With 360 degree vr panoramas, especially vertically-wide
ones or spherical ones, optic flow is often a striking
phenomena. Looking around by dragging with the mouse,
causes the world in the scene to spin - dragging right say
will cause the objects in the view to move leftwards in
the frame (this is the normal convention with handling
mouse dragging with  vr panoramas). Dragging up so the
view is directly up and then dragging right again will cause
to view to spin round anticlockwise with the apparent centre
of rotation in the middle of the frame, look (drag) right
down and and drag right and the view spins clockwise.

This is how it is in the real world too if we stand on a spot
and turn to the right while looking up, ahead and down.
In general dragging in various directions with panoramic vr
content creates optic flow patterns in the viewer window
which are direct analogues of the real world visual events
that occur as we look around and up and down in actual locations.
(Zooming in a vr scene also has particular patterns of
optic flow events - they are not identical but similar to
real world experiences when we are navigating scenes,
approaching or receding from objects.)

    In the real world we might quite often stand on a spot and
look laterally all the way round but less commonly do we
do this while looking directly up and down. On a children's
roundabout about we might get a spinning view of the
visual world above or a jet pilot in a spiralling plane might
get a revolving view of the earth below. These optic flow
patterns are strong generators of felt sensations of tilt and
rotation. A Swedish psychologist Johansson has shown
that a very sparse pattern of vertically moving lights in
the peripheral vision of a subject is sufficient to generate
strong sensations in the subject that she
and her surroundings are moving together in the opposite
direction. So it seems that optic flow interpretation is
encoded in the human perceptual system at a quite
fundamental level. It would be surprising if was not an
important factor too in the 3d impression that can be
experienced so commonly with panoramic vr scenes.
http://www.erlbaum.com/380.htm
 

One reason perhaps why vertically wide panoramas in
particular can  seem very 3d  in a panoramic viewer
is that when the virtual camera that the panoramic vr
viewer provides is set to a wide zoom setting (only possible with
scenes with lots of inherent vertical coverage) the real world
types of optic flow patterns associated with turning around
and looking up and down are more apparent. When the virtual
camera in a vr panorama is level, for example, if the view is
wide vertically and horizontally the optic flow will consist
of a central band of flow with basically parallel and lateral
optic flow patterns while in top and bottom regions of the
frame  there will be offcentre circular patterns revolving
in different directions. There is a lot of research also that confirms
that people with tunnel vision (corresponding to a narrow
fov in vr panorama) have serious deficits in their pathfinding
capacity - perhaps the lack of distinctive optic flow patterns
is a factor in both situations.
See also on "Effects of Field of View on Performance with
Head-Mounted Displays":
http://www.cs.unc.edu/~arthur/fov.html
 

    Interesting new research indicates that  Alzheimer's patients
may have difficulty perceiving optic flow patterns - perhaps
this rather than memory problems accounts for their frequent
navigation difficulties - ie getting lost.
    http://more.abcnews.go.com/sections/living/inyourhead/allinyourhead_35.html
This finding suggest that such patients may have difficulty too in
navigating vr panoramas especially if there not other cues in
the subject matter to horizon, verticality and zoom level. One
could test this with a random dot spherical panoramas maybe.
 

     We spend much of our time moving from place to place in the
world rather than just looking about. The optic flow seen by an
observer moving through the
world in a static scene are characterised
by a pattern of expanding lines (the focus of expansions FOE) in the
direction the observer is heading. In a presentation consisting of two vr
panoramas linked by a linear movie such as this one constructed by
Bill Meikle - the contrast in the optic flow patterns in the two types
of content (panorama and linear movie) are obvious.
http://www.mountain-inter.net/~bmeikle/lineartrans.mov
(click on the show hotspots button to find the link to the next scene)
 

for general recent psychological research on optic flow fields:
http://www.perceptionweb.com/ecvp99/mon.html

for spatial perception experiments using virtual reality environments
http://www.mit.edu/people/beall/ve.html
http://www.net1plus.com/users/ohl/thesis/discuss.htm

on the design of virtual reality environments
http://www.man.ac.uk/MVC/SIMA/vrml_design/toc.html
this page in this paper on the "Art of Memory" seems
particular stimulating wrt multinode qtvr environments:
http://www.man.ac.uk/MVC/SIMA/vrml_design/usesf.html
- from these papers:
"If we survey a wide angle on a scene, using 'undistorted'
3P-perspective, the resulting image tends to have what
we subjectively regard as distortions at the edges -
things seem overlarge, or stretched, at the perimeter of the
view (Dubery and Willats 1972, p87-89). However, if we
distort the objects, as though they were projected onto a
curved picture-plane, we create what appears to be a
more 'natural' view (ibid p89-90). In addition, we gain the
advantage, as Van Gogh does in his painting Vincent's
Room, that viewers 'sense' their own location in relation to
the scene - a key requirement of virtuality. Vincent's
Room feels like a small, intimate space, because of, not in
spite of, the 'distortions' imposed on it."
this provokes the question - what kind of distortions
away from rectilinearity and central unitary projection
might be useful for  panorama vr viewers - earlier versions
of QuicktimeVR had "partial correction" corresponding to
an architectural shift lens - this was very suitable for some
built environment vrs.

The Smoothmove viewer (now iMove)
has the potential for extreme wide angle real time rectilinear
view generation. At settings over say 130 degrees HFOV
the distortions are extreme even though the view is rectilinear.
Some subjects though look good even with extreme HFOV.
On rectilinear perspective/rectangular frame
as a convention in computer games
and very interesting discussions of alternative perspectives and
camera conventions in computer animations see
http://jupiter.ucsd.edu/~manovich/text/cinema-cultural.html
(in particular see his references to Waliczky)  see also:
http://www.ntticc.or.jp/preactivities/gallery/trilogy/exh/duguet_e.html
http://www.ntticc.or.jp/preactivities/gallery/trilogy/exh/shirai_e.html

On the psychology of immersion in film display see:
http://iml.umkc.edu/comm/faculty/gutenko/papers/pursui.htm
he discusses field of view, projection geometries, frame rate
and  camera mobility as they affect perceived transparency, 3d
and physiological impact - see also his article on time compression
http://iml.umkc.edu/comm/faculty/gutenko/papers/speed.htm

JJGibson - the American psychologist was the first to note
the importance of optic flow and observer mobility in space perception
- there are very interesting (unpublished) papers here on
various space and object perception topics:
http://www.trincoll.edu/depts/psyc/perils/
in http://www.trincoll.edu/depts/psyc/perils/folder6/exploring.html
(unpublished) he perhaps anticipates vr panoramas in his remark -
"A completely panoramic "image" of the environment would be
equivalent in some ways to the ambient optic array." (1972)

There are  intriguing psychological problems involved in event
and spatial perception in
panoramic movies. Seen as a conventional movie a series
of panoramic (say cylindrical) frames at video rates of say
people entering a room and sitting down at a table filmed from
the middle of the table will likely have people moving out of
frame at one end and sliding into view at the other. Seen in dewarped
partial fashion in a vr view of course it seems perfectly natural.
WarpTV (used to be www.warp.com - had a demo like this).
Behere's panoramic video iVideo
http://www.behere.com/
looks like this if you dont have the right addon for your Real player.

A panoramic view even if uncorrected can still contain lots of
information - this applies to panoramic movies too:
http://www.dres.dnd.ca/Products/RD96006/
note this system gives you a number of options for viewing panoramic
movies - "These include a panoramic 360 degree strip view, "situational
awareness" and "zoom" windows or an immersive virtual reality format"
 

discussion of "visual ecologies" to follow

Copyright: Peter Murphy 1999