Short answer: headroom is the space between the loudest peak in your audio and the maximum level before it clips. In digital audio that ceiling is 0 dBFS, so if your loudest peak sits at -6 dBFS you have 6 dB of headroom above it. Leaving some room is what keeps your peaks clean and gives the later stages of the chain somewhere to work.
The numbers below are common starting points rather than strict rules, and a few dB either way is fine. Levels are written in dBFS, decibels below the point where digital audio clips.
Headroom is the space between the loudest peak in your audio and the maximum level the signal can reach before it clips. It is the room you leave at the top, measured in decibels below the ceiling. If your loudest moment peaks at -6 dBFS, you have 6 dB of headroom before the signal runs out of room.
In digital audio that ceiling is a fixed point called 0 dBFS, the loudest a sample can be. There is nothing above it. A signal cannot go past 0 dBFS the way an analog level can push into the red on a meter, so anything that tries to gets flattened at the top, which is clipping. Headroom is simply how far your peaks stay below that hard limit.
It is worth being clear that headroom is about peaks, not loudness. Two masters can feel about as loud as each other while one leaves plenty of headroom and the other has almost none, because loudness is an average over time and headroom is measured from the single highest peak. When you check your headroom you are reading the peak meter, the highest instant the signal reaches, not how loud the track feels.
dBFS stands for decibels relative to full scale, and 0 dBFS is the ceiling in digital audio, the loudest a sample can be. Go over it and the signal clips, a hard and ugly distortion you almost never want. Every level in this guide is written as a number below that ceiling, so a peak at -6 dBFS is 6 dB of headroom left before you hit it.
Headroom matters for two reasons: it stops your peaks running into the ceiling, where they clip into harsh distortion, and it leaves room for the processing and the mastering stage that still have to happen. Both come down to the same thing, giving the signal somewhere to go instead of pinning it against the top.
The first reason is clipping. When a peak crosses 0 dBFS the top of the waveform is cut off flat, and that produces a hard, cheap distortion, nothing like the warm saturation people reach for on purpose. It tends to sound harsh and crunchy, and once it is baked into a bounced file there is no clean way to take it back out. Leaving headroom means your loudest peaks have somewhere to sit below the ceiling, so they never get clipped in the first place.
The second reason is room to work. Almost everything you do after setting a level can add to it. An EQ boost lifts a band, a compressor with makeup gain pushes the whole signal up, and parts stack and sum on a bus so the total sits higher than any single part. If your signal is already touching the ceiling, there is nowhere for that added level to go. Leave headroom and each of those stages has space to do its job without clipping.
Mastering needs that space most of all. If you plan to master the track yourself, or hand it to someone who will, the master needs a few dB of headroom so the mastering chain has room to add level and shape the sound. A mix delivered already slammed against 0 dBFS gives the next stage nothing to work with. Keeping headroom at every step is really the whole point of gain staging, the habit of setting sensible levels all the way down the signal path.
Headroom is not wasted space, it is working space. Peaks need room below the ceiling so they stay clean, and every stage after them, EQ, compression, bus summing and the master chain, needs room to add level without clipping. A mix with no headroom cannot be worked on cleanly. Leaving a few dB is what keeps the door open for the rest of the chain.
A common starting point is to keep your mix-bus peaks a few dB below 0, often cited around -6 to -3 dBFS, and to hand off the master a few dB down so the mastering stage has room to work. These are approximate targets rather than exact marks, and a few dB either way is fine. The point is room above the peaks, not a precise number.
| Stage | A sensible level | Why it helps |
|---|---|---|
| A single track | peaks around -12 to -6 dBFS | Leaves headroom at the start, so parts are not already hot before they stack. |
| The mix bus, while mixing | peaks a few dB below 0, often around -6 to -3 dBFS | Keeps the summed mix clear of the ceiling with room to spare. |
| The master, handed off for mastering | peaks a few dB below 0 dBFS | Gives the mastering stage space to add level and shape the sound without clipping. |
| True peak on a finished master | at or below -1 dBTP | Leaves about 1 dB so lossy encoding does not push a peak into clipping. |
That last row is about true peak, which is a slightly stricter measure than the peak meter you watch while mixing. True peak accounts for the way a signal can rise higher between samples than any single sample shows, so a track that looks like it just touches 0 dBFS can actually push past it once it is turned into a streaming file. Holding true peak at or below -1 dBTP on a finished master leaves a little headroom to absorb that. The numbers for the final stage are covered in our guide on how loud a master should be.
Do not chase these figures with your eyes at the expense of your ears. They exist to keep you out of trouble, not to be hit exactly. If your mix bus peaks at -4 or -8 instead of -5, nothing is wrong. What matters is that the loud parts have room above them and nothing is pinned against the ceiling on the way to the master.
Headroom is the space at the top, between your loudest peak and the ceiling at 0 dBFS. The noise floor is the opposite end, the quiet hiss and background at the very bottom, below which nothing useful sits. Headroom is how much room you have before the signal is too loud, and the noise floor is how quiet the signal can get before it disappears into noise.
The gap between those two ends is your usable dynamic range, the distance from the quietest sound that stays clear of the noise up to the loudest peak that stays clear of the ceiling. Headroom protects the top of that range and the noise floor sits at the bottom of it. In the analog days you had to mind both, because pushing a signal too low buried it in tape hiss and circuit noise while pushing too high clipped it.
Modern digital audio changes the balance. At 24-bit the noise floor sits so far down that a quiet track carries essentially the same detail as a loud one, and you will almost never hear it, so in practice the end you actively manage is the top. That is why headroom, not the noise floor, is the thing you keep an eye on when you set levels. You leave room at the top on purpose, and you let the very low noise floor take care of itself.
Think of your signal as living in a band with a hard lid and a soft carpet. The lid is 0 dBFS, and headroom is how far your peaks stay below it. The carpet is the noise floor, the quiet level below which sound is lost in hiss. Your music sits between them, and in digital there is so much room down to the noise floor that the top is the end worth watching.
Headroom is a mixing habit. It only matters once you have parts worth mixing, and writing and recording those is the harder job. The numbers on this page are simple to set; having something clean to point them at is the work.
The Collection is our three instruments together, ARGISH, SILT and REHEAT, with three separate licence keys. This page is about keeping room at the top of your signal, and that job starts easier when the sound going in is already clean: our instruments output at sensible levels with clean peaks, so your project keeps its headroom instead of arriving already slammed against the ceiling. ARGISH is a self-playing drone synth, SILT is a tape-loop instrument, and REHEAT writes acid lines, and whatever you make with them is yours to release. Each one ships as AU, VST3, AAX and a standalone app on macOS, signed and notarized, and a VST3 on Windows.
You can hear one free in your browser first, no install and no account, then decide.
We write these when there is something worth writing down. One email when a new one lands or a new Tunary instrument ships. No newsletter, no schedule.
Tunary