Simply put, if you treat mixing and mastering as separate stages, your tracks are likely to sound much better. In this feature, we'll look at which plug-ins and effects you should use at this final stage, and in which order. But first and foremost, we'll see that postponing their deployment—so they're used only when your track is already sounding great—will lead to better results.
How do you prepare the optimum mix file for mastering? Ideally, without any plug-ins in the output channel. You're looking to produce a mix that peaks at somewhere between -2 and -3 dB, so you'll want to leave some headroom to work with in your mastering session. If your mix is peaking higher than that, particularly if it's overloading, you can begin to understand why a separate mastering session is a good idea.
If your session looks like this one, drop all of the fader levels within your track by the same amount, to bring the overall level down. Alternatively, placing a gain plug-in on the output channel or bringing down the master fader are good ideas. Then, bounce the stereo file in high quality—a 16-bit, 44.1kHz (at least) WAV or AIFF file, rather than an mp3, will always yield better results. Also, ensure that any automatic normalizing is switched off. As you're trying to leave headroom in this file, you don't want your workstation processing it at the bounce stage to arbitrarily drag the loudest point to 0 dB full scale.
Let's think about the chain of effects we're likely to need, and cover a few that might also suit the mix, depending on the music content. Setting up a channel of effects in this context is different from setting up one for an individual sound within a mix, as we'll need to place effects in the chain that you might not adjust until later. For instance, I'd recommend placing an EQ at the top, ideally one with multiple bands, so you can target a number of narrow frequencies at the same time. The likelihood is that you won't need to make adjustments to tone right away—after all, if a mix is deficient in treble or bass, for example, you'd have addressed these issues at the mix stage.
So why is EQ needed at all? Well, it's needed in preparation for what will come further down the chain of processors. A huge part of mastering concerns dynamics processing, specifically compression and limiting. And as you'll have noticed when you compress or dynamically shape individual sounds in a mix, compressors have a habit of affecting the tone of a sound as well as its volume. This shouldn't come as a surprise when you think about it. An individual sound such as a synth lead will be rich in harmonic content, with plenty of bass, mid-range and treble likely to be present in the sound. So as a sound is compressed, all frequency content will be re-prioritised as it's processed dynamically, often leading to sounds becoming brighter or duller as they decay, for instance. This means that when you insert a stereo or multiband compressor into the output chain (we'll come to both of these shortly), there's a good chance there will be knock-on effects for tone. This in turn means that having an EQ ready to make adjustments, to offset any unwanted tone change, will be important. So insert an EQ as your first mastering processor but leave its settings alone for now.
Compression comes next. There are two approaches to compression at the mastering stage, the simpler of which is to set up a regular stereo compressor that will dynamically process the entire mix, all the way across its frequency range.
Again, there's a forward-thinking approach to consider here, as the main purpose of stereo compression at this point is to make life easier for the limiter, which will follow later. The limiter's job is to add volume, and this is easier when the overall dynamic range of a track has been reduced a little. How much compression you'll require will depend on the contours of your track, but if your drums are prominent, there's a good chance that these will form the loudest content within your track, and you'll be able to see this as a series of spiky transients in your mix's waveform. Catching these, so that your compressor's threshold is set to duck a little each time one plays, will bring their levels closer to the average of the mix.
You're not looking to reinvent the wheel here or to completely squash the dynamics of your carefully mixed record—as a starting point, look to reduce no more than 2 or 3 dB from these loudest peaks. You'll barely notice the difference in level, but your limiter will thank you later on. If your mix is dynamically wide, with huge peaks and a much quieter average level beneath it, drop the threshold to duck the level of these a little more. But remember the more you do this, the less dynamically rich the mix will be.
That deals with the threshold point, but what about ratio? This controls how much those peaks will be reduced in volume. A ratio of 2:1 is designed to halve the volume of signals above the threshold point, and with the threshold set high enough to only catch peaks, you may find it sufficient. Indeed, some mastering engineers don't like going above ratios of 1.5:1 to ensure that dynamic processing is as subtle as can be. Again, use your ears to judge what sounds best for your track.
The second, more complex approach to dynamic processing is to use multiband compression. This processor has the added benefit of splitting up the frequency bands of your mix file, so that each can be compressed individually. Suppose the bass end of your mix needs more compression than the treble. A stereo compressor's global approach won't allow you to make frequency-based tweaks. If you like the idea of multiband compression, remember to deal with each band in turn, rather than setting up lots of bands and trying to figure out what they're all doing at once. Once you've set up a band, set its threshold and ratio and then adjust output volume for each until you've got a good level, before tackling the next band. Remember, too, that you'll probably need to adjust the default crossover points between bass, low mid-range, upper mid-range and treble to suit your track.
The final essential mastering processor, and the one you've probably found most tempting at the mix stage, is an output limiter. Let's pause for a moment to understand exactly what these do. Before the advent of digital recording, masters were printed to tape, which allowed you to go 'into the red' (between +2 and +4 dB) at louder levels without immediately obvious distortion. Indeed, many popular tape saturation plug-ins emulate the pleasing warmth and colouration you can hear when tape is pushed toward distortion without actually fully overloading. The extra harmonics this generated are musically pleasing, so engineers were frequently tempted to try this when running off a mastered mix. In the present day, now that masters are digital, life is much simpler: 0 dB provides the threshold point above which digital mixes overload. But this is where confusion takes over. If your mix has been bounced so that its loudest moment catches 0 dB, how on earth can you make it louder? Surely by turning it up, those peaks will cross the line into the red to produce distortion. That's true, but it doesn't take into consideration how we hear.
Suppose you slam the door in your studio. Then later in the day, you walk past some workers drilling in your street. Let's also suppose that, through lucky chance, the loudest moment of the door slam happens to the same volume as the guys drilling. Which sound would be louder? The drilling, of course. Why? Because drilling is a sustained sound—the high volume is held for a period of time—whereas the door slam very quickly decays to silence.
The rough settings we've used so far simply highlight that mastering requires going back to earlier processors in the chain, once you've got a basic limiter level you like. Our mix now sounds quite woolly at the bottom end, the result of the way the limiter is reacting to bass frequencies. Let's consider our EQ options for removing or reducing unnecessary content. We're using DMG's Equilibrium EQ. It allows us to set up multiple frequency bands, which makes it a great tool for mastering. We'll focus on four low frequency bands, the first of which is actually a high-pass filter.
It's a common misconception that high bass levels equal excitement and energy; in mastering sessions, often the opposite is true. As we've already seen, limiters target bass content first, often consuming the mix in murky, unfocused sound. Removing or reducing bass content you don't need will provide clarity and sonic richness, and you can often start by scooping out the busy low frequency content that most speakers won't even be able to play back.
Beyond EQ, compression and limiting, which other processors are regularly called upon at the mastering stage? To answer this, it's helpful to look at an all-in-one mastering plug-in like iZotope's Ozone. It takes a modular approach, with independent processors for EQ, dynamics and maximizing, in addition to a reverb module, a harmonic exciter, stereo imaging and a second EQ for use after the dynamics stage. (Remember, all of these modules are optional, so if you're an Ozone user, don't just switch on all of the power buttons next to each processor—you'll lose track of which processors are doing what to the sound.) Unless you're mixing quite sparse, dry music, it's unlikely you'll need the reverb module, particularly if reverb has already been used at the mix stage. However, both harmonic excitation and stereo imaging could prove very useful, if they're used carefully.
As you can hear through these examples, some mastering settings are dramatic, and others produce very subtle changes. Limiters have an obvious effect, while EQ, compression, multiband compression, harmonic excitation and stereo imaging can be extremely hard to hear. Listening carefully, though, you can pick out how they control peaks, add extra flavour, pull elements out of the mix and add control or drama. If you're new to carrying out your own mastering, it's worth breaking ranks from the usual advice for a moment. I'd actually recommend calling up a preset plug-in chain as a means of familiarising yourself with the steps a mastering engineer would take. You can then bypass various modules and listen carefully to the differences. Once you've auditioned different presets, subtly change its settings to suit your track. Also listen to your unmastered mix and think about what it needs. Is it bloated? Is it over-bright? Is the bass too wide? How loud do you want it? These are but a handful of questions, but through analysis and careful listening, you can add this final level of processing to make your mixes sound better than ever.
All of which begs the question: is DIY mastering advisable, or should you have your tracks mastered by a dedicated professional? As with all aspects of music production, if there is an opportunity to outsource a stage of the process to someone with a more established skill set, this is always a good idea. Just as working with a session singer might bring about a better result than you singing yourself, employing a mastering engineer brings huge benefits. One of these has access to a wider range of equipment than you're likely to have yourself—including expensive hardware for handling EQ, multiband compression and limiting, which may well bring additional sonic richness. A fresh pair of ears at the mastering stage is also beneficial. But most of all, if you hire a mastering engineer, you're hiring someone who takes this process very seriously, and who listens to all sorts of music day-in and day-out.
Of course, this costs money, which is often the sticking point. This feature focuses on what settings might be appropriate, assuming that no budget is available for mastering and, perhaps just as importantly, if your tracks need this extra stage of completion more quickly than a mastering engineer can deliver. If you've prepared a fresh mix and you're DJing tonight, it might not be possible to enlist mastering services and get your track back in time, in which case this process should help your cut fit into the mix. Of course, it's rare to find a commercially released track that hasn't been mastered by a professional, so whenever that opportunity presents itself, it's to be grabbed.