My research on this subject led to an inescapable conclusion: The only way to get the kind of efficiency desired is to build the alternator yourself.
There is basically no possibility of buying an 'off-the-shelf' alternator that doesn't make tradeoffs you're not interested in.
An alternator's efficiency can be improved in several ways, but none of them are very likely to have been found worthwhile for commercial uses. Using the strongest magnets possible is a big one (rare Earth magnets), designing the rotor and stator to be as close together as your methods permit is another, using three-phase power (more efficient generation) is a biggie... and biggest of all; if you design an alternator with a specific turning speed, friction load, and output in mind, you can get much further along than any 'savings' in time or effort that would accrue from buying your alternator 'as-is'.
I can see how you might think alternator design and building is a daunting task, but I assure you, it's worth it... and probably not as difficult as you might first think.
Let me put it this way: If you try everything else suggested and still see only minor improvements/ negligible output, then how reasonable would it sound to suddenly achieve what you wanted at the cost of 'merely' building an alternator? Put in a thousand man-hours with nothing to show for it and I'll wager that even a "herculean" 100-hour-effort will seem like nothing, if that's the price for success.
There are useful data on this subject available; let me know if you're interested.