How to Comp a Vocal Take
Published · updated · by Loopin
Comping isn't picking your single best take — it's building one flawless-sounding take out of the best line, phrase or word from several. Almost no one nails a whole verse in one pass, and no one needs to.
Record more takes than feels necessary
Three takes gives you almost nothing to comp from — you're stuck choosing between three imperfect versions instead of building one good one. Aim for 5-8 full takes of the same section before you start comping. Performance fatigue is real, but so is the fact that take 6 often has the one line take 2 was missing.
Don't stop the moment you think you have "the one." Record two more anyway. The best comp usually pulls from at least three different takes, not one take with a couple of punch-ins.
Comp by phrase, not by word
Cutting between takes mid-word almost always creates an audible seam — a change in tone, breath, or room sound that's obvious even if you can't name why. Cut at natural breath points and phrase boundaries instead, where a small timing or tone shift is far less noticeable.
If two takes have different energy and you need to combine them mid-phrase anyway, look for a consonant to cut on rather than a vowel. A hard consonant masks a splice far better than the sustained tone of a vowel does.
Listen for consistency, not just correctness
The instinct is to pick whichever take hit the pitch and rhythm most accurately, but a technically perfect line that doesn't match the emotional intensity of the lines around it will stick out more than a slightly imperfect one that flows. Comp for the story the vocal is telling across the whole section, not phrase-by-phrase perfection.
Play the comped section back start to finish, not one line at a time. Problems that are invisible in isolation — an energy dip, a tone shift — become obvious the moment you hear the whole thing in context.
Match room tone and mic distance across takes
If takes were recorded at different times or the vocalist moved relative to the mic, comping between them can introduce a subtle but audible shift in room sound or proximity. Before you comp, listen to the raw takes back to back for tone consistency — not performance, just sound.
If one take is noticeably more distant or has more room in it, either match it with EQ before comping or avoid pulling from it for anything except full phrases where the shift won't be as jarring.
Crossfade every edit, even the ones that sound clean
A hard cut between two takes can sound fine in isolation and still produce a click or a tiny discontinuity once it's sitting in a full mix at higher gain. A short crossfade — even just 5-10 milliseconds — at every edit point removes this risk for almost no cost to the performance.
Zoom in on the waveform at each edit point before finalizing. If the waveform doesn't line up smoothly at the crossfade, nudge the edit point a few milliseconds to a place where the two takes' phase and amplitude are closer together.
Frequently asked questions
How many vocal takes should I record before comping?
5-8 full takes of the same section is a good range. Three takes rarely gives enough material to build a genuinely strong comp -- you end up choosing between imperfect options instead of combining the best parts of several.
Where should I cut between vocal takes?
At natural breath points and phrase boundaries, not mid-word. If you must cut mid-phrase, cut on a hard consonant rather than a vowel -- consonants mask the seam far better than a sustained tone does.
Why does my comped vocal sound inconsistent even though every line is technically correct?
Technical accuracy and emotional consistency aren't the same thing. Play the whole comped section back start to finish rather than judging line by line -- energy dips and tone shifts that are invisible in isolation become obvious in context.