Gain staging means managing the level at every point along the signal chain so that each stage receives a signal that is loud enough to stay above noise and quiet enough to avoid distortion.
Noise floor and headroom
Every device adds a small amount of unwanted noise, called the noise floor. The useful signal must sit well above it. At the other end, every device has a maximum level it can handle before distorting. The space between your normal level and that maximum is headroom. Gain staging aims to keep signals comfortably between the two.
dBFS in digital audio
In digital audio, levels are measured in dBFS (decibels relative to full scale). The maximum value a converter or fixed-point file can represent is 0 dBFS, so all other levels are negative: -6 dBFS, -18 dBFS and so on. A signal that tries to exceed 0 dBFS is clipped: the waveform is flattened at the ceiling, producing harsh distortion that cannot be undone.
Clipping most often happens at the points where a signal enters or leaves the digital domain: the analog-to-digital converter on input, and the final export or digital-to-analog conversion on output.
Floating point inside a DAW
Most DAWs process audio internally in floating-point format, which has a very large dynamic range. This means a track can go well above 0 dBFS inside the software without being clipped at that point. The overload only becomes a problem when the signal reaches a stage with a fixed ceiling, such as the output converter or an exported fixed-point file.
That does not mean levels inside the DAW are irrelevant:
- Plugins that model analog gear are designed to respond to a particular input level. Feed them too hot or too quiet and the saturation, compression or tone changes, so the plugin may sound different from how it was designed to.
- Some plugins have fixed internal ceilings and may distort when overdriven.
- Meters and fader resolution are easier to use with sensible levels.
Common level conventions
These are conventions, not rules, and they vary between engineers and systems.
| Situation | Common guideline (approximate) |
|---|---|
| Tracking peaks | around -18 to -10 dBFS |
| Average level on analog-modelled plugins | around -18 dBFS, often treated as 0 VU |
| Mix bus peaks | enough space left for mastering, often several dB below 0 dBFS |
The idea that -18 dBFS corresponds to 0 VU is a common calibration convention, used because many plugins are modelled on that reference. Some systems use different reference values.
At 24-bit, recording at moderate levels costs little, because the theoretical dynamic range is so large that the noise floor is far below the signal. Peaking near 0 dBFS gives no real benefit and risks clipping.
Practical workflow
- Set the preamp or interface gain so peaks stay out of the red while the typical level sits near the target above.
- Use clip gain or a utility plugin to bring recorded clips to a consistent level before processing.
- Check the level going into and coming out of each plugin. Where a plugin has input and output controls, match the level so the processed sound is not simply louder.
- Leave headroom on the master bus rather than pushing it close to 0 dBFS.
Clip versus bus headroom
An individual track may peak low and still contribute to a loud mix, because levels add when tracks are summed. Check the meter on the group and master buses as well as on single tracks. If the master is too hot, lower all channels together or lower the individual faders rather than turning down only the master, so each stage keeps proper levels. See the signal flow article for where each of these controls sits in the chain.