Gain Staging Calculator: Set Your Recording Level Correctly
Track with peaks around minus 12 to minus 6 dBFS and an average around minus 18 dBFS RMS, which is the digital equivalent of 0 VU on analogue gear. At 24 bit that leaves 18 dB of headroom for free, because the noise floor sits more than 100 dB below anything you will hear.
Definition: Gain staging is setting the signal level at every point in the chain so each stage runs in its designed operating range, neither so quiet that noise becomes audible nor so loud that the stage distorts.
Gain staging goes wrong in two directions and only one of them is fixable. Too quiet and you can turn it up later, losing a little noise floor you probably cannot hear. Too loud and the take is clipped, permanently, and the only fix is doing it again. This calculator works out the gain your source actually needs and tells you whether your preamp can deliver it.
What is the target level, and why that number?
The target every part of this site uses is -18 dBFS RMS as the average and -12 to -6 dBFS as the peak. That is not a taste preference, it is a calibration.
Analogue equipment was built around a nominal operating level marked 0 VU on the meter, with roughly 18 to 20 dB of headroom above it before the circuit began to distort. When digital converters arrived, manufacturers aligned them so that an analogue signal at 0 VU produced -18 dBFS in the digital domain, which meant the two systems could be patched together without anyone recalculating anything. That alignment then propagated into every analogue-modelled plugin ever written, because those plugins model circuits that expect a signal at that level.
The practical consequence is that a compressor or preamp emulation fed at -3 dBFS is being driven 15 dB harder than the circuit it models ever was. It will still make a sound. It will not make the sound the model was tuned for, and it is why plugin chains sometimes feel harsh for reasons nobody can locate.
Why recording hot stopped being good advice
It was correct advice, and then the hardware changed underneath it. At 16 bits the quantisation noise floor sits about 96 dB below full scale. Record a quiet passage at -40 dBFS in 16 bit and you are using only the top 56 dB of the available range, which on a quiet source with a wide dynamic range was audible. Getting the level up was worth real effort.
At 24 bits the theoretical floor drops to roughly 144 dB below full scale. In practice converter electronics limit you to somewhere around 110 to 120 dB of real usable range, which is still far more than any microphone in any room can deliver, because the room's own noise floor arrives long before the converter's does. Recording 15 dB quieter costs you 15 dB of a margin you cannot hear the bottom of.
| Format | Bits | Theoretical range | Typical real range | What it means for tracking |
|---|---|---|---|---|
| CD audio | 16 | 96 dB | 93 dB | Level discipline matters. Track closer to -12 dBFS peaks. |
| Standard recording | 24 | 144 dB | 110 to 120 dB | Headroom is free. Track at -18 dBFS RMS with confidence. |
| Float recording | 32 | over 1500 dB | Converter limited | Clipping in the file is effectively impossible. The preamp can still clip. |
Note the last row carefully. A 32 bit float file cannot clip, but the analogue preamp in front of the converter absolutely can, and no file format saves you from that. The DI box and pad switches exist for exactly that stage of the chain.
The gain mistake I still catch myself making, and the one I see most often in other people's sessions, is setting the level from a spoken sound check. A singer says "check one two" at conversational volume, you set the gain so that reads healthy, and then they perform and hit you with 12 dB more. Always set gain from the loudest thing the performer will actually do, which means asking them to sing the loudest phrase in the song before you touch the knob. It costs thirty seconds and it saves the take.
When do you need more gain than your interface has?
Microphone sensitivity varies by more than 20 dB across common types, and that is the whole problem. A large diaphragm condenser puts out a strong signal and needs modest gain. A passive ribbon or a low-output dynamic like the Shure SM7B puts out far less, and on a quiet source you can run out of clean gain on a budget preamp before the level is usable. What you hear is hiss, and the hiss is the preamp's own noise being amplified alongside the signal.
Two solutions, and they are not equivalent. An inline FET preamp or a Cloudlifter adds around 25 dB of clean gain right at the microphone, before the cable run and before the interface preamp, so the interface preamp is now amplifying an already healthy signal and contributes far less noise. That is the cheap fix and it works. The expensive fix is an interface with a better preamp and more of it, which also buys you lower noise across everything else you record.
If you are choosing an interface and you own or want a low-output dynamic, published maximum gain is the spec to read first. Compare figures in the interface spec database, or work through the whole decision in how to choose an audio interface.
Gain staging through the whole chain
The preamp is the stage that matters most, but it is not the only one. Every point where the level changes is a gain stage, and each one has a range where it behaves.
- Source to microphone. Distance is a gain control. Moving from 12 inches to 6 inches adds roughly 6 dB, and on a cardioid it also adds proximity effect in the low end.
- Microphone to preamp. Set from the loudest phrase, targeting peaks around -12 dBFS. Engage the pad if the gain is at minimum and still too hot.
- Converter. Nothing to set. This is where dBu becomes dBFS.
- Channel fader and plugins. Keep the signal near -18 dBFS RMS going into analogue-modelled plugins. Use trim, not the fader, to correct a track that was recorded loud.
- Mix bus. Leave headroom. A mix peaking at -6 dBFS gives mastering something to work with.
Related tools
- Latency buffer calculator: the other setting you change between tracking and mixing
- dB reference levels chart: every level standard on one page
- How to set recording levels: the full workflow, step by step
- Signal flow explained: what each stage in the chain is doing
Frequently asked questions
What level should I record at?
Aim for peaks around minus 12 to minus 6 dBFS and an average level around minus 18 dBFS RMS. Minus 18 dBFS is the digital equivalent of 0 VU on analogue gear, so it is the level most plugins are calibrated to expect, and it leaves 18 dB of headroom for a louder-than-expected take. Recording closer to 0 dBFS gains you nothing at 24 bit and risks a clipped take you cannot repair.
Why is minus 18 dBFS the standard tracking level?
Because it is where analogue 0 VU was aligned when digital converters were designed. Analogue gear was built around a nominal operating level with roughly 18 to 20 dB of headroom above it before distortion, and converter manufacturers matched that so an analogue signal reading 0 VU produced minus 18 dBFS in the digital domain. Every analogue-modelled plugin inherits that calibration, which is why it sounds right when fed at that level.
How much gain does an SM7B need?
About 60 dB for a normal speaking or singing voice at close range, and more for a quiet source. The SM7B is a dynamic microphone with a low output sensitivity, so many budget interfaces run out of clean gain and you hear preamp hiss before you reach a usable level. An inline gain booster adds around 25 dB of clean gain and turns a marginal preamp into a comfortable one.
Does recording louder give better quality at 24 bit?
No. A 24 bit recording has about 144 dB of theoretical dynamic range, and even a good converter delivers well over 100 dB of usable range, so the noise floor sits far below anything you will hear. Recording at minus 18 dBFS instead of minus 3 dBFS costs you nothing audible and buys you 15 dB of insurance against a take that gets suddenly louder. The old advice to record hot came from 16 bit, where it mattered.
What is the difference between dBFS and dBu?
dBFS measures level inside the digital domain against a fixed ceiling, where 0 dBFS is the highest number the format can represent and every real signal is negative. dBu measures analogue voltage against a reference of 0.775 volts, and it has no ceiling until the circuit distorts. The two meet at the converter, and the common alignment is that plus 4 dBu on the analogue side produces minus 18 dBFS on the digital side.
Should I use the pad switch on my interface?
Use it whenever the source is loud enough that the preamp is near the bottom of its gain range or the input is clipping with the gain fully down. A pad attenuates the signal before the preamp, typically by 10 to 20 dB, which lets the preamp work in its comfortable middle range instead of being fought against. Loud sources that commonly need it are guitar amps close-miked, snare drums, and any line-level device plugged into a microphone input.
Working out your own room and signal chain? The Home Studio Build Planner is the paid version of these pages: 8 printable worksheets you fill in with your own numbers, plus the full PDF, $29.