BRD idea: Songs with MP usage

[Shurrikhan]

Just going to repeat a suggestion I've thrown out occasionally since Stormblood release: use ramping MP-cost penalties for sustained use paired with activation costs to enforce rotation in a far softer way.

E.g.

• Give each song an initial cost of 200 MP and a further drain cost of 50 MP per second.
• For each second a given song is active, its per-second cost increases by 2.
• For each second the song is not active, its per-second cost decreases by 1, to a minimum of its original cost. I.e., if you spend twice as long off the song as on the song, its cost refreshes to its original, and therefore using a song for more than a third of the time comes at a net cost.
• Since you generate ~66.7 MP/sec and can start with as much as 10000 MP, you can therefore sustain a song for as much as 108 seconds at a time, or afford up to minimum and maximum average swap times of 25-90 seconds (still in even thirds) for 5 minutes (or can spend a more uneven portion of time on a single song), 30-75 for 7 minutes, 35-60 for 10 minutes, or per 45 seconds to rotate indefinitely.
• That is to say, you have considerable freedom in how short or long you play a song or in what proportions, with costs increasing the further you stray from a 45-second rotation or from equal use per song.

:: These costs can easily be shuffled around to entice use around a different maximum efficiency or push players nearer to the most-efficient pacing from the start, reduce the penalties of disproportionate use, etc. I just went with these since it changed less (descending towards our current [nominal] 45s durations, assuming equal thirds, etc.).

:: My own preference would offer a bit more leeway to disproportionate use by having song efficiency return to normal in less than twice its use time, as to better recover from mistakes and allow more sway, but would probably escalate a bit more sharply to bring the maximum time down to 90 seconds or so. I'll mock up some numbers to that end soon enough.

ASIDE: Another consideration is whether BRD should be able to sacrifice flexibility in song use for flexibility in direct support actions. Presently, I'm slightly leaning away from that, at least in any directly MP-consuming form, since I don't want Bard to feel too much conflict between enjoying the nuances of optimizing its own damage and providing situational support to its party, but I would very much like to see at-cost support options in other ways.

Why at cost? If you get a capacity for "free", what you're really doing is pre-paying for it, as you will then be balanced around using that utility. You can still come out ahead in fights that make better than average use of that utility, but you then lose out in fights that make worse than average use. Having ways to divert those expenses to more general fall-back options (much like Scholar can get at least some use out of its would-be excess healing capacity by spending it instead on Energy Drain) helps to steady "supportive" jobs' contribution while also adding some further depth to their kits as a side-effect.

To be clear, some "free" (or, more representatively, pre-paid) utility is fine, and what we have now isn't even the half of what we're taxed for so there's plenty of room to add on utility without fitting damage loss, but I would rather any powerful/impactful additions be more flexible and reduce the tax instead.