"Not at all like a roller coaster"?
I could see such being true for a final product, where a lower acceleration could be applied for a longer time. But here...
They went 395 meters (or possibly metres) in 15 seconds. Average speed is about 26 m/s, so peak speed is 52 m/s. (This assumes you accelerate halfway, then decelerate the other half, resulting in a consistent acceleration. Anything else will require a higher peak acceleration somewhere along the path.)
Anyway. To reach 52 m/s in 7.5 seconds, we're talking about roughly 7 m/s^2 acceleration, or about 70% of earth gravity. At startup, you'd suddenly feel an acceleration of sqrt(1^2 + .7^2) = about 1.2 gravities, and the direction of "down" would shift by atan(.7) = 35 degrees from vertical. After 7.5 seconds of that, it'll switch around (so that the direction of "down" will swerve back through the vertical to 35 degrees the other way, a change of 70 degrees total), still at 1.2 gravities. You definitely want to be sitting through all this, and I suspect you want to have not eaten much recently. Or at least want the person sitting next to you not to have eaten much recently. Or both.
I assume the acceleration was changed a little more smoothly than this (i.e., the "jerk" -- the time derivative of the acceleration -- was kept to a more reasonable level). Which would mean more acceleration. I don't see how you'd make this a stomach-settling experience.