[pedrocr]

I understand 1G as also requiring the vector to be pointing down on the aircraft. If you allow it to point in any direction as long as the vector has length 1 passengers will notice and not all mechanical parts will be guaranteed to work so the fact that it's technically 1G is not enough to guarantee any airplane can do it. I think that's not the usual definition of 1G either as we even talk about negative Gs (and airframes are particularly sensitive to them) which wouldn't make sense in that case.

But it would be cool to see if there's a theoretical path that we can recognize as a barrel roll than keeps exactly 1G of acceleration in any direction needed.