Good virtual reality creates a sense of presence - the feeling of being in the virtual world. But what causes presence to happen? We know that certain features of a virtual environment help generate presence, but other features can hamper or break it entirely.
Our experience will demonstrate how being embodied by an avatar can be a key element of successful immersive virtual reality. Our demonstration will build on recent academic work in psychology and neuroscience on how users understand and interact with avatars - both an avatar that represents themselves, and avatars representing other characters that respond to them.
A team from UCL's Virtual Environments group has designed this immersive experience to be compelling and engaging - but we also hope to gather data from it; we want to make a very large scale presence experiment. We will make public our findings on what worked well in this experience.
In our demo users can experience a highly immersive virtual environment, while taking part in an experiment on presence in virtual reality. Each user will be given the option of taking part in an experiment, or just seeing the demonstration.
Users will sit and put on the Gear VR. They will see a bar scene where a singer will perform. Each user will see one of a fixed number of versions of the scene, which are subtly different. We expect some of these versions to generate a higher sense of presence than others. We will gather some data from their behaviour in the scene, and we will also ask a questionnaire. Logs and questionnaire results will be uploaded to a server at UCL.
If you have already seen the application, or can't experience it, you can read about the science behind the study here.
A longer video showing how we used face and body motion capture to make the demonstration is available on our project web site.
We have received permission from our university's research ethics committee to conduct this study. Participation in the study is optional and the environment will work anyway if the user decides not to volunteer their data.