Collaborative Virtual
Reality
People have been sharing
virtual worlds of different kinds for quite a while now
One way to do this is to have a large shared display like a CAVE or CAVE2 showing a single virtual world where multiple people in the same location share that virtual world where one person controls the viewpoint but one or more people might have interaction control. The participants see each other's real bodies, and talk to each other directly. This is co-located collaboration. |
If the
participants are not co-located then more infrastructure is
needed. In this case each participant is in their own virtual
world (whether via HMD or a large display) and is connected via
a server to other participants sharing the same space.
Participants have avatar bodies - representations of their
tracked selves in the virtual world for others to see - and
voice (and / or video) is streamed between the participants.
Back in week 3 we showed a video on Dactyl Nightmare from 1991 where players wearing HMDs could see each other and shoot each other - https://www.youtube.com/watch_popup?v=dji9YiPZ4AM
Local Work from the mid to late 1990s exploring a range of options:
Caterpillar - audio and video on moving screens that showed the position of the various collaborators - the screens were located where the users were located in the space, and the videos always oriented themselves (billboard) towards the user on the local site. The lack of articulated avatars made it difficult to point in the virtual space, but the trade-off was that you had a richer view of the expressions of the remote users. At the time available bandwidth only allowed low resolution greyscale video, but today you could stream much higher resolution as in skype or google hangout. Video of this kind works very well with fish tank VR setups, less well with room scale setups, and not very well at all with HMDs. | |
Virtual Temporal Bone - full body avatars would get in the way of this educational environment so only different coloured pointers are used showing where the user's tracked wand is in the space. Full body avatars also tend to give a sense of up and down which may not make sense in worlds such as this. Audio connections between the remote sites allow the doctor to explain the anatomy and students to ask questions, while being able to point at what they are talking about. This kind of collaboration works fine in any VR display. | |
CALVIN - simple
articulated avatars were used in this design environment
that encouraged people to work at different scales to set
up a configurable room. Tracking the head and hand of the
user and providing an audio link allowed people to see
where each other were looking and pointing. This kind of
collaboration works fine in any VR display https://www.youtube.com/watch_popup?v=ZYY8JdFgCAc |
|
NICE - more
articulated avatars in an educational setting, and usage
by more casual users (issues of being able to see
yourself, issues of 'equality', emergent social patterns) https://www.youtube.com/watch_popup?v=ihGXa21qLms and a short paper about the issues of scaling this up a bit (nametags, colored costumes, etc.) |
|
TIDE - pointers with
static photographs attached. One issue with pointers (and
avatars) is knowing who is who. Attaching a name to a
pointer is one solution where you can quickly talk to the
person with the appropriate pointer since you see their
name (assuming a shared character set). If you already
know the people in the collaboration you can use photos to
identify the pointers. https://www.youtube.com/watch_popup?v=viA1EOGLrN0 |
|
Round Earth -
participants in a shared experience may not have the same
capabilities in the shared world. Here one person was on
the surface of a small asteroid moving forward / back /
left / right, while the other orbited above the asteroid.
Having a body for the astronaut on the surface allowed the
person guiding from orbit to know where the person they
were guiding was located and which way they were facing
(though a bright arrow made this more obvious).
Negotiating and remapping directions were key to
accomplishing the mission. https://www.youtube.com/watch_popup?v=Hh-6rK9TP28 |
Audio as the most important channel
Video as a window, video
as the avatar, video for just the avatar face
You can also scan people in and have 3D realistic, but not mobile, 'statue' avatars
Today you can also get yourself scanned in a 3D scanner and generate an articulated avatar of yourself
for example - https://www.youtube.com/watch_popup?v=DYllOiFmwdc
Some more avatar bodies:
- you may want to choose one that shows where you are from
- another to make you look a bit more 'professional' - note the virtual 'name tag' hovering over the avatar to help identify who this is and where he/she is looking
Another issue
to keep in mind when you have multiple avatars in a space is
whether they will collide with each other and/or the walls and
objects in the space. Trying to get a bunch of people moving
down a compact hallway can be extremely challenging for people -
think about a group of people all trying to get off an elevator
at the same time.
Avatars in Augmented
Reality
Heterogeneous Perspectives
People may not be seeing the
same exact shared world, especially in science and engineering
domains.
Asynchronous Work
While
collaborative VR across continents is rather easy to do
technologically, somebody has to stay late or come in early to
collaborate synchronously with people on the other side of the
planet. Sometimes asynchronous collaboration is better, but how
do you coordinate your work?
One way is to record your
voice and gestures in the virtual world and then allow others to
play them back later on
Virtual Harlem / V-Mail