I always though that the end of the loop was the total number of samples in the BRSTM. It seems you actually put the "loop end" as what would be the total number of samples, plus the loopstart.
Like if the loop should have been 14336 to 500000 or something, you would have put 514336.
So, if I subtracted your loop start from the loop end you put, that's what it should be right?
So for Meta Ridley you have:
Start Loop Point: 315,392
End Loop Point: 2,166,241
So if I subtract 315,392 from 2,166,241 that is what the End Loop Point should be, right?