Not entirely true...all these graphics that insist the two ends MUST be parallel are just dead wrong. The important thing is that the angles of the pinion and trans/transfer case output need to match. Up, down, left, right..doesn't matter.
However, and this is the big one, USUALLY, running them parallel is what is going to result in the easiest match within the packaging of an automotive chassis, and also often easier to figure out where to set these things.
This all ties in to the nature of how the u-joints work, in that they're constantly running through acceleration and deceleration while rotating. Having the angles match is how the cancellation works, and by having two different angles on the joints, they will no longer be able to cancel each other out.
For further proof that having both the pinion and transfer case pointing up can work, look at the way Rover did their stuff in the 70s on the full time 4WD rigs. Because the engine was tilted up, this meant that the front output on the transfer case was also tilted up. In addition, the axle pinion was tilted up as well. Then they assembled the driveshaft to be slightly out of phase, somewhere a bit less than 45* out. This all worked out to be pretty effective in cancelling out the vibrations on that shaft. It wasn't perfect, but did work out pretty well.
View attachment 253335
View attachment 253337
View attachment 253333
Realistically, though I've never tried it, I suspect that given the relatively short length of the intermediate shaft on our Samurais, as long as the angles are close to matching, vibrations will likely be minimal regardless of where it's set. I've never even checked the angles on mine, despite numerous drivetrain changes, and it's never been an issue (at least not on the intermediate shaft anyways...getting that stupid short rear driveshaft set up on a centered diff 14 bolt was a heck of a lot of not fun though, lol)