I remember the first time I ruined a vocal take because the input meter was red the entire time. I thought louder meant better. It does not. Clipping and noise floor are the two extremes every recorder has to manage, and learning how to set recording levels is the skill that sits between them. In this guide I will walk you through the exact workflow I use on every session to keep signals clean, loud enough to mix later, and free of distortion.
If you have ever asked why your recordings sound harsh, thin, or hissy even when you bought an expensive microphone, the answer is almost always in the gain stage. By the end of this article you will know the recommended dB targets, how to set levels on any audio interface, and what to do when a take goes wrong.
Table of Contents
What Is Audio Clipping and Why Does It Ruin Recordings?
Audio clipping happens when your signal tries to exceed the maximum level your system can handle. In a digital system that ceiling is 0 dBFS, which stands for decibels relative to full scale. Once a waveform tries to push past 0 dBFS, the top of the wave gets sliced off flat, producing the harsh, buzzy distortion that engineers call clipping.
I think of the digital ceiling like the roof of a room. Sound can bounce around freely underneath, but the moment a peak smacks into the ceiling, data is lost forever. That is why clipped audio cannot be fully repaired by turning the volume down later. The information simply is not in the file.
Clipping usually occurs at one of three points in your signal chain: the microphone preamp, the analog-to-digital converter inside your audio interface, or a plugin inside your DAW that does not have enough headroom. Finding the weakest link is part of learning how to set recording levels properly.
How clipping shows up on a meter
On a typical level meter, green means safe, yellow means hot, and red means you have hit or exceeded 0 dBFS. If your peaks touch the red even for a millisecond, distortion is already baked into the recording. Always watch the peak meter, not just the average level, because clipping is a peak event.
What Is the Noise Floor and How It Affects Your Recordings?
The noise floor is the low-level hiss and rumble that exists in every recording environment and every piece of gear. It comes from microphone self-noise, preamp electronics, room air conditioners, computer fans, and electromagnetic interference. The noise floor sets the quietest sound your recording can capture cleanly.
If you record too quietly to avoid clipping, the noise floor creeps up next to your signal. When you later boost the track to a usable level, you boost the hiss with it. That is why setting levels too low is just as bad as setting them too high. The goal is to find the place where your signal sits well above the noise but still leaves headroom below clipping.
The distance between your signal and the noise floor is called the signal-to-noise ratio. A healthy signal-to-noise ratio for vocals is around 60 to 70 dB, which is why most engineers aim for peaks near -6 dBFS on the loudest moments.
Recommended Recording Levels: The -12 dB to -6 dB Sweet Spot
After years of testing, the audio community has settled on a reliable target. Aim for your loudest peaks to land between -12 dBFS and -6 dBFS, with the body of the signal hovering around -18 dBFS. This range gives you clean headroom on the top end and keeps the signal well above the noise floor on the bottom end.
When I record vocals, I usually set peaks to land around -6 dBFS on the loudest sustained note. For acoustic guitar, I aim closer to -10 dBFS because the instrument has a much wider dynamic range. Podcasts and voice-over work tend to land near -12 dBFS because the source is more controlled.
Quick reference for common sources
Lead vocals: peaks at -6 dBFS, average around -18 dBFS
Acoustic guitar: peaks at -10 dBFS, average around -20 dBFS
Electric guitar (mic’d cabinet): peaks at -8 dBFS to -6 dBFS
Drum overheads: peaks at -10 dBFS, average around -22 dBFS
Podcast voice: peaks at -12 dBFS, average around -20 dBFS
These are starting points, not rules. Adjust based on how dynamic your source is and how much processing you plan to add later.
How to Set Recording Levels Step by Step?
Here is the exact workflow I run before every recording session. It takes about three minutes and prevents 90 percent of the level problems I used to deal with.
Step 1: Set up your monitoring chain
Open your DAW, create a new track, and route the input to the correct channel on your audio interface. Make sure the track’s input monitor or direct monitoring is on so you can hear what the microphone is capturing. Lower the fader on the track all the way down for now. You are setting input gain, not track volume.
Step 2: Perform a loudest-pass test
Ask the performer to sing, play, or speak the loudest part of the song or script. If you are recording yourself, push your own performance to the dynamic ceiling you expect during the real take. Watch the input meter during this test pass.
Step 3: Adjust preamp gain until peaks hit the sweet spot
Turn the gain knob on your audio interface up slowly until the loudest peaks land between -12 dBFS and -6 dBFS. If peaks are still climbing past -3 dBFS, the source is too hot for that microphone or interface and you need to either move the mic back, switch to a pad, or use a different preamp. Never try to solve clipping by turning up after the fact.
Step 4: Confirm during a full dry run
Run a full rehearsal of the song at performance volume and watch the meter the entire time. Real performances almost always have surprise peaks you did not plan for. If anything clips during the dry run, drop the gain by 2 to 3 dB and repeat. This step alone will save you from ruined takes.
Gain Staging Explained: Building a Clean Signal Chain
Gain staging is the practice of setting the right level at every stage of your signal chain so nothing overloads and nothing falls into the noise. Each stage, including the microphone, preamp, converter, DAW input, and every plugin, has its own ceiling. If you set the level too hot at one stage, you can clip the next stage even if your meter looks fine.
A useful mental model is to think of each stage as having a sweet spot of -18 dBFS for average signal and -6 dBFS for peaks. This comes from the analog days when VU meters were calibrated to 0 dB corresponding to about -18 dBFS in digital. Modern plugins still use this standard because it leaves room for transients and EQ boosts.
In practice this means you should set the loudest peaks to about -6 dBFS at the input, then let your plugins and faders work within the headroom you have created. If your master bus clips, the answer is almost always that an earlier stage was running too hot.
Analog vs Digital Clipping: Key Differences
Analog clipping, like the kind produced by a tube preamp or tape machine, rounds off the top of the waveform in a smooth curve. This produces even-order harmonics that the human ear tends to find pleasing, which is why some guitarists actually chase analog clipping as a tone choice.
Digital clipping, on the other hand, lops the top of the waveform flat, creating harsh odd-order harmonics that sound buzzy and unpleasant. Worse, digital clipping is irreversible. The missing data cannot be recovered with EQ or compression. That is why every DAW meter shows 0 dBFS as a hard ceiling with nothing above it.
If you want the warm sound of saturation, use a dedicated saturation plugin that lets you control the amount. Do not rely on accidental clipping at the converter to deliver that character.
Common Mistakes That Cause Clipping and Noise Problems
Over the years I have made most of these mistakes myself. Here are the three I see most often and how to avoid them.
Recording too hot because louder feels better
When you record into a fresh track, a strong waveform looks impressive and a quiet waveform looks weak. That visual bias pushes many beginners to push gain higher than they should. Resist it. Peaks at -6 dBFS are loud enough. Anything above 0 dBFS is distortion, not volume.
Ignoring the preamp on budget interfaces
Cheaper audio interfaces often have noisy preamps that hiss at high gain. If you set levels conservatively at the interface and then boost later in the DAW, you will amplify that hiss. On a budget interface, getting close to the -6 dBFS ceiling and using the cleanest gain setting on the interface is usually the right move.
Forgetting about high dynamic range sources
Vocals that go from a whisper to a belt can have a 30 dB or wider dynamic range. If you set levels for the loud belt, the quiet parts disappear into the noise floor. If you set levels for the whisper, the belt clips. The professional solution is compression during recording or riding the fader manually, but the simplest fix is to set levels for the loudest moment and accept that quiet passages will need later gain automation.
How to Fix Audio Clipping if It Already Happened
Sometimes a take clips despite your best planning. If only the very top of one or two peaks is clipped, a de-clipping plugin like iZotope RX De-clip, Audacity’s Clip Fix, or a similar tool can sometimes reconstruct the missing waveform. The result will not be as clean as a fresh recording, but it can save a performance you cannot redo.
If the clipping is severe, repeated, or audible as obvious distortion throughout the take, do not waste time repairing it. Re-record. A clean take at slightly lower level will always mix better than a clipped take at higher level. No amount of EQ or limiting can fully undo digital distortion.
Frequently Asked Questions
How do I prevent clipping when recording?
Set the loudest peaks between -12 dBFS and -6 dBFS by adjusting your interface preamp gain during a loudest-pass test. Always watch the peak meter, not just the average, and leave at least 6 dB of headroom below 0 dBFS to absorb transients.
What level should my audio stay under for clear sound while recording?
Your peaks should stay under -6 dBFS and your average signal should hover around -18 dBFS. This range keeps your signal well above the noise floor while leaving enough headroom to avoid clipping on transients.
How do I record loud audio without clipping?
Use a pad switch on your preamp, move the microphone further from the source, or switch to a dynamic microphone that handles high SPL. Then set the gain so peaks hit -6 dBFS on the loudest expected moment.
How do I record without picking up background noise?
Get the microphone close to the source, use a directional pattern like cardioid, treat the room with absorption, and record at healthy levels between -12 dBFS and -6 dBFS so the signal sits well above the noise floor.
Final Thoughts on Setting Recording Levels
Learning how to set recording levels comes down to three habits: aim for peaks between -12 dBFS and -6 dBFS, always do a loudest-pass test before tracking, and trust your peak meter more than your ears when the room is loud. Once those habits are in place, clipping and noise floor stop being problems and start being non-issues. Next time you sit down to record, run the four-step workflow above and watch how much cleaner your sessions sound.