trends in VR non-fiction

a retrospective study of virtual reality nonfiction experiences from 2012 - 2018

Analysing a representative sample of 150 different types of VRNF (VR Non-Fiction) experiences released between 2012 - 2018 the researchers identified 64 different characteristics of the medium during this period.

1. The VR medium is emergent and fragmented, influenced by documentary and journalism practice, as much as computer games and immersive theatre.
PRODUCERS
2. Producers range from international news outlets, specialist studios to independent artists, distributing through specialist festivals.
FEATURES
3. Full immersion is a distinguishing feature (whilst acknowledging that no embodied experience can be said to be completely immersive), as opposed to mixed reality which does not filter out visual representations of the real world.
4.Terms often associated with the medium include 'embodiment', 'immersion', 'presence', 'plausibility illusions' and 'empathy' (the researchers note that this latter term is contentious given that VR is more likely to offer a close dialogue with challenging experience, rather than a complete insight into that experience, or lifestyle change). 
5. Talk of VR as empathetic media was 1st promoted by Chris Milk who produced Clouds Over Sidra (2015), a 360 video presentation of a 13 yr old girl's experience of a refugee camp.  Milk described VR as the 'ultimate empathy machine'.
MEDIA
6.Typically, VRNF uses a combination of interactive elements and traditional visual storytelling techniques (video, animation, audio, editing, voiceover etc.) presented as a 360 degree panorama via a VRHMD (head mounted display).
PLATFORM
7. 83% of the sample were platform independent, 360 video presentations.
PARTICIPATION
8. In 69% of the sample, viewers assume a passive observer role.  In some cases viewers transition from passive observer to more active roles whereby they can activate cut scenes (active observation) or interact with characters in the scene without changing the story (passive participant).
POINT OF VIEW
9. All of the titles used one or more of three distinct perspectives, or points of view (POV): first person (9%), fly-on-the-wall (18%) and omniscient (23%).  50% of the sample incorporated 2 or more of these.
10. There were several instances of direct address in an objective point of view context, which the researchers suggest is unhelpful as it breaks the 4th wall.
ANIMATION
11. 73% of the titles were entirely or predominantly composed using 360 live action video footage.
12. Just over half (53%) of the titles included at least some CGI/animated content, with 27% being composed entirely using animation.
13. In some titles a real world environment, as well as real people are digitally reconstructed using photogrammatry (e.g. environment: Nefertari: Journey to Eternity, 2018, actors: Buzz Aldrin: Cycling pathways to Mars, 2017 or After Solitary, 2017)
GRAPHICS
14. To support narrative, 64% of the titles included non-diegetic graphical annotations (such as overlaying the name of a character who is speaking), while 27% included diegetic graphical annotations such as text appearing on walls. 
AUDIO
15. Only 8% of the sample used a spatial audio sound-field.  Examples include Easter Rising: Voice of a Rebel (2016) and I am a Man (2018).  Notes on Blindness: Into Darkness (2016) used binaural sound.
16. 82% included a non-diegetic score, where background music did not emanate from within the virtual world but instead was overlaid upon it.
GUIDANCE
17. 3 strategies to guide viewer attention were identified (i) prompt by an actor within a scene (46%), (ii) use of a graphical directional prompt, like offset subtitles towards direction of new scene (12%) (iii) directional audio prompt like a loud noise in that direction.
EMBODIMENT
18. Viewers were provided with the opportunity to embody body parts (e.g. control virtual hands) in only 10 titles, in 7 of these instances the body parts were rendered immobile/static.  There were nevertheless several cases of sensory deprivation or enhancement representation to enhance the narrative e.g. blurred vision
MOVEMENT
19. In 29% of the titles, the position of the viewer remains entirely static.  Viewers could teleport in only one title (Nefertari: Journey to Eternity, 2018).  In 40% of titles the viewer is placed in a moving object like a car, or boat.  First-person POV representation of movement is rare (11%).
20. In 66% of titles, the viewer is encouraged to rotate their body to see content that is placed in the rear 180 degrees of view at some point during the experience.
INTERPERSONAL SPACE
21. 85% of the titles included the presence of other people in the world.
22. Virtual actors were commonly spaced at the standard 'social' 1 metre apart indicating a close friend (27%), arms length (acquantances) less so (11%) and public distance over 5 metres even less (3%)
23. In 18% of titles the viewer is placed below eye level of the virtual actor.
TIME
24. Time is manipulated in 35% of the titles, with only 5% occuring in real time.  Speeding up time was common (22%), and there are also instances where viewers can experience multiple time speeds by moving around a scene e.g. Isle of Dogs: Behind the Scenes in Virtual Reality 2018.
RECOMMENDATIONS
The researchers recommend further exploration of spatial sound, embodiment and participatory observation. 

HIGHLIGHTS

- Hunger in L.A. (2012) used animated visuals to illustrate audio recordings of an incident in a queue for free food in L.A.
- Blindness into Darkness (2016) also married animated visuals with spoken word entries from a diary to evoke the experience of life without sight.
- Carne y Arena (2017) an ambitious room-scale installation, which includes several 'rooms' requiring the viewer to navigate through real space and interact with actors.



Bevan, C., Green, D.P., Farmer, H., Rose, M., Cater, K., Stanton Fraser, D. and Brown, H., 2019, May. Behind the Curtain of the" Ultimate Empathy Machine" On the Composition of Virtual Reality Nonfiction Experiences. In Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems (pp. 1-12).