actually, this isn't going to require cutting after all, that's not how i did it - i went along and pasted the sample a dozen times in a way that just mildly overlaps with the end of itself. so i don't have to worry about finding the perfect cut, i can just go along and find the place for the next sample by ear.

this should be fairly simple to construct very similarly, minus the improved sound quality.

and, in fact, i'm going to use cool edit for this part, as i think the blurry sound of the mix is an artefact of it. the song needs to be blurry.

i downsampled too, lol. i remember now that the noise was too intense, otherwise. it was a trade off between losing some of the subbass and softening some of the mid-high noise generators.

i may be able to mix the subbass back in, but one thing at a time....

i mean, now that the thing is downsampled, it will sound like 16bits if upsampled back to 32. so, i could then take the original file at 32, lowpass it and mix it in to bring some of the kick back.

that's a mixing stage thing, right now i'm reconstructing....

the fadeover happens to coincide with a bass beat. i'm almost certain that's what i did in the first place. so the syncing issues i was concerned about for this part are really non-existent.

the eyeing it out part has more to do with comparing the level in the waveform of the mp3 with the level in the reconstruct. how much did i amplify the fade? what percentage did i use for the paste over? i'm lucky that the possibilities are binary, as it allows for a finite number of possible choices. i can be confident it's extremely close to the original.