This is a follow on to an earlier discussion about pilot perception of G forces compared with actual aerodynamic loads on an airframe. Years ago, the Boeing 707 prototype famously did not one, but two 1G barrel rolls to impress potential buyers. I would imagine the expert pilot just kept the airframe at 1G positive load through the entire maneuver, which would put him weightless at the top of the roll (plane inverted). Is this what actually happened? How is this done?
You can't do a barrel roll at 1G. It's a loop pulled out into a coil, and you have to pull pitch to get it started so you are going to be 1.5 to 2 starting out, declining to 1G to .5 at the top, and pulling 1.5 to 2 on the pullout.
So the sensation at the start is like a hard climbing turn (which is what you're doing), pressed into your seat with about the force of a steep turn at the start, declining to normal gravity or bit less (giving a bit of that giddy becoming weightless sensation, but not coming out of your seat) as you ease off the pitch coming around the top, and about the same pressed into the seat force as at the start, or a little more, on the pullout.
A good barrel roll should be no more than 2Gs peak, but how much is a function of pilot skill and of the airplane's ability to keep up the energy through the first half. A jet with oodles of power to keep the speed up in the steep rolling climb can get away with a very gentle pull and will see only about 1.5 or less on the pull up and pull out.
But it's always going to be more than 1 starting out and finishing off and when Tex Johnston said in an interview it was a "1G maneuver", he really meant "not much more than" as in less than 2 (and no worse than a steep turn, unless you screw it up and fall out). Watch the film of the barrel roll and you see a firm pull up at the start with a fairly steep nose down attitude on the exit, which would have required a pull of at least 1.5 G in returning to level flight without over speeding.
By common terminology, when we speak of an "X G manuever" (say X=1 so a "1 G maneuver"), we mean that 1 G is the up-and-down acceleration indicated on the G-meter on the panel, and it also is the up-and-down acceleration (force per unit mass) acting to create stresses or strains on the aircraft's structure (and the pilot's body.) In other words, it is the "felt" acceleration or force per unit mass, or at least the up-and-down component thereof, in the pilot's reference frame. To get the ACTUAL up-and-down acceleration component (or force per unit mass) in the reference frame of the aircraft and pilot, we need to subtract the 1 G acceleration of gravity, or more precisely, the component of the the 1 G acceleration of gravity that acts in the up-and-down reference frame of the aircraft and pilot. Why? Because gravity exerts an equal acceleration (force per unit mass) on every molecule of the aircraft and pilot simultaneously and thus creates no stresses or strains within either, nor does it tend to accelerate the pilot toward or away from the aircraft and seat, so it is not a "felt" acceleration.
Two examples-- straight and level flight -- "felt" acceleration 1G, actual acceleration 0G. "Zero G" ballistic flight-- "felt" acceleration 0G, actual acceleration 1G downward (i.e. -1G).
So, in any maneuver that would commonly be described as a "1 G" maneuver, the G-force perceived by the pilot is +1 G, by definition.
Can a barrel roll be a "1 G" maneuver in this sense? Sure, as the aircraft goes over the very top of the roll. Can it be a "1G" maneuver in this sense from start to finish, including the initial pull-up and final round-out back to level flight? No it cannot. The barrel roll involves an arcing trajectory, and the G-load must be greater than 1-G during the initial pull-up and final round-out.
By the way, the "felt" acceleration as described above, is nothing more or less than the net aerodynamic force generated by the aircraft in whatever axis we are speaking of, or in all three axes if we want to be comprehensive. Normally we are most interested in the forces acting in the up-down direction in the aircraft's reference frame-- that's what the G-meter on the panel indicates-- and this is overwhelmingly dominated by the lift force generated by the wings. So when we speak of "feeling" 1 G (or "perceiving" 1 G), we are just saying that the wings are generating a net force equal to the mass of the (aircraft + contents) times the gravitational acceleration constant times two.
It may be sometimes a bit ambiguous as to whether the term "G-load" is being used to refer the TOTAL "felt" acceleration, or just the component that acts in the upward / downward direction in the pilot's reference frame. Either way, if we label a maneuver a "1 G maneuver", the logical conclusion is the the pilot is perceiving a 1-G acceleration-- i.e. the aircraft is exerting a 1 G force (per unit mass) on the pilot. Just as in straight and level flight.