I’ve been thinking about the hands free interaction.
I think the head gestures is going to have to be the only option, with hardware (like the wireless keyboard) as an extra for personal viewing optimisations.
A vertical slider of some kind could be the good option.
That way, the user tilts their head up and down until some kind of reference image looks appeasing.
I think we have to assume the user will need some kind of button (magnetic or, like the Cardboard V2, touch) in order for configuration to be intuative.
One way to do the configuration WITHOUT a button would be to do the above, but allow the user to pan horizontally outside a panning threshold (illustrated by say a slider across the centre of the screen) this of course is going to require some technological investment and tweeking. It would also not be that intuative for the users? But just an idea.
Otherwise have you considered launching without the personal viewing optimisations features, or just simple UML buttons, for now?
Best