
Preparing your music tracks for mastering
1. This is a concise entry with high information density about how to prepare your tracks and subsequent exported/bounced files for mastering.
Why your mix should not be hitting the top of the meter, what −12 dBFS on the kick drum does for you, and why 24-bit means you have nothing to lose. Written for people who mix at home, by someone who hears the results every day.
When starting a new track, set your kick or snare drum to peak at around −12 dBFS on its channel and on the master, then build the mix around it. You are very unlikely to clip, you will have ample headroom for export, and your mix will probably sound clearer while you work. That is it. The rest of this page explains why.
Digital audio systems have a finite way of representing your signal. The maths ends at 0 dBFS on your workstation’s stereo output. The moment a signal exceeds that, it is no longer being represented accurately; the waveform is clipped, flat-topped, and no mastering engineer can un-clip it. I still receive files that have gone past digital zero, or that were clipped inside a plug-in before the master output. In both cases the gain structure is wrong.
Much of this comes from habit. In the days of 16-bit recording you needed to keep the signal hot to stay clear of the noise floor. At 24-bit that is no longer true: the theoretical dynamic range is 144 dB, so peaking at −12 dBFS still leaves you 36 dB more signal-to-noise than 16-bit ever had. There is simply no reason to run hot.
Working at a lower level has a side benefit on modest equipment. A budget USB-powered interface has limited analogue headroom in the small op-amps driving its outputs, so distortion can creep in well before the digital maximum. Send it a more modest level and you will hear less distortion and more clarity, and you will probably make better mixing decisions as a result. Many analogue-modelled plug-ins are also designed around a nominal level, often close to −18 dBFS, so this approach brings you nearer to how they were meant to be driven.
Professional engineers have always optimised gain structure. On a large-format analogue console, channels were set with the PFL button so that each signal peaked at around 0 VU, a nominal level that sat midway between the noise floor and the point of distortion. 0 VU is roughly equivalent to −12 dBFS on a digital meter, referenced to +4 dBu (1.23 volts RMS). In other words, the level I am suggesting is the level a professional desk has always run at; your DAW’s meter scaling just makes it look low.
For information, +4 dBu is about 12 dB hotter than the −10 dBV consumer reference. If your analogue mixer outputs +4 dBu and your interface is set to −10 dBV, you will overdrive its inputs; switch one or the other to match. When taking levels from a band, peak at about −3 VU, because they always play harder on the take.
Recording or mixing through an analogue console, press PFL on each channel and adjust the channel gain trim so the signal peaks around 0 VU on the PFL meter. Do this for every channel and the gain structure through the desk is optimised: out of the noise, with ample headroom before distortion.
“Gain structure” by Barry Gardner, first published 2012, revised 2026.

1. This is a concise entry with high information density about how to prepare your tracks and subsequent exported/bounced files for mastering.

The 2 most important controls on a compressor are not what you might think.

As we all know mixing is hard work and a very skilled activity in the realms of music production. Mixing music successfully (and choosing…
Mix appraisal is free on singles and EPs if you want a second pair of ears on the result.