Part of the appeal of Reason is the idea that you can create tracks from start to finish in one integrated environment. However, for a while the program struggled to compete sonically with the final mixes possible in packages featuring bundled 'mastering' plug-ins. Reason 3 addressed this shortcoming with the introduction of several high-quality processors, meaning that now you really can go from scratch to a polished final track in Reason.
Last month, we looked at how to improve your mixes in Reason, and mastering is, in some ways, an extension of this topic. However, mastering is also a specific discipline that exists for two purposes: tweaking the final stereo mix for creative reasons, and optimising the output to suit the final delivery medium. These days, the stereo mix will either be destined for a CD or MP3/AAC delivery format, both of which will usually be created from a 16-bit, 44.1kHz stereo audio file, so mastering aims to contain your mix within the limits of this format. In the days when vinyl was the main music format, mastering was particularly aimed at the technical limitations of the format, such as stopping the needle from jumping out of the groove! Nowadays, when most musicians talk about mastering they tend to be thinking about working on the sound of the mix, rather than its technical aspects.
Mastering involves listening to and altering several attributes of the final stereo mix signal. The frequency composition of the mix is scrutinised, and EQ may be used. The spatial (stereo) characteristics of the mix may be altered, to create space or improve definition, and for technical reasons. Finally, level and dynamic range are massaged to make the most of the technical limitations of the final medium, to suit the type of music and to create a consistent, polished-sounding product. Each of Reason's mastering processors takes on one of these aspects of the mix.
The devices we're going to look at are the MClass Equaliser, Stereo Imager, Compressor and Maximiser. These devices can be used in the rack like any others, but in the context of mastering they should be inserted between the main mixer's stereo outputs and the Hardware Interface device (see screen, right, for the rear-panel cabling). If you wish to use all the devices, you can choose to create the MClass Mastering Suite (see screen at the start of this article), which is actually a Combinator patch that includes all the MClass effects, cabled and ready to go. The Mastering Suite also lets you play with some preset mastering patches to get an idea of what can be achieved.
The MClass Equaliser (see screen below) is a much smoother processor than the older PEQ2 EQ device, and is ideally suited to making adjustments to the final mix. The device offers low and high shelf EQs, two fully sweepable parametric EQs and a low-cut filter. Each EQ stage can be switched in or out with the small red button next to its name.
The first thing to say is that you may as well leave the low-cut filter switched in. It simply rolls off everything below 30Hz, which is almost certainly only going to be nasty stuff that you don't want to retain from samples. By far the most useful EQ modules are the two parametrics, as you can configure exactly how they work. They're your main tools for identifying and solving frequency-related problems in your mix and they have the following main controls:
The Frequency control, which centres the frequency that the EQ is boosting or cutting.
The Q control, which determines the range of the effect, from a wide area to a reasonably narrow peak/notch.
The Gain control, which provides +/-18dB of boost/cut — way more than you'll ever need.
The next issue is what to do with them. First, have a good listen to the mix you're working with and try to identify anything that is niggling you. This could be annoying boominess, boxiness, muddiness, resonance, sharpness, harshness, or lack of clarity or presence. These are all words often used to describe problems in the frequency content of a music mix, and they tend to be associated with certain areas of the frequency spectrum. If you've never done this type of listening before, now's the time to learn how — it's a lot easier than you might have been led to believe. When you've identified once which frequency band is associated with a particular problem, it will always stand out to you in the future and you'll have an idea what to do about it.
Listen to one of your songs and enable one of the parametric EQs. You might already be able to hear some areas of the spectrum that poke out, especially if you're comparing with a commercial CD (see 'The Other Suite Of Four Mastering Tools' box).
Whack the Gain control up on your enabled EQ.
Sweep the Frequency control slowly up and down until the offending area sounds much worse. What you are trying to do is find where a problem is and exaggerate it, which makes it really obvious and gives you confidence that you really were hearing a problem.
Now turn the Q knob so that you're only boosting the 'nasty bit' and not its surrounding areas, and pull the Gain control down, so that you're cutting slightly at this frequency. If your mix is fairly good in the first place, you should only have to cut by 2-3dB to smooth things out. If you're cutting areas by much more than this — say, 6dB — you should consider going back to an earlier stage of your mix and trying to sort out why there is an annoying peak or clutter at this frequency. Normally, EQ is used to cut frequencies, but you can boost an area if necessary. (See 'What's The Frequency, Kenneth?' box for some useful EQ starting points.)
Now that we've basically covered the use of the parametric modules, what about those shelves? The high shelf is pretty useful for giving a little boost to mixes that lack high-end life and sparkle, but the low shelf needs to be used very cautiously. It can be employed to roll off very low-end bass, although you're probably better off using the more precise parametrics. The problem with the shelf is that unless you have Q at minimum, there is a frequency boost around the cutoff frequency when you're cutting. This is pretty normal for an analogue-emulating shelf EQ, but this bump tends to boost the nasty boomy and boxy frequencies between 200Hz and 500Hz.
The last thing to say about the MClass Equaliser is that it doesn't have an output gain control to counteract any overall level change caused by your treatments. This means that changing the EQ could alter the response of the Compressor or Maximiser that comes afterwards in the Mastering Suite. Adjust the input gain on the Compressor to counter this.
Mastering For The Masses
You're probably waiting for the bit where I say that all these mastering tools are no substitute for a man in a professional mastering studio with 10-grand speakers, 'golden ears' and a beard. Certainly, if you're releasing a commercial CD this kind of highly skilled post-production is a valuable investment. But there are great reasons for all producers and musicians to learn to include mastering in their studio work. For a start, lots of music is now released non-commercially or independently, via the web or podcasts, and these tracks will not get the benefit of pro mastering. Secondly, many engineers will run their mixes through a maximiser to reference how their mix will sound when it's been mastered. Most importantly, if you are mastering with the full mix available to you, you have many more options than a mastering engineer working on a two-track mix. You can take what you learn from listening and tweaking at the mastering stage and go back and make changes to the mix, rather than just processing the stereo signal.
The MClass Stereo Imager (below) lets you tweak the stereo width of your mixes in quite a sophisticated way. Width can be narrowed or widened separately in two different frequency bands. A crossover control sets where the high- and low-frequency portions of the mix are split, and you can solo each portion for reference.
Typically, this processor is used to ensure that the low-frequency parts of the mix are close to mono, which keeps the mix tight and is also technically advisable for playback on stereo speakers, and for vinyl. The higher frequencies can be left at their original width, or even widened, if desired. Don't overdo widening, as this can just sound weird and chaotic. The Stereo Imager splits the mix into two frequency bands that have no crossover, a trick that can be used for other purposes than those intended. For example, check out the Mastering Suite preset patch 'Dual Band Compressor', which sends the high and low components separately to two compressors.
What's The Frequency, Kenneth? Some EQ Guidelines
Low, bassy boominess usually occurs around 100-250Hz. Try cutting here, and also compressing the instrument(s) responsible. This will make the low-frequency sound sources punchier and more defined.
A boxy, 'roomy' mix will benefit from a cut at around 350-600Hz. This can make a mix sound very tight, but if overdone may make it too 'dry'.
Muddiness can be addressed by reducing peaks anywhere between 600Hz and 1kHz.
Cutting at around 1kHz can reduce sharpness, and make mixes sound more powerful, but can also be overdone. If this happens, the mix may sound artificially processed, like the effect created by the 'loudness' button on a dodgy hi-fi.
Shrill and piercing sounds tend to live at around the 2.5kHz mark, so a cut here can smooth a mix. Conversely, if your mix is a bit flat and limp, a boost here can liven things up.
Mixes that lack presence, shine or the sought-after 'air' might need a subtle boost between 3kHz and 6kHz.
Compression, limiting and maximisation are all provided for in the MClass Mastering Suite. A decent mix will have made good use of compression before the mastering stage, but a little extra on top from the Compressor can help bind the mix together. Try using a medium threshold and a ratio of 1.5:1 or less. As I said last month, the MClass Compressor is probably more useful as a general-purpose tool in Reason than as a mastering processor (see screen below).
The MClass Maximiser (bottom of page) is where most of the magic happens, and where you can get the results that most people probably associate with mastering. The Maximiser does three things (which we'll look at in more detail in a moment): it boosts the overall average level of the mix; it prevents the signal from clipping (hitting 0dB); and it adds some subtle 'analogue-style' distortion.
Maximisation is achieved in the Limiter section. The technique behind maximisation is limiting peaks in the mix so that you can push up the quieter sections, increasing overall loudness. This doesn't mean turning up quiet sections of a song and turning down the loud bits. A maximising limiter compresses the transient peaks within the signal from moment to moment. This means that sections of the song that were already peaking at maximum can be turned up. The brain doesn't notice if fast peaks are limited, so if you don't go overboard the mix can sound a lot louder, while peaking at the same level and without sounding more compressed. The screens on the right show two waveforms of the same short section of music. The first waveform is the finished mix, but with no maximisation. The second waveform shows the mix with heavy maximisation. Both waveforms peak at -0.2dB but the second sounds a lot louder. Applied at this extreme level, maximisation can be heard: the mix will sound as though it's constantly flat out, straining at the edges and almost breaking up. This is the kind of limiting used on most commercial music radio, or TV commercials. More subtle maximisation, however, gives a louder mix with no perceived loss of dynamic range.
Looking at the Maximiser (see previous page), the signal flows from left to right. The Input Gain control effectively pushes the limiter harder and harder, so this is the main control for setting how loud you wish to make your mix. However, there is an Output Gain control, which can then push the peak-limited signal further into the Soft Clip section. Any boost you give the signal at the Output stage is basically clipping the signal (going over 0dB), but the Soft Clip stage is transparent enough that you can squeeze a few extra dBs through, if you really want them. The Attack and Release controls really need experimenting with for each song. Often, the slower you can get away with setting these, the more transparent the results. Set them too fast and the mix will pump and sound compressed. Set them slow and the sound is smoother, especially if you've managed to achieve a pretty loud mix in the first place. If sharp peaks (such as drum hits) suddenly poke through, you'll need to use a faster response speed.
The MClass Maximiser can stop your mix from ever clipping in two different ways. Firstly, if you set the Attack Speed to Fast, and enable Look Ahead mode, the mix will be prevented from clipping (provided you don't turn up the Master Output above 0dB). Look Ahead mode imposes a 4ms delay on the signal, which the Maximiser uses to get a head start on any sudden transients. This stops any sharp peaks from 'surprising' the limiter and getting through. The second way in which the device imposes a so-called 'output ceiling' is by using the Soft Clip stage. If you switch this in and set the Soft Clip Amount knob at minimum, the signal will not 'break the rules' by hitting 0dB, but will be able to 'flatline' if pushed hard, which amounts to the same result sonically. However, if you bring up the Soft Clip Amount, the top few dBs of the signal will be rounded off, more as if you were overdriving an analogue device. Unless you really push the mix, this will sound like a very subtle warming distortion.
The Other Suite Of Four Mastering Tools
Reason provides four MClass mastering processors, but the other 'big four' are ears, reference CDs, speakers and headphones. The best way to create a mix you'll be proud of, and learn a huge amount at the same time, is to stop and listen to some CDs that have a sound you'd like to get close to. Listen via the speakers you're using with Reason, and also with headphones. Try to listen from a technical point of view, especially for things that you might find surprising. How loud is the mix, and is it louder when there are more parts playing than just one or two? Can you actually hear any compression working in the mix? How much low end is there, and how does the overall frequency content compare to your mix? What else is different, and what might you try to make your mix sound similar?