Tugger, what Fireboat has said is pretty much spot on. Further to what he's saying, you might want to know that the majority of transmitters use a value of around 5000-Ohms for each gimbal pot. If this is the case with your Tx, then it's possible you might hit a few problems with the resistance values that Fireboat has shown. Nothing wrong with his schematics - merely a possible hiccup with the calculations, is all. You might find that your servos will glitch. It's probably because the timing sequence of the encoder pulses will be out of sync. But eyeballing the actual pulse on a 'scope shows that the majority of Tx gimbal pots have been pre-set to a value of approximately one-third total pot resistance. This puts the channel pulse at 1.5mS of the encoder's frame rate and corresponds with the servo being at neutral (centre) position. Fireboats' suggestion of halving the total pot resistance means you'd likely experience those out-of-sync glitches. If I'm not stepping on Fireboat's toes, might I suggest the values shown in the attachment. Being that a 2-position switch is used, it means the servo can be switched either fully CW or fully CCW.
You also mentioned that you need a switch with 'center-off' action. This can be done. It would result in, say, a servo being fully CCW with the switch in the 'up' position; servo neutral at 'center' position, and servo fully CW at switch 'down' position. But it needs a slight change to the way the resistors (and their values) are connected. You'll need a SPST 'center-off' toggle switch.
Fireboat, I'm still using dial-up. Well worth the wait to see them vids, though.
Nice work!