now, something about the dense part:
there are some unpleasant anchors (reporting only big ones, 5 notes and more)
00:34:582 (34582|2,34714|2,34846|2,34978|2,35109|2) -
00:34:648 (34648|4,34780|4,34912|4,35109|4,35241|4,35373|4,35505|4,35636|4) -
00:36:295 (36295|5,36427|5,36559|5,36690|5,36822|5) -
00:37:613 (37613|0,37744|0,37876|0,38008|0,38140|0) -
00:39:589 (39589|4,39721|4,39852|4,39984|4,40116|4,40248|4) -
00:41:829 (41829|0,41960|0,42092|0,42224|0,42356|0) -
00:15:917 (15917|5,15983|6,16005|3) - the snaps here are incorrect, also I believe there's one additional note (ghost note)
00:09:285 (9285|2,9285|1,9351|4,9351|3,9417|6,9417|5,9417|0) - I feel like a chordtrill would feel much better here
the snaps are still incorrect and I believe this can be resolved after closer audio analyzing.
In order to figure out the snaps better, this section should be slowed down even further - I provided the audio files in your Discord DMs.
After analyzing the audio files, you can distinguish two parts, the first section and the ending.
The first section uses the same snap compared to the ending part.
In the ending part, there are 2 "steady" signals, instead of many of them, just like in the first part.
My suggestion is to express the first part with 1/8 snap, and the other one with 1/4 I believe
I also sent you a screenshot of (I believe) correct snaps), not a suggested pattern.
now that I look at it again, the patterns aren't difficult and combined with this density, it's not that huge deal
changed 00:26:970 (26970|2,26970|4,26970|0) - into double [13] and moved note in 00:27:144 - to 4 column