Maybe you’re a singer-songwriter who mixed a song and want to release it quickly. Or perhaps you’re an EDM producer working at a fast clip, and you don’t want to shell out the money to bring every new track to a competitive level. Maybe still you’re a mixing engineer looking to give a client a rough master that sounds better than the average limiter-slam.
That last guy used to be me. And it was frustrating because I had tools at my disposal that I knew mastering engineers used regularly. I watched tutorials; I talked to my friends in the mastering world; I read books on the subject.
I suspect many of us have.
What follows are the tricks and tips I’ve gleaned over the years to master one track at a time. The heavier work of mastering—blending together a cohesive album; creating DDPs and whatnot—that’s for other articles.
If you’re a musician, songwriter, producer, or mix engineer looking for tips on how to deliver your own master, read on; you might very well find something useful in my process.
You wouldn’t build a house without checking out the neighborhood first, would you? Otherwise, you might end up with something incongruous, like a yurt in the middle of some McMansions, or the Guggenheim Museum in the middle of the Upper East Side of Manhattan.
Similarly, you’ll want to have an idea of where your song will end up before you master it—especially these days, as Spotify, YouTube, Tidal, and Apple Music have embraced (to various degrees) loudness normalization. This means that if you deliver a master at a higher level than a streaming service’s target, the service will lower your song’s level to match the rest of their tunes.
Chances are, if you’re mastering one tune for consumption, you’re going to stream it—so use this to your advantage: Streaming services tend to operate between -12 and -16 LU, depending on the service, and this gives you more room to create a breathing, musical master.
You can see how loud your master is by checking your meters often in a plug-in such as Insight, observing your short term and integrated loudness readings to see if you’re in the ballpark. Need more info? We dive into common loudness myths and what to consider when mastering for streaming services here. If you’d prefer a video on the subject, check out the Are You Listening? episode on limiting and loudness in mastering below.
Two outliers here are SoundCloud and Beatport, which both boast no loudness standards and offer lower-quality audio than the other services. Here you might want to master more aggressively, as these services do not limit the loudness of recordings you submit.
There is also the context of your genre to keep in mind, which brings us to our next point.
Your next move is to locate tunes similar in tempo, genre, and arrangement. Try to obtain a lossless file if you can, such as .WAV. Lower res file types like .AAC or .MP3 will give you a picture of how a song should sound, but once you’re getting granular, you’ll notice the draining quality of lossy codecs, and the comparison will be harder to achieve as a result. Try to avoid them unless it’s the only option.
Your own mixes also make great references, as you’re already familiar with the material. The reason you want a reference track is because your master needs to compete on a commercial platform. It doesn’t need to match a reference perfectly, but the two must meet for a game in the same ballpark. Their overall level, frequency content, and dynamic heft must gel. Think, “would this work on the same playlist?”
We have many different guides on reference tracks, from tips on how to use them correctly, to genre specific primers with reference curves you can download and use yourself in Ozone’s Tonal Balance Control (which you can demo free for 10 days).
Meters are critical in mastering because they give us visual context that can confirm what our ears are hearing. Used in conjunction with a reference track, and meters can be an enlightening resource. You’ll need an objective check on your choices.
Displaying information in LUs (loudness units), these meters can measure the short term (momentary) as well as integrated (average) level of your track. Integrated loudness is typically used to ensure compliance with broadcast standards, but is useful for spot-checking the overall loudness level of your music tracks as well. Short term loudness is useful for checking the dynamic range between the loudest and quietest sections of your mix.
These display the frequency content of your mix, and can be useful in discerning where you need to add low end, take away from the upper mids, and more. They are a good objective check on your ears and can clue you into what your room might be obfuscating. For more of an explanation, check out the metering section of this article.
This tool gives you a good feeling for whether your master is as wide as your reference, or conversely, if it’s entering dangerous territory, which can happen if you’re employing widening trickery.
I also tend to check my master in mono frequently, listening to whether important elements have disappeared. If they have, that’s a clue that something is awry.
If you have a meter that gives you a pictorial view of your loudness over time (such as the one pictured below), this can give you an excellent readout of your dynamic range.
Luckily iZotope makes a metering system that handles all of these—and in a customizable interface to boot. I speak, of course, of Insight, which I use on every mix and master. It’s a hell of a plug-in.
For your purposes, start with these three tools: EQ, compression, and peak limiting at the end of the chain. That’s it. That’s all you likely need. Whether you put the EQ before the compression or after is up to you—I tend to go EQ before, but that’s me.
Limit yourself to these, and you’re in far less danger of messing things up. Seriously, there’s a reason Ozone’s Master Assistant doesn’t automatically slap stereo-width tools and spectral shaping on your master, and it’s because they are garnishes, not staples.
Much fancy trickery can actually be accomplished with these three tools. Take stereo width adjustments: Any EQ with mid/side capabilities can help you widen or narrow your mix. Simply raise the frequencies you’d like to hear on the sides with broad, wide EQs to achieve a boost in width. Be careful, of course, because you can easily change the mix’s intention. But the difference between a multiband stereo-width plug-in and an EQ used well in mid/side mode is not as big as you’d think.
Working in reverse means you try to achieve the desired level first, then apply dynamics-shaping as needed, and move on to EQ after that.
Why work this way? After all, many tutorials promote the opposite, advocating EQ moves first, followed by compression, and then bringing everything up with the limiter.
I used to work that way. I also used to deliver masters that sounded far more squashed than the ones I deliver now. Also, the choices I made regarding frequency-shaping and dynamic movement didn’t hold up when the overall level was boosted at the end—I found the moves I made at quieter levels sounded off, or the limiter’s innate aesthetic drastically impacted the timbre, undoing all my precious work.
When I reversed the order of operations, everything changed. The process became quicker, and the results tended to sound more natural, so much so that people began to hire me for the occasional mastering job.
Yes, you do lose that dramatic moment of making the tune louder at the very end of the process, but c’mon, that was always a cheap thrill anyway. You also get a sense of a song’s innate loudness potential earlier in the process—and believe me, every tune has one. This impacts your choices down the line, giving you more room to operate and more of a structure to work within.
But if you’re limiting straight away, won’t you hit the limiter too hard? Won’t you introduce distortion into the signal immediately? Sure, you may notice some distortion at this phase. But now you can use the other tools at your disposal—the compressor and the equalizer—to avoid these distorted artifacts. This, in conjunction with corrective tonal choices, becomes the crux of your work.
Onto compression, which you’ll use in one of two ways: to control the dynamics of the material or to add color. For tips on how to achieve the former, see this article—particularly the section on how to fine-tune attack, release, ratio, and threshold controls. Both processes should help you tame distortion before the limiter, as the dynamic range will be addressed either way, and the limiter won’t have to work as hard.
Now onto EQ: As you presumably have some knowledge of how to mix, your mix probably sounds good already. Your job now is not to make it sound better, but to allow the mix to shine across all sorts of listening environments and platforms. You do this, as a rule, by implementing very subtle moves—a 1 dB cut at 800 Hz or so if it’s nasal, a subtle shelf of 1 dB at 8 kHz if it’s too dull, and so on.
As you can use a compressor to keep the peak limiter from doing the heavy lifting, you can also use the equalizer to back off frequency ranges that cause the limiter to clamp down in a particularly distorted manner. You can better do this by implementing the following suggestion.
A mastering suite like Ozone will allow you to do this handily, but if you’re mixing and matching plug-ins, try the following:
Put an unprocessed copy of your mix on a dedicated track. Feed the track and your master-in-progress to the same auxiliary channel, and slap a loudness meter on this aux. Next, get your master in the loudness-ballpark. You can use Ozone referencing to do this, and even level match during it.
Now, as you loop the loudest section of the song, note the loudness reading of the unprocessed mix, switch to the mastered track, and bring its fader down till you’ve achieved the same meter reading as the mix. (Note that I said level match your master to the mix and not the other way around—if you were to boost the mix, you’d hit the digital ceiling and hear nasty distortion.)
As you listen to the two tracks at comparable levels, see if what you’ve done has been an improvement, or if it’s made things worse. This might not be the best moment to judge EQ aesthetically, but you’ll surely hear the differences in dynamic range: if something sounds too squashed, it’ll jump out here, and you can back off your compressor, or make limiter-based equalization decisions (i.e. backing off 100 Hz or so if the kick triggers the limiter too hard).
As soon as you like your results, move on to the next step, which is:
Much like you did before, import a reference track into Ozone, or set up another track with a reference mix on it. Now, level match that reference track to your master. Track referencing is an incredibly helpful tool to keep in your back pocket for times when your ears might be failing you.
As you switch between the reference track and the master, you can make aesthetic decisions regarding EQ. Does the reference feel brighter than your master? Then go ahead and try to match it with a subtle shelf. Does it have cleaner low mids? Try a dip in midrange to uncloud the master. Check compression as well: Does the master feel more squashed than the reference mix? Pull back on the compressor, or don’t hit the limiter so hard.
It’s a mixing engineer’s job to make the mix as good as it can be. A mastering engineer’s job is different in one key way: the mix (or group of mixes) must sound good in as many disparate rooms as possible. Thus, you must make sure your master translates to various places.
How, as a novice, can you do this? By employing every monitoring environment at your disposal. Work off your trusted rig, for sure, but then monitor the mix any which way you can, including through studio cans, computer speakers, laptop speakers, clock radios, your car, your best pair of consumer headphones, and your worst pair of earbuds. Buy terrible speakers for the sole purpose of monitoring in terrible conditions. Seriously—it’ll go a long way.
Take the average here and look for acceptability: if you’re consistently finding everything harsh on consumer-grade headphones, well, you need to tame that high end. If your midrange is underrepresented on all your smaller speakers, you need to give it some juice. And so on.
For your main setup, it’s quite important to have a fixed monitoring level, as listening at different levels can lead to inconsistent choices and ear fatigue. You might boost the lows when listening quietly, then turn up the level, decide the mix is too boomy, and wind up boosting the highs; you could’ve left everything alone in the first place!
Your monitoring level should be comfortably loud—as in, loud enough to hear the effects of the lows and trebles, but not so loud that you fatigue your ears. Some engineers prefer 80–83 dB SPL. Whatever you choose, a fixed monitoring level will help you make objective decisions, because it will be an immutable, stable point of reference in your studio. You’ll begin to know, over time, that if a master sounds good at this level, it will translate to other systems at other levels.
It’s also useful to have a monitoring level about 12 dB quieter on hand (often called a dim position), to make decisions about how the master sounds when listened to more quietly. Flick to this setting occasionally, and then come back. You can set one up in your DAW with any simple gain plug-in; just be sure to leave it off when you bounce.
Mastering engineers, as a rule, work very fast. It makes sense: working quickly helps you stay objective—you don’t find yourself diving into the weeds if you actually avoid the weeds. One way to work fast and achieve good results is to cycle through these steps again and again, making quick changes in short order.
It could look like this: I’m happy with the way the mix is slamming, but the master feels a little thick in the low mids. So I make a change to the EQ. Now, I check it against the reference track, and I find they’re in the same ballpark. I proceed to check the master against my unprocessed mix, but notice my low-mid dip has had an adverse effect on the compressor. Here, maybe I adjust the comp’s threshold, or I change its sidechain filter a bit so that different frequencies are triggering compression.
It goes on and on until I walk away, come back ten minutes later, and find myself satisfied with the results.
Once you’ve gotten the master how you like it, it’s time to bounce, render, or export the file. There are a few considerations here, and once again they depend on your destination. For CD and the majority of online aggregators, the name of the game is 44.1 kHz / 16-Bit WAV files. For all else, I recommend brushing up on this article.
You’ll want to apply dither to the final master, and if we had another 3,000 words, I could explain why. We don’t, so I’ll just refer you here. For what it’s worth, I tend to export my files in their original sample rate and resolution, and then apply sample rate conversion, add dither, or export to lossy codecs from a standalone application like RX. This gives me the freedom to move in any direction I choose, as well as excellent SRC tools.
Why did we save these topics for other articles? For one thing, we don’t have the space, and for another, you most likely can get a great-sounding master without these tools. I, for one, believe that you can—after all, you’re an iZotope reader, so you’re passionate about making informed decisions all along the chain, from the mixing to the mastering process. It’s our hope that this article will provide another useful perspective on your journey to audio mastery.
This article has been updated to reflect the newest versions of iZotope products.
Get top stories of the week and special discount offers right in your inbox. You can unsubscribe at any time.
Copyright © 2001–2019 iZotope, Inc. All rights reserved.