Yep, you have two possibilities here. The phase angle, and the balance weights.
A friend of mine builds driveshafts, and I have watched him balance them. He mounts it in the balancer. Hits a button, which locks the up and down motion of one end, then spins the shaft. The released end vibrates, and a strobe light shows the light spot. A readout tells how much weight is needed.
He has a big bucket of various scrap washers. He weighs different ones until he finds one the right weight. He tapes it on and tries again.
When its the right weight and in the right position, he spot welds it on and double checks the balance.
If its right, then he locks that end and respins it and balances the other end.
After that, he unlocks both ends and gives it one final spin.
So, in theory, each end is balanced. But you have to figure that even balanced this way, the weight on one end has to affect the balance of the whole shaft. And the shaft was balanced as one piece. So, when you turn the ends differently from the way they were balanced, it has to have some effect, just no idea how much.
The ideal situation would be to load up the shaft in a balance machine, check the balance and then change the phase and check again. But it would mean finding a driveshaft shop willing to do the experiment. Either for free or for a decent price.
I just setup the rear section like you would on a 1-piece shaft and left the intermediate section to fall where it ended up. I don't notice any vibrations but a 2-piece shaft with a CV is expensive.
