Please be more careful with your optimization. There are a lot of minor issues that add up:
-You have lots of poorly optimized rests. [r1]X should be [r2]X instead, and make sure that it's at least [r2]8 before looping them.
-The cymbals can probably be label looped
-Your macros are redundant. Something like K8K16K8 redefines the instrument and q command each time, which is not necessary and bloats the port's size. Do K8c16c8 instead
I quickly fixed the last point about percussion macros, and that alone brought the size down from 0x4D0 to 0x471, and I'm sure you can save more by fixing the rest.
Also, I lowered the global volume down from w140 to w110 since it was a bit loud in game. For future reference, and this isn't a moderation issue or anything, but something that I've found helpful, I would suggest using a higher global volume in combination with lower v values instead, since it gives you more room to make things louder, without having to rely on $FA $03 $XX.