen 2003/4/13

Outline • VIDEO COMPRESSION – VIDEO CODING BASICS – MOTION COMPENSATION • MPEG FAMILY – – – MPEG-1 VIDEO standard MPEG-2 VIDEO standard MPEG-4 H. 264 MPEG-7

VIDEO COMPRESSION • Why do we use compression? y – Temporal Redundancy – Spatial Redundancy x time

Two Stage Coding Process Stage 1 Stage 2

Reducing Temporal Redundancy • Segment a frame into macroblocks, not search by pixels • Output energy( bit rate) is increased with the degree of temporal redundancy • Interframe coder ( only use temporal redundancy reduction)

Reducing Spatial Redundancy • Using DCT coding • Intraframe coder (use only spatial redundancy reduction, only for current frame) • Hybrid( intra-/ inter - frame) coding method

Motion Compensation • Often used and developed by MPEG family • 將相鄰的frame中出現的移動物件進行互相補償的過程，因為該物件移動範圍不大且形狀在短時間內不易變動 • It is preceded by motion estimation Search Range W Motion Vector Reference Frame Current Frame

MPEG-1 Introduction • • • DCT for intra- and inter- frames Block-based motion compensation Huffman Coding Source input format : YUV, why not RGB Y : luminance component(亮度) U(Cb), V(Cr) : color difference or chrominance component（色差）

MPEG-1 Picture Type • Intra-Picture( I ) • Predicted- Pictures( P ) • Bidirectional-Predicted-Picture( B )

MPEG-2 • 與 MPEG-1 通用 • Difference from MPEG-1 – – – 好很多的畫質更多的輸入格式( ex: RGB, more YUV form) 在雙向傳輸時有較短的delay time Higher bit rate We can adjust our pictures quality