Should I Normalise My Mix Before Mastering?
Usually, no. You do not need to normalise your mix before sending it for mastering.
If your mix is clean, unclipped and sounding the way you want it to sound, you are already most of the way there. There is no particular peak level or loudness target you need to hit before sending it over, and there is no need to make it artificially louder just because you are about to have it mastered.
In most cases, the best approach is simply to leave normalisation switched off when you export and send the mix at the level it naturally comes out of your session.
If youβre looking for a complete export checklist rather than just the normalisation question, read How Should I Export My Mix for Mastering?.
Do I Need to Normalise My Mix Before Mastering?
No, not normally.
It is easy to see why there is confusion around this because plenty of DAWs put a Normalize or Normalise option right there in the export window. It sounds like the sort of thing you might be expected to switch on to make the file properly prepared for mastering.
In reality, I do not need your mix to arrive at a particular level before I can work on it. If the mix is quieter than I need, I can increase the gain. If it is louder than I need but still clean and unclipped, I can reduce it.
So in most cases you can simply leave normalisation switched off and forget about it.
What Does Normalising a Mix Actually Do?
Normalisation is essentially a way of adjusting the overall gain of an audio file so that it reaches a chosen target.
With peak normalisation, the software looks at the highest peak in the file and then turns the whole thing up or down by the amount required to make that peak reach a specified level. If your mix peaks at -4 dBFS and you normalise it to -1 dBFS, for example, the whole file is increased by 3 dB.
The important point is that straightforward peak normalisation is not remixing the track. Your vocal does not suddenly become louder relative to the drums and your guitars do not move forward in the balance. Everything is simply turned up by the same amount.
Loudness normalisation uses a measured loudness target instead, often expressed in LUFS. You may have come across this in discussions about Spotify, YouTube and other streaming platforms, but there is generally no reason to make your unmastered mix hit a particular LUFS figure either. If youβre wondering what LUFS or overall loudness your mix should actually be before mastering, Iβve covered that separately in How Loud Should My Mix Be Before Mastering?.
The final loudness of the song is something that can be dealt with during mastering, so there is no real benefit in trying to manufacture it beforehand.
Why You Donβt Need to Normalise to a Specific Level
A lot of this question comes back to the old advice that a mix supposedly needs to peak at -6 dBFS, -3 dBFS or some other exact number before it can be mastered.
Thankfully, it does not need to.
If your mix is otherwise healthy, changing its overall gain is easy. Normalising it simply to make the meter land on a particular number does not give me anything useful that was missing from the original file.
That does not mean levels are completely irrelevant. Clipping is a different matter because that can actually affect the audio, but there is a big difference between avoiding a real problem and chasing an arbitrary target because you think a mastering engineer expects it.
I have covered that distinction properly in How Much Headroom Should You Leave for Mastering?, but the short version is that there is no magic amount of headroom you need to create before sending your mix.
So if the only reason you are about to click Normalize is because somebody told you your premaster has to peak at -6 dBFS, there is no need to worry about hitting that target.
What If Iβve Already Normalised My Mix?
This is not usually something to panic about.
If you have already exported the track with straightforward peak normalisation switched on, and the resulting file is still clean and unclipped, that does not automatically mean you have damaged the mix.
Normalisation may simply have turned the entire file up by a fixed amount. The balance between the instruments is still the same, the dynamics are still there, and I can reduce the overall gain again if I need to.
If your project is still open and it takes ten seconds to export another version with normalisation turned off, I would generally send that one instead. It removes an unnecessary step and gives me the mix exactly as it leaves your session.
But if all you have is the normalised version, send it over. The fact that a file has been peak-normalised is not, by itself, a mastering emergency.
Normalisation vs Limiting and Clipping
This distinction is worth making because normalisation, limiting and clipping often get grouped together as different ways of βmaking something louderβ, when in practice they are doing quite different things.
Straightforward normalisation changes the overall gain of the file. Turning a mix up or down by the same amount yourself is also a gain change.
A limiter behaves differently. It reacts to peaks in the music and applies gain reduction when they reach a certain level. Push that hard enough and it can change the transients, dynamics and feel of the mix, which means I cannot simply turn a limited mix down later and recover what was there beforehand.
If you have a limiter on your master bus, that is therefore a separate question, which I have covered in Should I Remove the Limiter Before Mastering?.
Clipping is different again. If distortion from clipping has already been printed into the exported file, turning it down afterwards only gives you a quieter clipped file. It does not restore the peaks that have already been lost.
That is why I am much more interested in whether the mix is clean and healthy than whether its highest peak happens to land on some particular number.
What Should I Do When Exporting My Mix?
For most mixes, keep it simple.
Leave Normalize/Normalise switched off and export the mix at the level it naturally comes out of your session. There is no need to deliberately make it louder because you are sending it for mastering, and no need to change the overall gain purely to make the file peak at -6 dBFS, -3 dBFS or another number you have been told a mastering engineer expects.
Make sure the mix is not clipping and send the version that actually sounds the way you want the song to sound. That gives me something much more useful to work with than a file that has been altered just to satisfy an arbitrary meter reading.
If you are not sure whether your current mix is ready for mastering, you are welcome to send it over. I am always happy to have a listen and let you know if there is anything worth changing before we get started.