Don't worry you are not the only one confused.
With 12 poles of the 2 x 1 x 1/2 magnets on a nominal 12" disc you have an alternator that is pretty stiff for 8ft and a bit on the light side for 10 ft. You can play changes with the cut in speed and tsr to get what you want to some extent, there is no perfect winding that will out perform anything else.
At 8ft with 24v and short thick leads you may have to increase air gap and possibly add a bit of line resistance if you cut in at 7 mph. The 12v case usually keeps out of stall with more diode drop and line loss.
At 12v you may have a tough time holding things down fairly well with a 10 ft prop, but you should make it with 24v as long as you keep the air gap tight.
You will most likely have a similar peak output power from 10ft as from 8ft as the temperature will be the limit.
The 10ft machine will gain by reason of its swept area and will produce more power at all wind speeds up to the thermal limit ( furl earlier for 10ft).
When you have your magnet rotors sorted out you can play with a test coil, your winding ability will have a significant effect on the size of wire you can squeeze in for a given number of turns and your moulding ability will affect the air gap that you can safely work with.
When you are satisfied what wire you can squeeze in then you can select your cut in speed to suit your wind area. If you are prepared to forgo the lower winds you can use less turns of thicker wire and achieve better efficiency in the 12 mph upward winds. You will find that if you do this you will be running faster than the normal stalled state that you reach with a low cut in speed.
For 10 ft in a good wind area I would aim for a cut in speed of 180 rpm at 24v and use the thickest wire you can while maintaining a safe air gap. If you get sloppy with air gaps at 10 ft you will start running into high speeds at full load.
Not sure if this has helped but there are really no exact performance details for any of these machines as far as I know. Constructional ability, magnet grade, line loss are all factors that are a bit unknown, and I suspect that individual interpretation of the blade profile may have as much effect as all the other factors and you haven't even considered that.
Flux