These days, many of the production tasks we carry out when making music blur into one seamless process. Once upon a time, creating a track would have involved separate musicians, programmers, engineers and mixers, as well as a dedicated mastering engineer, to say nothing of the producer overseeing the whole project. Nowadays, programming and production are one and the same, and as we load channel strips of effects to help steer sounds as we work on them, mixing becomes a consideration from an early stage. So it's no surprise that mixing has incorporated much of the mastering process as well, since many of us begin to put effects on the output channel long before our track is officially mixed. The desire to hear our music at pumped-up, maximized volume is overwhelming, and if output channel processors get us there quickly, then why not use them in the programming, production or mixing stages?

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.

Mix to master: When your mix is done, drag the stereo bounce into a new, empty session within your DAW and place it on a new track. Ensure that the output level of the master track is at 0 dB—the processors you'll use will ensure that levels are optimised, rather than anything you do with the faders, so leave both the source track and the output faders at zero. Here's where we're starting.

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.

Mix with stereo compression: As you can hear from our track, it's only subtly different after we apply stereo compression, with the ducked peaks dropping the overall level by 1 dB or so.

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.

Mix with limiter: That loudest transient in our mix is a little like the loudest moment in a door slam. If you could leave that where it was level-wise but bring the quieter moments closer to this peak, the average level would increase accordingly, producing a more sustained level of high volume. This is exactly what limiters do—they set an output level (sometimes referred to as an output ceiling) above which no volume increase is possible, and bring up the quieter bits by squashing the dynamic range. The input level will control how much volume you're going to add to these quieter moments, so the bigger the number, the more dynamically squashed your mix will be. Lots of mastering engineers take 0.1 dB off the output ceiling level, too, as the debate rages as to whether or not maximum 0 dB level represents the last volume step before digital distortion begins or the first step on that path. As no one will hear a tenth of a dB's difference, it's not a bad idea to drop the output ceiling by this amount. Here's our mix with a limiter increasing overall volume.

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.

Mix with EQ: We're setting up a high-pass filter with the cutoff set to 25 Hz, then adding a low shelf to roll off frequencies from 110 Hz by 1.5 dB. The bloated sounds in mixes often come from frequencies around 200 Hz, and we're adding two more bell shape filters to reduce these and frequencies below, at 100 Hz. None of these changes are hugely dramatic; unless the mix file is compromised by extremely loud levels in particular frequency areas, changes between 1 and 3 dB should be plenty.

Mix with multiband compression: It would also be beneficial to go back to the multiband compressor to enhance the performance of each of our separate bands. Even the subtlest adjustments here—we're focusing on the bass and low mid-range particularly—can make a pleasing difference.

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.

Overexcited: Exciters add artificial harmonic content to input signals, increasing the busyness in a specific frequency band and, as a result, "exciting" it. Be careful, as particularly after a long and concentrated listening session, the temptation is always to add more energy and brightness, and harmonic exciters will give you this by the bucket-load. As you can hear in this example, lots of extra excitation has been added to the treble end of the spectrum, and the result is horrible: tinny, thin and overly sharp.

Stereo imaging: This is the final bit of processing to consider. Stereo imagers work by spreading or scattering a group of frequencies across the stereo space, as if separate components of that frequency band were split up and panned anywhere from dead centre to hard left or right. Ozone uses a multiband approach, so it's possible to spread different frequency bands by different amounts. Again, if you've been careful to create a balanced bass sound, it's unwise to upset this by adding stereo imaging to low frequency content. However, it could be advantageous to set a negative value, which will make the bass end more mono and even. Otherwise, as ever, use your ears and decide whether there's a more impressive stereo image when upper frequency content is scattered a little more from side to side.

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.