Next: Part II Middleware
Up: Coding and Compression
Previous: Processing Requirements for Video
In this chapter we have looked at Video and Audio coding and
compression techniques. We have also introduced some preliminary ideas
in conferencing and conference control protocols, by virtue
of their being part of one of the dominant standards for compression,
namely the H.320 family: more of this in chapter six.
Important lessons for people using multimedia
include:
- 1.
- Multimedia is large! The same amount of information can often be represented
for human consumption in text taking far less storage or transmission
than in graphical form.
- 2.
- Multimedia is highly redundant, and compresses very successfully
- 3.
- Human perception is frail, and cognition smart. We can take advantage
of it - note that the auditory and visual systems, however, are quite
different, with persistence, averaging, and so on happening at
different stages in each.
- 4.
- Users of some critical applications cannot tolerate
any artifacts, and therefore prefer lossless compression - examples
are typically in the area of medical imaging
(Radiology, Ultrasound, Functional imaging from magnetic resonance
imaging and so on).
- 5.
- There are a number of ways to multiplex a set of media together, and a
set of engineering tradeoffs associated with these. Different
industries have made different choices for good reasons. However, this
means that in the Internet, we have to unpick some of the multiplexes
chosen in the broadcast and telephony industries before gaining the
advantanges of the Internet's full generality.
- 6.
- Last but not least, it is very important that the more sophisticated
the compression technique (at least for video) the harder it is in
processing terms; but decompression is massively less expensive than
compression. This means that one-to-many applications are cost
effective.
Next: Part II Middleware
Up: Coding and Compression
Previous: Processing Requirements for Video
Jon CROWCROFT
1998-12-03