Recording vocals at home: mic choice, room, gain and comping

Most home vocals that sound "amateur" are not let down by the singer or the microphone price. They are let down by a reflective room, a mic placed by habit, a level set too hot, and a final take that was really just the last one recorded. This guide takes those four in order. Everything here is sourced to the manufacturer documentation or the magazine article we read on 24 Oct 2026, and where a common rule of thumb has no source behind it, we say so.

If you have not treated your room at all yet, read our guide to home studio acoustic treatment on a budget alongside this one. It covers the bass traps and panels; this piece is about the few square metres around the singer.

Which microphone should I use for vocals at home?

There are two common families: dynamic mics and condenser (capacitor) mics. Shure describes its SM58 dynamic as having a cardioid pattern that "isolates the main sound source while minimizing unwanted background noise". We found no source that names one type as always better for home vocals, so treat it as a fit question. Our view: a sensitive condenser earns its place when the space around it is quiet and dead, and a close-worked dynamic is more forgiving when it is not.

Whatever you choose, most vocal mics are directional, and that brings proximity effect: the closer the source, the more bass. Shure's SM58 user guide puts a number on it, saying cardioid microphones "progressively boost bass frequencies by 6 to 10 dB below 100 Hz" at about 6 mm (1/4 in.) from the source, and that this "can be used to create a warmer, more powerful sound". The SM58 itself is voiced with a "bass roll off to control proximity effect", according to the same guide.

So distance is a tone control. The Shure guide's placement table sets out the trade-off:

The table is written for one dynamic mic, so read it as a guide to the direction of change, not exact numbers for your mic. The same guide also advises placing microphones "as far as possible from reflective surfaces", which leads to the room.

How do I stop my room sounding on the vocal?

Close miking reduces room sound, but it has costs. Julian Rodgers, writing in Sound On Sound (Mar 2026), notes that working very close to a capacitor mic brings "over-prominent breaths, lip and mouth noises and excessive proximity effect". The better fix is to deaden the space around the singer so you can step back a little.

From the same article:

• Blankets work, heavier ones better. Duvets give some absorption, but heavier packing blankets usually do better, and purpose-made acoustic blankets are more predictable.
• Thicker is better for low end. For panels and portable booths, Rodgers writes that "thicker panels invariably provide better low-frequency control".
• Reflection filters have limits. A curved filter behind the mic shields it from some reflections, but avoid "placing the microphone too far back inside the curve of the filter", which colours the sound.

A practical order: hang something heavy behind the singer first (that is what a cardioid mic faces towards), then add absorption to the sides, then think about a filter.

How loud should I record vocals?

Lower than most people think. Matt Houghton answers this directly in Sound On Sound (Apr 2026). When recording, he writes, the goal is "a healthy signal (relative to the noise floor) without clipping your interface's converters". Peaks around -6 dBFS will do that, but "6dB isn't a huge amount of headroom", and "as long as you're recording through 24-bit conversion (never mind 32-bit), you don't need to sail so close to clipping".

The often-quoted -18 dBFS figure, he explains, "would refer to an average (RMS) level, rather than the peak", leaving almost as much headroom as professional analogue gear.

In practice (steps 3 and 4 are our own practice, not from the sources above):

1. Record at 24-bit (or higher, if your interface offers it).
2. Have the singer perform the loudest part of the song, at full commitment, while you set the preamp gain.
3. Set gain so those loudest moments still sit comfortably below 0 dBFS, with the average well below that. In our experience a singer's level can creep up over a session, so we leave extra room.
4. Set the gain on the interface's preamp, which is what sets the recorded level, rather than reaching for the DAW fader.

Our view: it is safer to record a little quiet than to clip, because a quiet 24-bit take can simply be turned up.

What is comping, and how should I do it?

Apple's Logic Pro guide describes the scenario plainly: you record several performances over the same section, then "select the best parts from the different performances and piece them together into a single take". That composite take is a comp. In Logic, takes stack into a take folder and you build the comp with Quick Swipe Comping. Other DAWs have their own comping tools; check your DAW's manual.

Our own workflow (this is our practice, not something the sources above prescribe):

1. Record full passes first. We record several complete takes before fixing individual lines, to keep the performance's arc.
2. Then punch in on problem lines rather than recording words one at a time.
3. Listen line by line, choose by feel first. We pick for emotion and delivery before pitch.
4. Keep the comp in context. Check each choice against the music, not soloed.
5. Keep the unused takes until the song is finished. You may want a different line once the mix takes shape.

What we think

The home vocal is the one part of a record where an independent artist can match a professional studio, because the thing that matters most, the performance, is already in the room. That is a reason to spend effort where the evidence points, not where the marketing does: a heavy blanket behind the singer and sensible gain cost almost nothing and fix more than a mic upgrade usually will.

We are also wary of over-comping. A comp should collect the singer's best moments, not invent a performance they never gave. If you need fifteen fragments per line, the honest move is usually another full take after a break. Listeners respond to breath, phrasing and small imperfections; those are the parts of a voice that belong to the artist, and they are worth protecting. On that note, if your recorded voice is going out into the world, our guide on protecting your voice from AI cloning is worth a read. When you move on to the mix, see common bedroom mix mistakes.

How we tested

This is a sourced guide, not a hands-on test. We read Shure's SM58 user guide, Apple's Logic Pro comping documentation and two Sound On Sound articles, and kept the factual advice those sources state; the steps marked as our practice or view are our own recommendations. We cut common figures we could not trace to a source.

FAQ

Do I need a condenser mic for vocals? No source we read requires one. In our view, in an untreated room a directional dynamic mic used close can be the more forgiving option.

How close should I sing to the mic? Closer gives more bass and isolation, further gives a more natural sound. Shure's guide puts the line at about 15 cm for its SM58.

What level should my vocal peak at? Sound On Sound's advice is to avoid clipping with room to spare; at 24-bit you don't need to push peaks near 0 dBFS, and -18 dBFS is an average-level reference, not a peak target.

Is comping cheating? It is standard practice, documented in mainstream DAWs. Keep it to choosing among real takes, and keep the performance's feel.

Sources

• Shure, SM58 User Guide
• Apple, Comping overview, Logic Pro User Guide for Mac
• Matt Houghton, Sound On Sound, Q. Should I gain-stage to -6dBFS or -18dBFS?
• Julian Rodgers, Sound On Sound, Better Vocals In The Home Studio

Updated 24 Oct 2026.