IBM SG24-4817-00 - Manual (Page 66)

This soft copy for use by IBM employees only.
4.6.1.3 Motion Interpretation and Improved Compression
MPEG-2 interprets motion between successive images and takes advantage of
motion to improve compression while sustaining the same level of perceived
image quality. The MPEG-2 motion interpretation uses a combination of two
kinds of frames to do this:
• An anchor frame:
−
Intra-Frame (I-Frame):
- Only exploits spatial redundancy to compress information
within the frame
- Contains all information to reconstruct the image; does not depend
on another frame
• A difference frame, which can be either:
−
Predictive Frame (P-Frame):
- Exploits temporal and spatial redundancy to compress video frame
- Must reference a previous anchor frame to reconstruct the image
- Can be anchors to other P or B frames
−
Bidirectionally Predictive Frame (B-Frame)
- Exploits temporal and spatial redundancy to compress video frame
- Must reference an anchor frame
- Cannot be an anchor to another frame
A typical encoded frame sequence might look like the one in Figure 19 on
page 51.
50
ATM Workgroup Solutions: Implementing the 8285 ATM Switch