Finishing
Almost everyone reading this has a folder of unfinished eight-bar loops. Not because they lack talent or knowledge — every other page on this site is about knowledge — but because finishing is a separate skill from making, and nobody teaches it. You get better at production by completing tracks, and each abandoned loop teaches you almost nothing, because everything hard about music happens after the part that's fun. This page is about getting to the end.
Why Tracks Stall
Tracks rarely die because the idea was bad. They die for a small number of predictable reasons, and naming yours is most of the fix — because the remedies are completely different and applying the wrong one keeps you stuck.
| It stalled because… | What it feels like | What actually helps |
|---|---|---|
| The fun part ended | The loop is great; building a song around it feels like homework | Accept that this is normal and universal. Structure it anyway — the reward is on the other side |
| No destination | Endless tweaking, no sense of what "done" would even mean | Decide the length, the sections and the deadline before you continue |
| Perfectionism on the wrong layer | Two hours on a hi-hat sound; no second section exists | Ban mixing until the arrangement is complete end to end |
| Fatigue disguised as doubt | "This is bad" — after four hours on it | Stop. Judge tomorrow. Nothing sounds good at hour four |
| Comparison | It doesn't sound like the reference, so why bother | Compare your track to your last track, never to a mastered record |
| It genuinely isn't working | Calm, considered, still true after a break | Harvest the good part, archive the rest, start again. This is legitimate — just make sure it's this and not fatigue |
The most common of these by far is the first, and it's worth stating plainly: the loop being fun and the song being work is not a sign you're doing it wrong. It's the actual shape of the job. Producers who finish things aren't enjoying the boring part more than you — they've just stopped expecting it to be enjoyable.
Escaping Loop Purgatory
The eight-bar loop is where the overwhelming majority of unfinished music lives. It's seductive because a loop can be endlessly improved and never has to be judged as a song. The escape is mechanical, not inspirational — you don't wait to feel like arranging, you follow a procedure.
Duplicate to full length first
Before anything else, copy your loop across three minutes of timeline. It will sound terrible and repetitive. Good — now the problem is visible and specific: "make bar 48 different from bar 16" is a task you can do. "Write a song" is not.
Mark the sections empty
Drop markers for intro, verse, chorus, bridge, outro before you've written them. An empty labelled container is a much stronger prompt than a blank timeline — you're now filling gaps rather than inventing structure.
Subtract to make sections
The fastest way to get a second section is to delete things from the one you have. Mute the drums for eight bars and you have a breakdown. Strip to the bass and vocal and you have a verse. New sections don't require new material.
Write the ending early
An outro takes ten minutes and makes the track feel finishable rather than open-ended. Having a real end on the timeline changes how the middle feels to work on.
There's a broader principle here: work in passes, not in perfection. Get a bad version of the whole song before you get a good version of any part of it. A complete, rough, three-minute track can be improved anywhere; a flawless sixteen bars can only be extended, which is the hard direction. The Arrangement guide and the Songwriter workbench are both built for blocking out that rough whole quickly.
When You're Stuck in the Middle
Being stuck is usually an excess of options rather than a lack of them. Constraint is the reliable cure — every one of these works by removing choices, not by summoning inspiration.
| Move | What it does |
|---|---|
| Change one constraint | Change the tempo by 10 BPM, transpose the whole thing up a third, or swap the drum kit. Radical, cheap, reversible, and frequently the thing that unlocks it |
| Delete your favourite part | Often the element you're protecting is the one preventing the song from becoming anything else |
| Set a timer | Twenty minutes to finish the second verse, badly. Deadline pressure defeats perfectionism more reliably than willpower |
| Work on a different layer | Stuck on melody? Do the drum fills. Stuck on drums? Write the lyric. Progress anywhere restores momentum everywhere |
| Steal a structure | Take a song you love, map its section lengths with the Song Analyzer, and pour your material into that shape. Nobody will recognise the skeleton |
| Bounce it and go for a walk | Listening away from the screen, with no ability to fix anything, tells you what's actually missing in about ninety seconds |
Notice that none of these are "wait for inspiration." Inspiration is what starts tracks; it has essentially nothing to do with finishing them, and treating it as a prerequisite is how people end up with a hundred beginnings.
Deciding It's Done
A track is never finished in the sense of being unimprovable. There will always be another edit. So "done" can't be a quality you detect — it has to be a decision you make, and having criteria for that decision in advance is what stops a track from being polished into the ground.
A workable definition: it's done when
It plays start to finish
No placeholders, no gaps, no "I'll fix that later" markers. A real intro and a real ending.
Nothing is obviously wrong
Not "everything is perfect" — just no clicks, no clipping, nothing out of tune, nothing badly out of balance. The absence of errors, not the presence of excellence.
It survives elsewhere
You've heard it on a phone, in a car, and on headphones, and it holds up. This catches almost everything that matters and almost nothing that doesn't.
Your changes stopped mattering
When you can't tell whether the last three edits improved it, you finished a while ago and have been rearranging deck chairs since.
Two practices make this decision much easier. Bounce a version and date it whenever the track reaches a coherent state — it removes the fear of losing the good version, which is a surprising amount of what keeps people tweaking. And let it sit for a week before the final call. Distance does something no amount of listening can: it tells you whether the track is actually good, or whether you'd just heard it two hundred times.
It also helps to remember what "done" is for. A finished track that's 70% as good as you wanted teaches you more than an unfinished one that was heading for 95%, because everything you'd learn is in the last stretch. Finish it, learn the lesson, apply it to the next one. That loop is the entire mechanism by which anyone gets good at this.
Getting Feedback That Helps
"What do you think?" reliably produces "yeah, it's cool," which is worth nothing. Useful feedback has to be asked for specifically, and the person giving it usually needs to be told what kind of information you're after.
| Ask this | You learn |
|---|---|
| "Where did you get bored?" | Arrangement problems, with a timestamp. The single most useful question you can ask |
| "What's the first thing you noticed?" | Whether your intended focus is actually the focus |
| "Could you hear all the words?" | A vocal balance answer you can't get yourself, because you already know the lyrics |
| "Does anything sound wrong or annoying?" | Harshness and resonances your ears have adapted to over hours |
| "Would you play this twice?" | The only quality question that means anything |
And a rule for receiving it: when someone tells you something is wrong, they're almost always right; when they tell you how to fix it, they're often wrong. Listeners are excellent detectors and poor diagnosticians. "The chorus doesn't hit" is real data. "You should add more reverb" is a guess — but the feeling that produced it isn't, and that feeling is what you go and solve.
Building the Habit
The producers who get good are not the ones with the best ideas. They're the ones with the most finished tracks, because the number of complete songs you've made is the closest thing to a measure of experience there is. Everything else is a proxy for it.
Deliberately finish bad ones
Take a mediocre idea all the way to a bounce on purpose. It costs a day and it teaches the entire second half of the process — the half you keep skipping. Nobody has to hear it.
Constrain the scope
One-day tracks. Five-instrument tracks. Two-minute tracks. Constraint makes finishing achievable, and a finished small thing beats an abandoned ambitious one every single time.
Separate the modes
Writing, arranging and mixing are different jobs using different parts of your brain. Doing them at once is why sessions stall. One mode per session, and don't let mixing start until arranging is done.
Keep a graveyard, and raid it
Don't delete abandoned ideas — file them. A loop that went nowhere on its own is often exactly the bridge a different track needs six months later.
One last reframe. Every finished track you make is a data point about what you actually sound like — and you can't develop a voice from fragments, because a voice is largely made of the decisions you only face at the end: how you end things, what you leave out, how long you let a moment sit. Those choices are the ones that make a producer recognisable, and you never get to make any of them in an eight-bar loop.
Copy the loop across three minutes. Mark the sections before you write them. Make sections by taking things away. Set a deadline, ban mixing until the arrangement exists, and bounce it even if it's a 70%. The next one starts better because this one ended.