Hi All. Mastering is something I know only about in the “hey, this sounds BETTER!” way, as I don’t know how to do it, so I asked Eric Oehler, who has pretty much mastered everything I’ve released since Booze Up and Riot and who’s not only an amazing musician but a really great tech geek (in the best sense of the word). The same blog can be found at http://blog.nulldevice.com/2009/08/mastering-101/
Part 1: Mastering – What It Is.
(paraphrasing what I’ve got on submersiblestudios.com)
Mastering is that final stage before going to disc (or tape, or TV, or whatever). It’s often spoken of in hushed tones, for fear that the Wizards of Mastering Engineering might overhear and smite thee with a curse. Everybody’s told they need it, few are told why, and even fewer really know what’s going on.
In reality, it’s less mystical and magical than one would think. It’s really just a form of audio processing done to add that final polish to an album. Of course, that’s just the basics.
It’s a finicky process, that requires good ears, good gear, and practice.
So why is it important for an indie artist? The common answer is “it makes things sound better.” This is part of the equation, but not all of it. The end result may even sound “better” before it goes to mastering – but the mastering engineer makes it sound good on a whole range of systems. A track may sound awesome on a set of studio montiors, but sound like complete pants when played through a car stereo or a TV speaker. Mastering engineers polish things up so it sounds as good as possible everywhere, from a giant club PA to a tiny radio. Sometimes that means a sacrifice on one system or a another.
Part of the process, of course, can be sweetening and shining the sound, cleaning up some rogue frequencies, making sure every track on the disc is reasonably consistent, etc. Usually, the process of making it sound good everywhere makes it sound better overall.
Part 2: Mastering – Why You Need It
Simply put, it makes things sound professional. Since you have very little control of the environment in which your work will be heard, having someone with some experience in optimizing your tunes can be tremendously helpful.
Mastering can make a well-mixed song sound great. It can also make a poorly-mixed song sound slightly better. It can’t fix everything – it’s not magic, despite what you may have heard – but it can glue things together sonically, make transitions between tracks on an album flow better, and clean up some things you might not have been able to hear. It also puts an extra set of rather finicky ears on your tracks, which can help immensely. Sonically, a mastering engineer can do a lot for a recording that you can’t necessarily do by yourself.
There’s also the element of the mastering process that’s not always discussed when people think of mastering – burning the master CD. While yes, you can make a master recording with iTunes or RecordNOW or Nero, a real Redbook CD – standard master has PQ codes, ISRC embedding, CD-Text, crossfades, etc. A good package can create DDP images and Edit Decision Lists for a replication house, too – not strictly necessary but often times they increase the reliability of the replication run and save everybody time and money. It’s not a huge deal for a demoes, but before replication of an actual album, it’s a nice thing to have. All those little extras can provide benefits on a variety of systems – and in some cases, for sales tracking.
In short – you need your stuff mastered because it makes everything sound polished, professional, and consistent. It brings your recording out of the studio and into the real world.
Part 3: Mastering – What Goes Into It
Most of what mastering is, from a masterer’s standpoint, is listening. And, in some cases, watching (those spectral analyzers can see those subharmonics and high frequencies that a human can’t). But that’s also sort of a douchebag answer, because it doesn’t tell anyone what the process is like.
The problem is, the process is difference for every track. Like writing a song, or recording a song, mastering audio is unique to the source material. There are a number of tools, though, that are commonly used.
Harmonic Coloration and Excitation – I’ll start with the most contentious one. Most pro mastering houses wouldn’t put this in as a step. But then most pro mastering houses are already using top-of-the-line gear that will handle some of this for them. Digital recording can be a little sterile or flat, so the addition of a subtle amount of coloration – whether by the use of real hardware or emulation software – can sometimes liven up a recording. Some systems have this as in independent stage, others work this into the compressors and EQs by emulating high-end gear. It is, of course, best done subtly, since over-application can turn a recording to mud (which is why it’s contentious). Excitation adds some upper harmonics to give a bit of brightness to a dull recording, and can add a little extra “sparkle” to a darker track – but when overused can set a listener’s teeth on edge. Exciters were used a lot during the mid-80’s – witness the teeth-shattering treble on most of the Stock/Aitken/Watermen records. At any rate, coloration in any form is to be used only as necessary.
EQ – the unsung hero of the mastering process. Smoothing out that one peak that makes track 5 sound so completely different from tracks 1-4, taming some out-of-control subharmonics, adding a little “air” to the top end, EQ helps bring everything together. In the digital realm, linear-phase EQs are often preferred because, while they are processor intensive, don’t risk screwing up phase balance in a track. However, some of the vintage EQ plugins do a nice job in some circumstances as well. And certainly, various bits of hardware like the mastering EQs from Manley and others add a nice character. EQ has the added bonus of reducing some of the overall levels (when you’re pulling down a frequency, anyway, which is the preferred way to do it) which gives the masterer more headroom to play with.
Compression – the bus compressor is the most misunderstood and abused link in the mastering chain. Judicious use of compression can “glue” a track together by selectively squashing out some dynamics and levelling things out. It can give a bit of a pulse to a dance track. It can also completely destroy a track by sucking all the life out of it. Multiband Compressors are even more dangerous in this regard – yes, they can tighten up some sloppy bass or bring out the midrange a bit better, but a 4-band compressor just gives a person 3 additional ways to destroy the dynamic of a piece of audio. Worst is when a track comes into the mastering studio with a lot of bus compression already on it.
Expansion – compression’s antimatter. It does just the opposite – expanding dynamic range instead of compressing it. It’s not used much, but occasionally it’s needed to emphasize some dynamics. It can be an antidote to overcompression, but only to a point – it can’t replace dynamic range that’s no longer there.
Stereo processing – sometimes a recording is very narrow and everything is clustered in the center of the stereo field. Sometimes it’s really wide and everything sounds washed-out. In cases like these, stereo imaging tricks can be used to fix things. It can be very, very tricky – since screwing with stereo imaging often affects phase, and not in a good way. The last thing someone wants with their music is for the vocals to drop out when the club speakers are too far apart, or for the song to consist of just a bassline when played in mono. What will happen often to fix stereo imaging is that, instead of actually adjusting the stereo image, the mastering engineer will split the track into a band representing the middle of the sound field, and a band containing the sides of the sound field. They can be adjusted independently, then, and mixed back together. A tigthter middle can make an overly-wide recording have more punch, and a louder side channel can make a track sound wider and deeper without actually expanding the stereo field.
Limiting/maximizing – two sides of the same coin. The final stage is usually some brickwall limiting to keep he signal from getting peaky, with the added bonus of increasing the perceived loudness of the track overall. Again, though, as a type of compression, limiting removes a lot of dynamic range from a track, so in addition to making the thing seem louder, it removes the quiet parts necessary to keep “loud” sounding loud. The human ear adjusts to constant loud noises, so without the contrast of a dynamic range, “loud” starts sounding “flat” after a while. That said, limiting and maximization are necessary parts of the mastering process, as they help even out albums-worth of material, add some consistancy, and give everything that room-filling sound. They’re often abused, even by big names, at the behest of artists or producers who want their track to be the loudest thing on the radio. That, however, is a separate discussion.
Dithering – much of what goes on in the mastering process occurs in the digital realm, at bitrates higher than the standard 16-bits of a CD. A straight conversion from a 24-bit recording to a 16-bit recording will likely sound fine, but adding a little bit of dithering – a process of algorithmically selectively adding certain kinds of noise to smooth the edges of a harsh downsampling – can make things sound pretty nice on a CD.
Metering – keeping tabs on levels, headroom, phase, clipping, RMS, peak, and spectrum is very important. A track may sound good right now but if it’s clipping or has some subtle phase problems that even the good mastering ears don’t catch, it may sound terrible on some system. The eyes see what the ear can’t. When I’m mastering, I like to keep various metering tools enabled between every stage of the process, so I can keep tabs on where anything’s going wrong in case my aging ears don’t detect a problem. Certain metering systems exist so that perceived loudness is baselined and can be kept consistent easily.
Testing, testing, testing – The real hard work. The mechanics of mastering are not just the technical tweaking of compressors and EQs. There’s an awful lot of critical listening. If you were to look at a photo of a real, high-end mastering studio, the first thing you’d see would be multiple sets of speakers – from tiny nearfields to giant, wall-mounted mains, and usually a set of midfields, and possibly even some battered old Yamaha NS-10’s. Since the goal of mastering is to make sure the track sounds good outside the studio, a wide array of listening systems are used to verify that, in fact, this club track doesn’t lose all the bass in a car, or distort like crazy through a set of crappy computer speakers. Those of us who do mastering on the cheap are usually limited to nearfields (since a good set of mains requires more money and installation time than any of us really have) so we’re forced to test by playing tracks in the car, or through the iPod, or whatever various systems we have handy. After a while, a masterer can develop decent instincts and know what’s going to pass the “car test” before actually trying it. Also, a good suite of the aforementioned metering tools helps, since they can see that, yeah, there’s a spike at 22hz that none of your speakers pick up, but a cheap subwoofer would make sound flabby.
You might notice that there is a common thread in all these tools – it’s all about careful application and critical listening. A lot of the more severe applications of effects that come so normally in mixing are usually inappropriate in mastering. EQ boosts and cuts are rarely more than a decibel or two, compression tends to stay on the mild side, and so forth. Mastering engineers also worry about signal phase a lot, since phase problems lead to cancellations, comb filtering effects, and muddiness. The whole process is just very fussy.
Of course, you may be thinking, “gosh, Eric, I’m already a super-fussy guy in front of my 64-channel Neve board, and everything is balanced and EQ’ed just so. Why do I need you to master this?” Well, you might not. A good mastering engineer also knows when to step back and say “yeah, this already works just fine.” Chances are, though, if you submit 12 tracks to the mastering engineer and one of them is great, the other 11 are going to need to be adjusted to match.
Part 4: Mastering – Can You DIY?
The conventional wisdom is currently that it’s de riguer to pay someone else master your tracks. This is primarily for two reasons – 1) you already think your tracks sound good, so someone else’s ears and opinions can correct things that you don’t notice and 2) people who do this professionally likely have better gear and acoustics than you do, specifically for this purpose, so they’ll be able to reproduce and test in a way you can’t. However, this being the era of both music-technology democratization and vastly reduced production budgets, it is entirely possible to DIY.
Most standard DAW packages come with everything you need for basic mastering in plug-in form. A compressor, a limiter, an EQ and a variety of metering tools. These are usually of varying quality, though, so you may wish to invest in better hardware or software. Regardless, these tools are not going to get you the same level of quality Bernie Grundmann studios is, since they have multi-kajillion dollar hardware and software for doing this. However, if you’re trying to keep costs down and this is really an indie release you’re selling at gigs or giving to indie labels, your own software is probably fine. There are economies of scale involved. Without the right tools and environment, you’ll never get that track mastered to sound as good as Coldplay’s does on the radio – but you also don’t have to pay the prices they paid to have that done. You can likely get close enough that it won’t sound like crap by comparison.
Where it falls apart for most people is in the arena of hardware. Many people kit out their home studios with fantastic software, plugins that sound like the best vintage analog gear, high-powered computers, super-featured DAWs – but then completely neglect to think about the listening environment or the output gear. If your monitors are crap (or even if they’re just “decent”), your I/O isn’t clean, or your room acoustics aren’t any good, then forget trying to DIY mastering. The tools for mastering are surgical, and having a insufficiently detailed listening environment is like trying to do surgery with a burlap sack over your head – you’re likely to do more damage than you fix.
This part of the equation can be costly, too. Good nearfield monitors are not cheap, nor are precision audio I/O systems. Acoustic treatment of a room requires time and effort (although it’s less expensive than one might thing – I have other posts on that). To really do things right, if you want to be all pro about it, your mastering room and your own mixing studio should be separate, and you should have midfield and mains monitors too, but that’s starting to get into the Grundmann teritory again. This is why it’s often more cost-effective to pay some guy a few hundred to master your stuff – it’s cheaper than buying new gear for it. There are a lot of low-end mastering houses, too – people with said gear who, while maybe not providing the fanciest of results, are still better at this than most and have invested in at least some of the gear.
If you’re fixated on doing it yourself, at a minimum you need good monitors and some acoustic treatment of a room.
You should also own Bob Katz’s book “Mastering Audio, the Art and the Science.” While Katz often gets off onto technical, beard-y discussions of jitter and compressor knees, his knowledge is still pretty indispensable. Also, he’s a vocal critic of the loudness war, and his K-system metering standard is an excellent reference system, since it is predicated on a calibrated listening environment.
What you SHOULDN’T do, EVER, is grab a cracked copy of Waves L2 or fire up Logic’s Adaptive Limiter and crank the gain knobs. It may sound awesome and space-filling for about 2 minutes but your ears will get tired and you’ll get really bored with it – and so will anyone who’s going to listen. This, unfortunately, has been the M.O. of too many people, leading to loud, squashed indie recordings that only perpetuate the loudness race, and piss off professional mastering engineers.
Mastering is really all about the ear and how it relates to the totality of the source material, so most of what you’ll be doing in the mastering process will be listening to a track, comparing it to another track, changing settings, and then doing it over and over again. A common myth is that mastering is mostly about making a track sound good – this is partially true, but it has to sound good in context. A great, pumping club-loud track is going to really stick out on an album of dynamic ballads, so even if the mastering job is good in isolation, it’s could sound bad in context.
There are also, of course, many all-in-one, master-your-tracks software packages. T-Racks and Ozone come to mind immediately. They’re all reasonably nice programs, but they do afford one the laziness of hitting “dance music preset #1″ and expecting everything to sound good. This rarely works (Ozone, for some reason, also includes a “mastering reverb” which seems like a Bad Idea) and often causes the end result to be a distorted, overcompressed mess. Certainly, this can sour a person on the whole concept of mastering, and the number of poorly self-mastered songs out on the internet is a testament to this. These tools are not to be dismissed out of hand, though – they often do have some very nice options hidden within. Ozone’s M/S processing is brilliant, for example. However, you can build a really flexible mastering chain out of a bunch of different pieces of software and hardware, to suit your own needs and workflow. PSPaudioware, Voxengo, Waves, Roger Nichols Digital, Wavearts, and many others make high-quality, mastering-chain software plugins. As for hardware, there’s all sorts of stuff out there, from Manley to UAD to DangerousMusic.
The biggest part of the DIY mastering chain (or even the pro one) is the one sitting in the chair by the mixing console or computer. The fundamental center of the mastering world is good, trained ears. It takes practice to hear a lot of things. It takes practice to compensate for your own high-frequency hearing loss from all those years as a club DJ. It takes relaxation – a tired listener means tired ears, and tired ears cannot perceive detail or loudness well. Dehydration, alcohol, caffeine – all these things affect listening, and should be taken into account during the process. Taking a break to rest the ears, to approach things fresh again after a while, can make a huge difference. Hell, it makes a huge difference when mixing, too, but it is really critical when doing the fine-tooth-comb work of mastering.
Note: Clint Sand of SYNNACK/MONOCHRONE/CUT.RATE.BOX also does mastering and his thoughts on the similar subject are at http://www.0xf8studios.com/index.php?page=faq#audio