Are You Listening featured image.jpg
May 30, 2019 by Jonathan Wyner

How to Master an Album, and More Audio Mastering Questions Answered

Expert engineer Jonathan Wyner answers five common audio mastering questions from viewers of ‘Are You Listening?’. Learn how to better manage your next session.

This article references previous versions of Ozone. Learn about the latest Ozone and its powerful new features like Master RebalanceLow End Focus, and improved Tonal Balance Control by clicking here.

To close out season one of Are You Listening?, we opened the floor to your questions. Out of those, we chose five that pro mastering engineer and iZotope Education Director Jonathan Wyner answers in the video below.

These tips run the gamut of audio mastering, and hopefully will be helpful in your own practice. While we couldn’t answer every question, we’ll be using all comments and questions to help inform future episodes. So please continue to comment, question, and be curious.

If you’d like to learn more, we have lots of educational resources for you to learn about mastering, mixing, audio repair, recording, and even psychoacoustics and career advice.

  1. What order should my modules be in a mastering session?
  2. How do you approach mastering an entire album?
  3. Are there different strategies or rules for mastering songs in different genres?
  4. What is crest factor, and is there a crest factor I should aim for?
  5. How many different exports do I make, and to what loudness standard?

1. What order should my modules be in a mastering session?

The first thing in your chain is probably an equalizer. It may not always get used, but it’s there to help correct any tonal balance issues or problems before you start dynamics processing. You’ll usually see an EQ, some kind of dynamic processing, and maybe another EQ after for overall enhancement and broad changes. This is a good starting point for your mastering signal chain.

Look at the Ozone manual, or some of our presets, and you’ll start to notice similarities between signal chains. There’s always a limiter at the end, and the limiter itself is probably the brickwall limiter. There may be other limiting happening beforehand, adding color and so on, but the brickwall limiter usually sits at the very end of your mastering signal chain as protection against distortion.

What goes in-between your initial EQ and final brickwall limiter is all to taste. There are lots of other modules in Ozone and other types of processing—like reverb (occasionally used in mastering), excitation, and stereo imaging (Ozone Imager is free!). Sometimes you may need to experiment, but if you start with a chain of an EQ, Dynamics, EQ, and then a limiter (called Maximizer in Ozone), you’re on a good path.

Ozone module chain ft maximizer.png

Example Ozone module chain

2. How do I approach mastering an entire album?

It can be challenging to make your tracks sound cohesive when they’re in different genres, styles, and intensities. There are a few things you can focus on to navigate this particular challenge.

Ask yourself, “What’s the lead instrument? What’s the focal point of this record?” Choose a track (generally a vocal) that might represent the lead artist, or maybe a vocal duet, or in the case of an instrumental album, maybe it would be saxophone or guitar. Think about how you will maintain consistency from track to track in terms of instrument tonality and level.

There can be some variance. A vocal might sound a little hotter in an open quiet track than it would sound in a fully arranged or denser track, but if the vocal wanders dynamically—if it gets quieter or softer by more than 6–10 dB—there can be a real disconnect for the listener where they have to listen really hard until the element they’re listening to suddenly disappears.

Another way to tackle an entire album is to think about dynamics. It is a challenge to think about dynamics when pushing a consistent level into a record. But even if you’re aiming for a high RMS level, you’ll want to leave enough room for some contrast. A loud track should sound a little bit louder than a quiet track, and, even more importantly, should sound loud if it plays right after a quiet track.

There’s nothing worse than reaching the end of a quiet track only to find a full-on screaming moment in a record that doesn’t scream and doesn’t sound loud. Thinking about dynamic contrast is really important. Beyond this, focus on a consistent mid-range, a good tonal balance. But remember, we’re making art in a sense. Or at least we’re making something creative, so you should allow for differences to happen.

3. Are there different strategies or rules for mastering songs in different genres?

Genre matters. There are basic boundaries around which we want to put the sound that we're creating, no matter the genre. Beyond that, it's less about rules than it is about understanding the genre. EDM records usually have a lot of edgy or aggressive synth and virtual instrument sounds that drive the music forward. Usually, you don't want to smooth out an EDM record too much, unless it happens to be spiky. So you wouldn't want to use a tube exciter or something with tubes in the signal path because they'll sweeten the track undesirably and add even order harmonics.

If you understand what your tools are good at, and what a genre requires, you’ll have an easier time answering the questions: “What's the sound I'm after? What changes, if any, do I need to make, and what tools will I need?"

4. What is crest factor and is there a crest factor I should aim for?

From a technical perspective, crest factor is an expression of the difference in peak level against the average or RMS level in a very short term measurement. If you look at the crest factor of a sine wave, classically, it’s a 3 dB difference between peak level and sustained difference. If you take a look at the crest factor of a square wave, there’s a 0 dB difference because it spends the entire time, the entire level, at the same value as peak. Technically speaking, this is what crest factor is.

In music production, we can think of crest factor as the difference between peak and average. When you look at a meter like what you might find in Ozone, you’ll see the two levels happening at the same time. The average level is the lower value that moves more slowly, and the peak level is the thing that lives quickly at the top of the dynamic range. Beyond this, if you look at a dense portion of a track, we can relate peak to average as being the drum strike of a kick or a snare and the sustain of melodic or tonal instruments.

This is where crest factor really starts to get interesting. If you have a very wide crest where the peaks are high compared to the average, you’ll have something that’s going to sound more spiky, more transient, perhaps more exciting. But here you’re also going to lose contact with the sustain, the tonal portion of it.

So there’s a sweet spot between peak and average, but it’s a little hard to come up with a spec. Beyonce’s Drunk in Love is a wonderful example of a very high crest factor that sounds perfect. The crest factor sits somewhere around 15 dB, which sounds very loud but if you listen to the track you’ll notice it’s very sparse. There’s a great deal of room between anything with transients to it and the bass, which drives strong peaks in the track. That level needs to be pushed up so you have a momentary bass peak and a much lower sustain that comes from when the vocal comes in.

In pop music, you might notice a crest factor or peak-to-average relationship that’s somewhere in the neighborhood of 6–8 dB. Any higher and you lose bottom punch and hear more high-end distortion. If you have to push the level to this point, understand that you’ll need to tackle these challenges; get a little bit more low end and manage the top end a little bit better. Most well-balanced recordings with dense arrangements in the louder sections max out in that zone.

Different genres also have different allowances. More dynamic music will have a higher crest factor, like an acoustic jazz album which might have a crest factor closer to 14—16 dB. While there's a rough relationship between crest factor and genre, you have to think more about what's going in the music to come up with a number.

5. How many exports should I make, and to what loudness standard?

It's unrealistic to optimize the level of your recording for every single distribution format and for every listening context. The levels used by most streaming services are very similar. If you make one version of your record that will work for loudness normalization, you should be in good shape. Most mastering engineers do a version of the record that’s as hot as they can make it, while still sounding good on a streaming service.

Say you’re aiming for -12 or -11 LUFS; if it sounds good and lively when it comes down to -14 LUFS on a streaming service, you’re in good shape. You can still go from there up a little bit if you need to if you're going to put something on SoundCloud and it needs to be a little bit hotter.

You can push the level up a little bit for a peak normalized playback, and then make a version with no limiting or at least very little limiting to have in your back pocket, so that if it gets sent out for cutting vinyl it will sound as good as it can. Vinyl cutters don't like limiting! Make three versions (maximum) of everything that you do. 

Any questions?

Hopefully, these tips help you better manage and conceptualize your next mastering session. Don’t forget to subscribe to our YouTube channel for more informative videos, and explore our mastering articles to further improve your craft and get your questions answered. Thanks for reading, and stay tuned for season two of Are You Listening?