Image Processing Reference
The MPEG-4 wired content is only demonstrated in prototype form at present
although it was being heavily touted as a new technology in 2001. The iVAST (now known
as DG2L) and Envivio companies both had tools available as early as 2002.
Tools for developing the systems-layer interactive components (the BIFS stuff) are
beginning to appear. There are online resources such as the MPEG Industry Forum
(MPEGIF) mentioned above, with links to supplier Web sites where you can download
open-source kits and demonstrations of proprietary tools.
Multi-Plane Display Model
The BIFS scene allows a multi-plane presentation to be created, such as floating objects in
front of one another and changing their transparency. All of these effects are available with
coding systems prior to MPEG-4 but special code would have to be built into the applica-
tion to manage the effects. Now, with MPEG-4, it is an integral capability of the player.
This saves you a great deal of time as a developer of playback applications because this
player will inevitably be built in as an operating system component—but not yet. When it
is, the intelligence will be carried in the content rather than having to delegate decision-
making control to the application.
Because a BIFS scene is constructed by inserting a video object, you may be able to accom-
plish the effect you are aiming for by deconstructing your scene into a series of nested BIFS
packages. MPEG-4 allows you to embed another MPEG-4 package within the scene you
are working on. This goes beyond the permitted complexity of the profile and level that
you are working to. You should check that your player supports this but provided the
aggregation is compliant with the standard, it should work. It has been proven to be prac-
tical with the EnvivioTV plug-in.
If you recall, Chapter 7 was concerned with the compression of linear audio that has been
ingested along with the video. MPEG-4 supports a variety of different audio objects.
The MPEG-4 standard describes a lot more than just the AAC-format audio com-
pression that is becoming more popular. It also provides for some structured audio that
allows speech and music samples to be delivered with instructions on how to re-synthe-
size the output at the receiver/player. This offers a way to deliver extremely low bit-rate,
high-quality audio, but requires that the player has a complexity equal to that of a MIDI
sequencer and software synthesizer. That is currently beyond what MPEG-4 players are
capable of outside of a few laboratories, but it is an interesting direction for future devel-
opment, especially when combined with the BIFS scene description to allow sound
sources to be placed into a 3D environment.
These structured audio components would very likely be authored in other tools
such as MIDI editors or text-to-speech renderers—tools that are already widely available.
MPEG-4 support merely needs to be added incrementally.