Summary of the study

The main objective of this study was to assess whether experiencing an environment
natural through immersive virtual technology could help individuals to
promote a positive affective state in them. There is a close relationship between promotion
of a positive affective state and stress reduction, because stress reduction involves
positive changes in psychophysiological states and often in behavior
and functioning, including cognitive performance.
In addition, the psychological component of stress reduction involves changes
positive in affectionate states, such as reduced levels of feelings to tone
negative (such as fear or anger) and increased levels of affect to tone
positive. Moreover, it seems synchronized with its physiological counterpart, so that the measurement of affect by physiological measures seems justified by both theories
stress reduction (Ulrich, R. S., et al., 1991) and attention restoration theories
(Kaplan, S., 1995).

The design of our experiment has allowed to compare the effect of different restoration environments reproduced according to different methodologies.

We reproduced three different virtual environments in terms of realism, dynamism, scene content and mode of exploration, which are described below. We hypothesized that since the three virtual environments are different from each other by their nature and composition, there would have a relative restorative effect on Activity-based markers of affective state brain of the participants, quantified by electroencephalogram.

​

Virtual environments

The virtual environments presented here are three faithful reproductions of natural forests located around Quebec City, Canada: Duchesnay Forest and Mount
Stoneham. All virtual environments have been reproduced under conditions
similar weather conditions, partly sunny and without breeze, in order to ensure that
all the scenes present the same constitution and the same structure of foliage.
Although lidar scans and immersive 360-degree video were not captured at
same time of year (lidar scans in spring and immersive video in autumn),
care was taken to ensure that vegetation, such as grass and foliage in the canopy, was not obscured by the dry weather and appears homogeneous among the three scenes.
In terms of content, the three scenes are characterized by the absence of close people
of the camera, closed but defined spatial configurations, the predominant presence
of native tree species, the exclusion of animal species and a first-hand perspective
no one.

​

The lidar forest
The lidar forest is a virtual forest entirely reconstructed from lidar data;
up to 21 different scans from a terrestrial lidar scanner were combined to reconstruct
an area of approximately 5 hectares (ha).
After merging the scans and coloring the point cloud from the images, the data
were downsampled and the Unreal Engine lidar plugin was used to visualize and
transmit the point cloud to the virtual reality head-mounted display.
The final environment is a static environment (i.e. without movement of
leaves or branches) but complex and with a spatial resolution of 1 point per half-
square centimeter. Participants experienced active exposure to this scene, as they could turn their heads freely, observe their surroundings, and voluntarily explore the path demarcated forest using the joystick.

​

The hybrid forest
The hybrid forest was designed using an innovative and personalized procedure that
combines the use of lidar data and three-dimensional models to reproduce
virtual environments.
To do this, the ground surface was reconstructed using only lidar data.
up to a height of about 50 cm, so that the elements adjacent to the ground, such as
rocks, dead leaves, branches and even small bushes and trunks
sectioned, can be represented by point clouds.
A trunk map provided the positions and diameters of the trees present in the forest
where the scans were taken; these data were used to place the trees in their
original positions through the virtual environment.
Three-dimensional models of sugar maple, basswood, balsam fir and birch to
paper were drawn using SpeedTree 3D software, then imported into the
Unreal game.
The final environment consisted of a forest environment with approximately 500 trees in the field of vision; the tops of these trees reacted to variations in the simulated wind and the interaction of light with tree branches could be appreciated. The
participants, as in the previous environment, experienced active exposure to this
scene, can freely turn his head and move freely along the forest path
bounded using the joystick.

​

360 degree forest
360 degree forest is a 360 degree video produced using an action camera
high definition GoPro Max and a DJI Mavic Pro drone.
The video simulated idling for about 10 minutes during which you could
see a stream, tree trunks and tree cover of native species. For
mask the sound of the drone's propellers, a second panning, this time along the
same path, was recorded to capture the sounds of forest and flowing water. Both videos were processed with the GoPro Player utility and edited with Adobe Premier
PRO.
The final scene consisted of an immersive video at 5.7k resolution, encoded with the
h.264 codec and presented to attendees with the VIVEPORT VIDEO app. A
particular attention was paid to immersive sound editing, where noise dominates
constant water and air, as well as the movement of the treetops.

The participants were able to experience this scene passively because, although the
participant can move his head and freely observe his surroundings, the movement of
the camera was unintentional and its speed could not be changed; therefore, the user also could not move at will.