Posted by Andy Brown on
Back in March we talked about the challenges of designing subtitle behaviours for 360° video content. We explained some of the challenges that arise and described four behaviours that we had implemented for testing. We have now completed the testing and analysed the results, and it's fair to say that we have been slightly surprised. So, what did we find?
As you might expect, there were significant individual differences, with one person's favourite being hated by someone else. However, across the body of participants, some general themes emerged. Taking each behaviour in turn:
(Subtitles are placed into the scene in three fixed positions, equally spaced by 120° around the video and slightly below the eye line.)
This behaviour wasn't well received - only one participant chose it as their favourite. The problem was that people found it hard to follow both the subtitles and the video; having to continuously switch from one to the other was difficult. Despite the fixed position (in the scene) of the subtitles, people wasted time and effort locating them and felt that they missed out on content.
Follow Head Immediately
(The subtitles are presented as a 'head-up display' always in front of you, and slightly below straight ahead. As you turn your head, the subtitle moves with you, always at the same location in the headset display.)
This was, contrary to our expectations, the most popular behaviour. It is a very simple behaviour to understand, and the subtitles were always available, no matter how much you explored the scene, and very easy to locate and read. Our concerns about the subtitles obstructing the video content were justified, and this was the biggest drawback of this behaviour. Feedback from our participants suggests that we need to lower the subtitles within the field of view. Our other concern with this behaviour was that it might induce 'VR sickness'; this was barely a problem (only one or two of our participants commented on it feeling slightly odd), although we need to test longer pieces of content to be sure.
Follow Head With Lag
(The subtitles follow your gaze around, but only for larger head movements: if you look slightly left or right it stays in place, but looking further will cause the subtitle to catch up with your head orientation.)
We thought that this would be the most popular behaviour, as the subtitles would allow exploration and be easy to locate, but it wouldn't feel as if they were 'stuck to your face'! In practice, although it was the second most popular behaviour, this one proved a bit too complex - the action of the subtitle 'catching-up' with the larger head movements was unpredictable, and this distracted participants and reduced their feeling of immersion.
Appear In Front, Then Fixed
(Each subtitle is placed in the scene in the direction you are looking at the time when it appears and remains fixed in that location in the scene until it disappears.)
Although this worked in some scenes, in others it left people confused and unable to follow the subtitles. The essential problem was that when people are having a quick look around a scene, the subtitles get placed part way through a head movement, and these positions are unpredictable to the viewer and not easy to relocate. The end result was that people either missed out on the subtitles and looked around the scene, or read the subtitles and missed out on looking around.
Overall, then, the simple obvious solution is the one that appears to work the best. Perhaps, as engineers we have a tendency to over-think problems, but this has certainly been a good reminder of the value of user testing.
What are our next steps? Clearly the main outstanding question is how do these behaviours stand up to prolonged viewing? Are preferences different for content that lasts 10 minutes or longer? Is the simple HUD behaviour still preferred, or does it cause discomfort or sickness. If people do find it comfortable and usable, then we have a solution, if not, we have more work to do.
More on Virtual Reality and 360° Video:
This post is part of the Future Experience Technologies section