Image Processing Reference
The Future Is Digital
Analog TV is essentially dead. Plans are being drawn up for the eventual switch-off and
replacement with digital TV services. In fact this needs to happen, otherwise the band-
width won't be available for the high-definition TV services that are being planned for
The city of Berlin has already turned off all the analog transmissions, and small com-
munities elsewhere in the world are already well on the way to migrating to an all-digital
Which Codec Will Win?
H.264 is clearly the future as far as video coding is concerned, at least for the next 10 to 15
years and possibly beyond that. The VC-1 codec may become relevant if the standards-
ratification process moves forward rapidly enough. This is running into some difficulties
and is taking far longer than first expected.
MPEG-4 as a family of standards is extremely important. It brings together many
diverse areas of work under one umbrella and standardizes the whole multimedia-author-
ing environment. Using that MPEG-4 framework with MPEG-7 and MPEG-21 will also be
important as we describe content with metadata and attempt to rights control its deploy-
ment in an interoperable way.
Most of what we need to make MPEG-4 content is already available. It's just that we
create multimedia in a variety of different ways on different devices. The changes required
to support MPEG-4 are largely incremental across the portfolio of digital-media authoring
tools. It simply requires some incentive to encourage people to manufacture the plug-ins
or minor upgrades.
The encoding process for any codec that expects to be dominant and long lived must
be capable of implementation at a price the industry is willing to pay. This requires that
the complexity of compression not become a burden on the CPU, in order to keep com-
pression times low.
The dominant codec must also provide the best available compression ratio. This is
the factor by which the output is smaller than the input data.
It is also vitally important that the decoding workload be capable of implementation
in very inexpensive hardware. Some HD tests have demonstrated that HDTV coded with
H.264 is borderline on some very powerful desktop systems when played back. The situ-
ation for VC-1 is even worse because it suffers from a higher decoder complexity. This is
not going to help its chances of being adopted.
Special hardware decoder chips will solve this, and it's not expected to cause any
great delay to the rollout of HDTV services delivered with H.264 coding.
Analog switch-off in the United Kingdom: