Professional Documents
Culture Documents
NAME: ANKITA NANDY Click to edit Master subtitle style DEPT.:CSE YEAR:3RD ROLL: 45 ST. THOMAS COLLEGE OF ENGINEERING TECHNOLOGY
&
6/11/12
OUTLINE
INTRODUCTION BASIC BUILDING BLOCKS OF COMPRESSION VIDEO COMPRESSION STANDARDS CONCLUSION
6/11/12
INTRODUCTION
Why video is compressed?
Raw video contains an immense amount of data Communication and storage capabilities are limited and expensive To reduce redundancy and irrelevancy
6/11/12
Types of redundancy:Spatial Redundancy: Neighboring pixels in a video frame are highly correlated with each other. Temporal Redundancy: Corresponding pixels in neighboring video frames are correlated. Psycho visual Redundancy: Some details in video frames are not perceptually important to the human eye.
6/11/12
BLOCKS
OF
Video clip contains a series of still images(frames). Hence, still image compression techniques may also be employed during video compression. Therefore we begin our discussion by explaining the inner working of transform based digital image compression algorithms like JPEG.
6/11/12
BLOCKS
OF
Block transform:
Partition image into 8x8-pixel blocks 2-D DCT of each block convert each pixel to frequency domain coefficient original block can be reconstructed using an inverse DCT
6/11/12
BLOCKS
OF
Coding:
Two types of coding is implemented Run-length-encoding-Consecutive zeros are grouped and the
number of zeros in the group is encoded. The DCT coefficients are encoded in a zig-zag fashion, starting from top-left and ending at bottom-right. Variable length coding-More frequently occurring symbols (either a run-length encoded number or a DCT-coefficient) are coded using fewer bits, and infrequent symbols are encoded using more bits.
6/11/12
BLOCKS
OF
Human eye has lower sensitivity to color than brightness The chrominance components U and V are reduced to half of their original value
Fig: Depending on the sub sampling, 2 or 4 pixel values of the chrominance channel can be grouped together.
6/11/12
BLOCKS
OF
Main addition over image compression is exploitation of temporal redundancy Here current frames are predicted based on previously coded frames Three types of frames are used: IFrame: Intra coded frames, coded independently of all other frames. P-Frame: predictively coded frame, coded based on previously coded frame. B-Frame: Bidirectionally predicted frame, coded based on both previous and future coded frames.
6/11/12
BLOCKS
OF
Steps involved in motion estimation are Frame segmentation Search threshold Block matching Prediction error coding Vector coding Block coding
6/11/12
BLOCKS
OF
artifacts:
Some visible artifacts may occur due to poor encoder implementation Two types of artifacts: Blocking Ringing Can be reduced by using low pass FIR(finite impulse response) filters.
6/11/12
COMPRESSION STANDARDS
The following compression standards are the most known nowadays. Each of them is suited for specific applications.
Standard JPEG H.261 MPEG-1 MPEG-2 H.263 MPEG-4 JPEG-2000 DV Application Still image compression Video conferencing over ISDN Bit rate Variable P x 64 kb/s
Video on digital storage media(CD-1.5 Mbit/sec ROM) Digital television Video telephony over PSTN Object based coding, content, Interactivity 1.5-15 Mbit/s 33.6kb/s syntheticVariable Variable 50 Mbit/s 6/11/12
CONCLUSION
Video
compression is important. It is not easy to implement. Video compression has come a long way but still definitely has room for improvement. New paradigms in computing will dictate future research directions
6/11/12
Thank You!!
6/11/12