Presence in Virtual Environments Mel Slater Department of Computer - - PowerPoint PPT Presentation

Presence in Virtual Environments

Mel Slater Department of Computer Science UCL

www.cs.ucl.ac.uk/staff/m.slater

Outline

• Evaluation of VEs
• Immersion and Presence in Virtual Environments
• Four experiments with the ‘pit room’
• Ways of Walking
• Latency
• Static Visual Realism
• Dynamic Visual Realism
• Experiment with virtual characters
• Real man meets virtual woman
• Teaching by punishment
• Conclusions

Evaluating VEs

• Task performance is an important measure
• f the effectiveness of a VE
• Task performance within the VE
• Task performance in the transfer of learning

from the VE to the real world

• Task performance – specific to applications
• ‘Presence’ is common across applications

and has been used as an overall measure of the effectiveness of VEs.

Immersion: the technology

• Inclusive – sensory experience from VE only
• Extensive – more sensory modalities
• Surrounding – from all directions
• Vivid – high fidelity
• Egocentric – first person point of view
• Plot – things are happening
• Proprioceptive match – between sensory data and

proprioception

Applications of VR in Scientific Visualisation

• Obvious applications of VR in visualisation, modelling, etc
• Above is collaboration with Andy Van Dam at Brown

University, Mavi Sanchez-Vives at UMH, Spain, …

• Visualisation of neurons obtained through confocal microscope images

Some Presence Definitions

• ‘The sense of “being there”’ (Held & Durlach,

Sheridan, Zeltzer: premier issue of PRESENCE, 1992)

• ‘A perceptual illusion of nonmediation’ (Lombard

and Ditton, 1997)

• ‘A mental state in which a user feels physically

present within the computer-mediated environment’ (Draper & Kaber, 1998)

• ‘The subjective experience of being in one place
• r environment, even when one is physically

situated in another’ (Witmer & Singer, 1998)

Presence Operationally

• Successful substitution of real sense data by

computer generated sense data

• ‘Successful’ – response is realistic, similar

to what would be expected if the events were real

• ‘Response’ –
• Low level physiological → high level cognitive

and emotional

• Includes verbal responses about ‘being there’

Measurement of Presence

• Questionnaires & Self Report Measures
• Behavioural responses
• Physiological responses
• Breaks in presence

The ‘Pit Room’ Experiments

Walking in Place

• Hypothesis that the correlation between

proprioception and sensory data is an important factor in maintaining presence

• Navigating a VE by using a ‘mouse’ breaks

this match

• Compared a ‘walking in place’ method with

a point-and-click (‘flying’) method

Slater M, Usoh M, Steed A (1995) Taking steps: the influence of a walking technique

• n presence in virtual reality. ACM Trans Comput-Hum Interact 2:201-219.

Experiment

• 16 subjects
• 8 in a walking-in-place group
• 8 in a point-and-click group
• Task was to take an object to the chair
• A between groups design
• Each subject had a virtual body that

they could see when looking down

• HMD display with Virtual Research

Flight Helmet 360 × 240 with 75 deg. horizontal FOV

• Polhemus sensors for head tracking

Some Measured Variables

• Presence Questionnaire
• Three questions
• The sense of ‘being there’
• Whether the environment the virtual room was remembered as

somewhere visited rather than only images seen

• The extent to which the pit room became the dominant reality,

and that the real lab was forgotten

• Each measured on a 7-point scale
• (1) Not at all
• (7) Very much so
• Final score is x = the number of ‘6’ or ‘7’ answers

given (x = 0,1,2,3)

• Path taken to the chair
• Extent to which they ‘associated’ with their

Virtual Body

Results

• For the ‘walkers’ – greater their association

with their VB the greater their presence score

• For the ‘pointers’ – no correlation
• If they associated with their VB then the

‘walkers’ reported higher presence than the ‘pointers’

• Path across the precipice associated with

lower reported presence.

Walking, Walking In Place, Flying

• A follow-up study compared
• (RW) Really walking (UNC wide-area tracker)
• (WIP) Walking in place
• (PAC) Point and click (‘flying’)
• Between groups experiment (11 subjects per group)
• Results showed (RW,WIP) >> PAC
• No sig. difference between RW and WIP

Usoh M, Arthur K, Whitton MC, Bastos R, Steed A, Slater M, Brooks Jnr F.P (1999) Walking > walking-in-place > flying, in virtual environments. (SIGGRAPH ’99), 359-364

Further Studies

• Meehan/Insko (2002)

exposed 10, 52 and 33 subjects in 3 different studies.

• Heart rate increase when

in the pit room.

• Static haptics further

significantly increased heart rate.

• Heart rate correlated with

subjective self-report of presence level.

Meehan M, et al (2002) Physiological measures of presence in stressful virtual environments. Acm Transactions on Graphics 21:645-652 Insko B et al (2001) Passive Haptics Significantly Enhances Virtual Environments. Department of Computer Science, p 111. UNC Chapel Hill:

Influence of Latency

• Lower and Higher latency

conditions in a ‘pit’ experiment.

• Skin response, HR and

presence questionnaire used.

• HR change most sensitive.
• Meehan et al VR’2003

Lower latency ∼ higher ‘presence’

Global Illumination – Who Cares?

• Why spend effort and resources on ‘global

illumination’ (eg, ray tracing, radiosity, photon mapping,…)?

• Scientific and practical question?
• What is its impact on people who are in

immersive virtual environments?

Visual Realism

• In a between-groups study utilising the ‘pit room’

Zimmons et al. (2003) used 5 levels of rendering:

• From wire frame to radiosity (global illumination)
• All subjects showed increased heart rate when approach

the precipice in all conditions

• No significant differences between the conditions in

heart rate or reported presence.

• So, from the point of view of presence is visual

realism not important?

Does Ray Tracing Help?

• What do we get with ray tracing?
• Shadows (umbras only)
• Reflections
• Cost – long rendering times!
• Real-time ray-tracing (parallelism + GPU)
• ffers chance to investigate impact of

dynamic changes (shadows+reflections) on presence

Experimental Design

• 34 people
• Displayed with V8 HMD + polhemus tracking
• Pit room rendered in two ways
• Real-time Ray Tracing with reflections and shadows of

virtual body

• OpenGL shading (no reflections/shadows)
• Virtual Lightfield Project at UCL
• Pankaj Khanna, Jesper Mortensen, Insu Yu

Experimental Design

• Subjects divided into two groups
• 17 experienced first RT and then OpenGL
• 17 experienced first OpenGL and then RT
• Both groups had relaxation periods for ‘baseline’

readings

• This design is both between groups
• Consider 1st results only
• Within groups
• Consider comparisons between the two results
• (with high danger of interference between the two

experiences!)

• (Experiment finished last week!)
• VIDEO of early pilot experiment

Measured Variables

• Elaborated questionnaire presence
• Skin conductance
• Heart rate and respiration
• Demographic variables – age, gender
• Background
• Programming knowledge, game playing, prior

experience of VR, etc

Early Results

• Consider the between groups experiment
• nly (comparing responses to the first

experiences only):-

• Reported presence is sig. associated with:
• Ray tracing (higher than for OpenGL)
• Prior VR experience (+vely related)
• Programming experience (-vely related)
• Age (-vely related)

• Most interesting VEs are
• those that involve other
• people
• Real online people
• Represented by avatars
• Virtual people
• Represented as humanoid

Interactions with Virtual People

Pertaub, D-P., et al. (2002) An Experiment on Public Speaking Anxiety in Response to Three Different Types of Virtual Audience, Presence, 11(1) 68-78.

Does ‘ audience’ behaviour influence anxiety response?

Fear of Public S peaking

Positive Audience Behaviours

pay attention lean forwards encouraging noises maintain eye contact Enthusiastic applause nod encouragingly Smile frequently standing ovation

Reactions

• ‘ It was clear that the audience was really

positive and interested in what I was saying and it made you feel like telling them what you know.’

• ‘ I felt great. Finally nobody was interrupting
• me. .. Here I felt people were there to

listen to me.’

• ‘ They were staring at me. They loved you

unconditionally, you could say anything, you didn’ t have to work’ .

Negative Audience Behaviours

yawn, cough,

mumble

fall asleep talk amongst

themselves

put feet on table read papers/notes avoid eye contact turn torso away

from speaker

frown and grimace walk out of room

Reactions

• ‘ It felt really bad. I couldn’ t j ust ignore
• them. I had to talk to them and tell them to

sit up and pay attention. Especially the man

• n the left who put his head in his hands; I

had to ask him to sit up and listen… . I entered a negative feedback loop where I would receive bad responses from the audience and my performance would get even worse… . I was performing really badly and that doesn’ t normally happen.’

Real Man Meets Virtual Woman

• Continuing research into social phobia
• Empathic Avatars project, UCL - Xueni Pan
• How do shy men respond to ‘advances’ from

(virtual) women?

• Can this be used in exposure therapy?
• Training?
• Analysing a study now

VIDEO from pilot studies (1:12) (2:53)(4:21)

Milgram Experiment

• Obedience to Authority
• Stanley Milgram (1963). Behavioral study of obedience. Journal of

Abnormal and Social Psychology, Vol. 67, pp. 371-378.

• Stanley Milgram (1983). Obedience to Authority: An

Experimental View. New York: Harper/Collins.

• Very famous – produced a huge debate that led to today’s

standards of ethics in human experiments.

• A version of this experiment was carried out in a VE
• Slater M, Antley A, Davison A, Swapp D, Guger C, et al. (2006) A

Virtual Reprise of the Stanley Milgram Obedience Experiments. PLoS ONE 1(1): e39. doi:10.1371/journal.pone.0000039

• Results showed strong responses at subjective,

physiological and behavioural levels.

Summary

• Immersion describes the technology
• Presence as ‘successful substitution’ of virtually generated

sensory data for real sensory data

• There is clear evidence that presence occurs
• There are many ways to measure it, none satisfactory alone

– has to be multi-level

• Very much a young and open field ready for new research

ideas.

• Higly interdisciplinary
• PRESENCCIA (www.presenccia.org) has neuroscientists,

computer scientists, bio-mechanical engineers, psychologists, etc.

What do you need to have?

• Curiosity about why ‘virtual reality’ works for people
• Knowledge of the capabilities and possibilities of VR

systems

• Ability to create VR scenarios
• Understanding of scientific method
• Hypothesis formation
• Experimental Design
• Statistical Analysis
• Interest in human responses
• Physiology (including brain activity – EEG)
• Behavioural (analysis of video, measurement of movement,…)
• Subjective and verbal (questionnaires, interviews)

Approaches

• Two complementary approaches to realisation of

presence in VEs

• Engineering – construct algorithms and devices that

create high fidelity virtual environments

• Perceptual/Neurological – understand what is needed in

perceptual terms to create presence.

• Sanchez-Vives & Slater (2005) From Presence to

Consciousness Through Virtual Reality, Nature Reviews Neuroscience, 6(4) 332-339