hello im sean: Hello again!
The capacitor foil in mind is a polymer dielectric with Aluminum cladding. You don't need both a 1.4um PET film
and capacitor foil. If you're planning to make your circuit via strips of a conductive material adhered to your film (think Aluminum leaf or voice coil wire) I'm picturing something that'd be extremely laborious. Those speaker guys can get away with this more easily because of the sheer size of the design and fact that they don't want too high of an impedance anyway. A headphone driver is much smaller so even to arrive at 8ohms like they aim to you'll have to increase your "trace length" by making each trace thinner and copying the serpentine design seen on some of these vintage ortho's. See
here. Can you imagine trying to glue voice coil wire into a design even half as complicated as that? It'd be "easier" in my mind to use capacitor foil and chemically etch your artwork into the Aluminum using a photoresist as an etch mask.
Magnetic force isn't that important, just try to get the flux evenly distributed and perfectly parallel with your circuit. The Wharfedale's use really weak magnets and like lots of electrical current to make up for it. You can, as many people before you have, get away with weakish ceramic ferrite if cost is an issue. Might be difficult finding them in the dimensions you want, though, which leaves flexible magnets which have a pretty low MGOe rating which then makes me question heat dissipation on that 1.4um film. Could just be me not having much knowledge about the subject, and I'm sure you could overcome all doubt with a thicker conductor, but then impedance becomes a topic for debate.
Spacing between .5 and 1.5mm sounds about ideal to me depending on magnet strength, diaphragm size and diaphragm material. The shorter the excursion of your diaphragm (usually the smaller it is the shorter its excursion will be) the closer you can place the magnets to it without foul play between the two. Unless using flexible magnets efficiency shouldn't be too huge of an issue so neither will the magnet spacing, although the closer you can get the magnets, obviously the higher the magnetic force will be and the easier they'll be to amplify, etc.