IEEE Consumer Electronics Magazine - July 2018 - 47

The first set of participants (group A) did not encounter
other characters while in the VE, and the instructions were
text based. The second (B) and third groups (C) encountered
a virtual salesman, who, in group B was an NPC, and in
group C was a 3-D human model controlled via real-time
motion capture based on the motions of a real actor. Groups
B and C received auditory scripted instructions, which were
prerecorded for group B but spoken in real time by the actor
via intercommunication for group C. All participants could
hear their avatar's footsteps as well as an ambient sound
resembling a silent room.
The experiment was carried out in two adjoining rooms of a
research laboratory; the first room hosted private, individual
participant sessions, whereas the researcher who wore the
motion-capture equipment and controlled the salesman for
group C, executed his performance in the second room. Participants encountered a female researcher who was responsible for
scheduling and monitoring sessions, including orientation and
interviews, and thus, did not encounter or have any knowledge
of the male researcher in the next room.
Upon arrival, participants were greeted by the female
researcher and were provided with a written description of
the equipment that would be used. This included an HMD, a
standard gaming controller, and an audio headset. Participants were not aware of the group divisions or the specific
characteristics of each of the three experimental conditions.
Before entering the virtual store, participants spent a few
minutes in a smaller, rectangular virtual room, referred to as the
anteroom, which only contained one full-body reflective mirror
(Figure 1). Verbal instructions and assistance were provided by
the female researcher who was monitoring the sessions. In the
anteroom, participants were provided the opportunity to practice orientation and virtual locomotion. The anteroom also
allowed participants to realize that their head movements were
being tracked to match the avatar's head movements; therefore,
when they looked down at the floor, they saw the avatar's chest
and legs, and when they were facing the avatar's reflection in
the mirror, all real-head movements were corresponding to the
avatar's head movements in real time. Finally, female and male
avatars were assigned to female and male participants, respectively (Figure 2).

FiGuRe 1. The top view of the anteroom.

FiGuRe 2. Female and male 3-D human models.

Once in the clothing store, the female researcher did not provide any further assistance, and participants had to follow
scripted instructions. The virtual clothing store was designed as
a square room featuring a few shelves with folded clothes, a
cash register, and two full-body reflective mirrors. The first mirror was opposite the register, therefore located in the central
open area of the store (CrM), whereas the second mirror, placed
to the right of the cash register, was partially hidden behind a
curtain, forming a small dressing room (DrM) (Figure 3). Once
behind the curtain, participants had no visual contact with the
cash register, and subsequently had no visual contact with the
salesman, who took his place behind the cash register as soon
as participants in groups B and C entered the store.
Upon entering the clothing store, the participants were
encouraged, via scripted instructions, to explore the room until
their preordered outfit appeared on top of the cash register
(which was programmed to occur after 40 s). Once the outfit
appeared, the participants were invited to approach the cash register and collect it by pressing a button on their controller. This
activated the virtual hand's animation, whereby the participants
could see it extending as if it was picking up the item (Figure 4).
The item then disappeared, presumed to be in their possession.

FiGuRe 3. The top view of the clothing store.

July 2018

IEEE Consumer Electronics Magazine

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - July 2018