This port starts from a very interesting idea and does some neat things with it. In particular, I love the idea of building upon a simple song and making it more complex. The pitch bends are added in all the right places and make the song funkier, and the @13 in the background gives a nice atmospheric feeling. Ultimately, however, there are a few issues that prevent me from accepting it.
First, the biggest thing I notice is the lack of optimization. There are a lot of patterns in the .txt that could easily be repeated, so I'd go over those once again. The loops you've currently put are a start, but there's still more work that needs to be done. Don't forget about using label loops.
As well, there are a few other things, such as:
-Long rests that aren't looped, such as at the beginning of channel #3. As well, it's slightly more efficient to loop rests with r2 instead of r1.
-Redundant instrument definitions, such as with @1 on channel #1.
However, there's a caveat to that last point; for vanilla percussion instruments (@[email protected]), it's actually preferable to define the instrument for each note, as otherwise, they detune. For those instruments, redefining the instrument doesn't take up more space, so don't be shy to do so. For example, @21<<cc>> would be better as <<@21c @21c>>
Now, moving on, the mixing and instrumentation aren't as great as they could be. For example, the @6 at the beginning on channels #1 and #2 are a bit loud and drown out the rest of the sound, making that section a little bit muddy. As well, if you're striving for accuracy, then @4 would be a closer sample to the original there; I also hear a few incorrect notes in the end part of that section. For the rest, the echo feels a little weaker than the original, which, along with the instrument choices, make the port feel a little emptier.
As for instrumentation, overall, I feel like the instruments you used don't blend in together very well, especially the low octave @2 and @4, @17, and @1. There's a workaround though, if you use ADSR with your instruments, it gives you more flexibility with how each of them sounds, and could lead you to some interesting results. There is no single answer on how to do this, so I'd suggest you experiment with ADSR. Sometimes you'll even get happy accidents.
Overall, maybe using the SMB2 Underground port that's in the SMAS soundtrack as a reference could help you get a better feel for the original's sound design and allow you to work off of that.
Next, the @0 and @13 are a really nice touch and help make the port feel really full; however, placing them on channels #6 and #7 is not the best idea, since those channels will cut out anytime any sound effect plays, and therefore noticeably cut out the sub-melody. Something like the @17 echo is more sacrificeable, for example.
Finally, I don't want to be too picky about the composition itself since that's a matter of personal taste, but I'd still like to point out some parts that could be improved in my opinion. These aren't removal reasons per se. Some of the sections, such as from 0:21-0:32, 0:54-1:06, and 1:12-1:21 sound a little bit too chaotic. This might sort itself out if you adjust the instrumentation, but it's still something I'd keep an eye out for after you do that. The first section I mentioned feels messy because of the percussion, while the third section I mentioned, feels chaotic because of the rapidly-changing instruments.
One last organizational note, make sure you include SPC tags with your submission to give your SPC a game name, etc. Information on the #spc tag can be found in the AddmusicK readme. Similarly, when titling your submission, it'd be preferable to include the full name of the original song, with the game of origin, and then follow it up with an indication that it's a remix, like this: "Super Mario All-Stars - SMB2 Underground (Recomposed)." Similarly, the "Source" in this case would be Remix, not Original.