Annotating Recorded Telemetry for Extracting Meaning and Insight from Scenarios in Virtual Reality (ARTEMIS-VR)

Producers of screen fiction face a daunting challenge in capturing the attention of audiences. The current gold standard used by industry CEOs and film content producers to understanding if a film or TV pilot will be commercially viable is audience testing. The structure of audience testing has changed little since it was first used in 1928\. A target audience is invited to preview content in a theatre. Afterward, the audience fills out questionnaires rating the show or film they have seen. A smaller focus group of twenty or thirty is then asked to stay and discuss their reactions in more detail, with questioning focused primarily upon ascertaining an overall rating (e.g., "What was your reaction to the movie overall?") and generalised aspects of the film (e.g. "list which scenes or moments of the movie you liked the most/least?"). There are a number of limitations with this approach:

* Since audience testing was conceived, our understanding of how emotional responses and individual differences shape how we evaluate experiences has advanced
* Audience testing was never designed for fully immersive experiences. It remains unknown how well existing audience testing methodologies can explain how audiences will consume 3D content
* Target audiences are selected on the basis of superficial characteristics (age, ethnicity) instead of the self-relevant traits that better determine interest and engagement
* It is prohibitively expensive for productions operating on a shoe-string budget

To address these needs we are running a feasibility study to test a new approach to audience testing. This will involve content creators labeling onscreen events for their desired emotional response and then testing if audiences show the desired response. To achieve the aims of this feasibility study we will:

* Investigate what features of 2D and 3D experiences content creators would like feedback on
* Develop a prototype system of SimLabVR for demonstration purposes sufficient to show proof of principle
* Run a pilot study to test if it is possible to identify a target audience using emotional responses alone, without reliance on superficial characteristics.