First-Class Sharing of Audio/Video
The previous project tries to realise sharing of a new kind of
``media'', virtual environments, that so far has not been shared.
Two other, related, media that can be shared are: audio and video.
Almost any conferencing system supports a simple form of sharing of
these media by allowing multiple participants to share video and audio
of the participants in the conference.
However, unlike the sharing of text/graphics objects,
this sharing is not ``first-class'' in many respects:
Arbitrary audio/video objects cannot be defined and shared.
For instance,
it is not possible to define a shared structured video object consisting
of (possibly hierarchical) scenes captured at different sites.
All users sharing an audio/video object are fully synchronized,
that is,
it is not possible for a set of users to compose and review a joint
message before transmitting it to their collaborators.
Finally,
flexible coupling.
concurrency control,
access control,
and
undo
are not available for
audio/video actions.
As mentioned above,
these operations are supported on the traditional text and graphics media -
the challenge here is to define them for audio and video.
Some of these problems have begun to be investigated by multimedia
authoring tools and multimedia research.
This project will extend this work by developing a new abstraction,
a multimedia object,
for defining shared audio/video objects,
a framework for controlled sharing and recovery of these objects,
a distributed shared memory model for flexibly
communicating these objects among
distributed processes,
a distributed architecture
for shared audio/video applications,
audio-video specific media adaptations added to
the collaboration-bus kernel to support interoperating audio/video
applications,
and a set of audio/video filters for defining the portions of
an audio/video clip that need to be shared/stored.
It will give you experience with research in
multimedia,
abstract data types,
interoperability,
concurrency control,
and distributed systems.
For more information, contact
Prasun Dewan.
or
Kevin Jeffay.