Here’s a little guide I wrote for Ozhiphop.com last year. I don’t know why I never posted it on here but, well… here it is! Oh, and it’s long.
FG’S NOT SO COMPREHENSIVE MASTERING GUIDE
I thought I’d spread some love in here with a little how to guide for something that doesn’t seem to have been covered yet: mastering. I’ll try to keep it short and sweet but stuff can get technical with this sort of thing. If I miss anything out, please feel free to ask or add. (Edit: This actually turned into a massive essay – lucky work is slow right now).
Also, I don’t claim to be the greatest mastering engineer of all time but I have learned some stuff that could be useful to others.
Part 1: What to think about before you begin mastering.
- What is mastering meant to do?
- What are the important questions you need to answer before you start?
Part 2: Key Tools and Environment
- What are the basic bits of gear/software that you will need?
Part 3: Preparing a Mix for Mastering
- How to set up your software for a mastering project.
Part 4: The Effects Chain – Getting the Right Sound
- What order should you run your effects in?
Part 5: Metering
- What to watch out for on your meters.
Part 6: Addressing the Whole Project
Part 7: The Basic Mastering Process
- What is the basic process for taking a mix to master?
Part 8: The Technical Side
- Preparing for Reproduction a.k.a The Birds and the Bees
Part 1 – Thinking about what you want to achieve.
Mastering is all about polishing up your final stereo (or other) mix for final transfer to the product format (such as vinyl, cd, mp3, streaming on SoundCloud etc.). For that reason, I like to think of mastering as having two related purposes: a) to create a technically sound audio product which will not fuck up when it’s pressed and b) to produce the sweetest sounding audio within the technical constraints of the final format.
It is important to, first, think about how this sound will be heard. Is it going to be pressed to CD and listened to on hifi systems and car speakers? Will it be streamed from online and listened to through laptops and computer speakers? Is it going to be a freely downloadable mp3 that people will plonk on their iPod and listen to with headphones? These questions are important (and the usual answer is “a bit of each, really”) because the average predominant listening environment will shape creative decisions you make later. Also, different formats have different limitations with regard to the sound that they can handle (for example, vinyl can NOT handle high level bass and especially not loud bass with a dynamic stereo field).
Secondly, what kind of mastered sound do YOU want? Are you looking for something loud and in your face that immediately grabs people’s attention? Or are you seeking a greater purity of sound that will allow connoisseurs to appreciate the subtleties of your art? The material that you are mastering will in part, shape these decisions. If you’re mixing classical music then dynamic range and movement are of supreme importance (quiet parts should be quiet, loud parts should be loud). If you’re making a club banger then “loudness” is crucial (be aware though, loudness can destroy the integrity of your musical progressions – “soft” parts should feel soft even if their RMS levels are nearly as loud as the “hard” parts).
Part 2 – The Mastering Environment
All major DAWs have the ability to master. Some are better than others but a professional product can be achieved with any of them, even using the native plugins.
Your key tools (in terms of software/plugins) will be:
Compressor – These come in many forms. Essentially, a compressor reduces the volume of loud sounds and increases the volume of soft sounds, thereby increasing the average level of a track. If you’ve reached the mastering stage you should be relatively familiar with compression. A particularly specialised form of compressor that MAY be used in mastering is the Multiband compressor. This operates by compressing different frequency bands within a recording to different degrees (for example, you may compress everything below 100hz at 4:1 with a threshold of -32dB whilst leaving all other frequencies uncompressed).
EQ – This adjusts the frequency content of audio (e.g. increasing the 2kHz to 3kHz range by 2dB). It doesn’t hurt to use a Linear Phase EQ during mastering because it will help to maintain the integrity of your mix. However, the difference between digital non-linear and linear phase EQs will be in audible in any single pass use. The important point is that you will not introduce any phase distortion using linear phase EQ that may (possibly) impact upon the final product once down sampled and dithered.
Limiter – A limiter is often called a ‘brick-wall compressor’ because it simply stops any levels from exceeding the threshold that it is set to. Aside from that they allow for the introduction of gain, which will push the RMS level of a signal higher (causing more of the peak energy to be limited and increasing the perceived “loudness”).
Additionally, it’s always handy to have:
An Aural or Sonic Exciter: This emphasises the harmonic content of a mix to varying degrees and generally creates a more ‘lively’ sound which is perceived as more dynamic and interesting.
A Saturator – There aren’t that many great software versions of saturation plugins. Most are simply too colourful. If you have the resources you could push your mix onto two-track reel-to-real and run it slightly hot for some authentic tape saturation – beware though, not any reel-to-reel will do. You want a unit with true two-track heads (it will only lay down two tracks on the tape and plays in one direction only). Some “two-tracks” are actually four track machines that lay down a left/right tracks forward and separate left/right tracks backwards. Also, you’ll want a machine that runs at 15 inches per second (or faster, but that’s rare). Lastly, you’ll want a machine that has been professionally serviced and aligned. Any less than those provisions and it is likely that you’ll actually degrade the sound rather than improve it.
Metering tools: Your basic level meters are not enough. You will also need something to clearly indicate the levels of various frequency bands (an audio analyser) and something to plot the stereo spread of your mix (a goniometer). One of the most important tools in your mastering arsenal, however, is a phase correlation meter – this tells you the degree to which the left and right channels of audio are in phase with each other. If they are out of phase then certain frequencies will simply disappear/cancel out whenever the mix is played back in mono. This happens a lot in clubs and it’s important to know about any phase issues before the recording is pressed because it cannot be fixed afterwards.
Whilst software is important the most crucial thing (and the thing which leads to most great or terrible masters) is the mastering reproduction (speaker system) and “space” (room). The requirements for a good mastering suite do not differ greatly from a good mixing room but they are even more necessary for mastering than mixing. A good mix engineer will learn how to make killer mixes on a substandard setup just through years of experience and compensation for the issues in the room. A mastering engineer is only partly concerned about the “sound” of the music, they are also concerned with the technical issues and, most importantly, how adjusting sounds to be technically correct impacts upon the “sound”.
In a practical sense, this means that you will need better quality monitor speakers for mastering than for mixing (but don’t be put off – there are ways to compensate). And you will need a room that is not too live. The rule of thumb is, if the room sounds nice, just put a bit of absorption and diffusion up. If the room sounds terrible – go somewhere else.
Part 3 – Preparing a Mix for Mastering
Before you can master your tune, you need to bounce it as a stereo mix. Actually, that’s a complete lie. With modern DAWs you can put the mastering plugins on the Output buss and master without mixing. Here are three big reasons why you shouldn’t:
a) You are just increasing the load on your CPU by running all of your mix plugins and all of your mastering plugins at the same time. You can, of course, freeze each track and reduce this load in which case see point b.
b) Mastering on the output buss of a full mix leads to one thing and one thing only – fiddling. You will spend most of your time adjusting your mix and then adjusting your master and then readjusting your mix and so on. It is simply not efficient and it’s a false economy anyway. If your mix is nice then it will master nicely. If your mix is in need of tweaking to make it master nicely, then it is not ready to master. Keep the two stages completely separate – it will be frustrating at first but it will actually make you a better mix engineer and a better mastering engineer over time.
c) Mastering on the output buss allows you to handle the creative side of mastering but does not allow you to handle the technical side. Ideally, you should master all the tracks for the same album in the same session/project. That way you can not only make each individual track sound good but also adjust the balance and relationship between tracks to master a great album.
Technically, the quality of the bounce down for your final mix should take into consideration the end product. If it’s for DVD then you should bounce down at higher than 48kHz (standard DVD) or higher than 96kHz (DVD-Audio) at 32bit (floating point). This should be done on your first bounce down, as you will be bouncing down again once you have finalised your mastering process. The reason why you should bounce down at higher quality than your intended end product is so that the maximum frequency and dynamic ranges will be represented in the final product. Of course, this all depends on what standard you have been working at during the tracking and mixing stage. If you recorded at 8bit 11,050Hz then bouncing down at 32bit, 96kHz will not magically make all of the missing frequencies reappear – they simply won’t exist.
Lastly, if your DAW gives you the option to “dither” your bounce down then DON’T do this for your pre-master mix. Dithering introduces random noise to your mix. The aim of this is to prevent audible quantize error, which can result from the behaviour of particular processing and/or AD/DA convertors, and it is very important – but ONLY in your very final bounce down.
Part 4 – The Effects Chain
The only rule to effects chain is this: use the chain that gives you the best result. For a guide, though, that’s a pretty shit rule to outline. So here are a few suggestions from my own experience:
Limiting should always go last. Limiting has one purpose – to make things loud without allowing the signal to clip. It offers little benefit if introduced early in the signal chain as you will need to reintroduce it again later esp. if you’ve added EQ.
The general rule for compression and EQ is this: compress first and EQ second. Compression will impact upon the perceived tone of a sound as well as the consistency of frequency content. Hence, compression decisions will impact on EQ decisions, usually in a positive way. A 100hz signal that wobbles between -10 and -15 dB may lead you to reach for the EQ to address it (it will sound muddy). If you apply a multiband compressor to that same frequency range you will reduce the “wobble” and create a tighter, less muddy sound that may remove the need for EQ entirely.
On the topic of EQ, it’s important to note the different purposes of EQ in the mastering stage as opposed to mixing. EQ is used during mixing and production as a surgical tool. We use it to completely strip away entire frequency bands, narrowly slice out resonant frequencies (such as a kick drum thud in a sample), and accent the dominant frequencies of individual instruments (such as the narrow 80 – 110Hz boost we apply to kicks). Generally, hip-hop producers wield EQ with a level of viciousness that would scare the shit out of rock/folk/classical/pop music engineers.
Mastering is different. The old EQ adage of “boost wide, cut narrow” applies much more closely to mastering hip hop than mixing it. Additionally, if you find that you are consistently applying over 3dB of boost or cut to your mixes then you need to rethink your mixing process as you’re using mastering to compensate rather than sweeten. When mastering, a 1dB EQ boost should cause a significant change to the overall tone of your tune.
Using an aural exciter is not necessary but may be worth looking at. This should come just before the limiter. Aural exciters can help with produced music because they help to create a sense of “live sound” but they are very, very easy to overdo – and they sound absolutely terrible when they are overdone. By way of example, I will often use an exciter (the one in Logic 9 is simple but actually really good) but I set it to work only on frequencies over 15 or 16kHz (top end of the adult hearing spectrum) and I tend to add between 15% and 30% of harmonic boost. If you drop the frequency threshold you run the risk of making your mix sound “tinny”, and if you increase the harmonics then you run the risk of making the sound very tiring to listen to. However, every mix is different so, quite literally, play it by ear.
Part 5: Metering
Metering is an absolutely crucial part of the home mastering (and professional mastering) process. There are a few things you need to be aware of.
Firstly, when aiming for loudness you will be looking to increase the RMS (Root Mean Squared) level of your songs rather than the Peak level. Generally, you should set your limited to limit any peaks over -0.1dB. Through compression and increased gain on your limited you will be able to increase the loudness of the song without increasing the peak level beyond this point. Obviously, you should listen to the “loudness” but level meters will help you in your decision making. You will need to use one that shows both Peak levels and RMS levels. Generally, if your RMS level sits fairly consistently between -14 and -10 then you have a mix that will translate as quite loud.
That said, I have to clearly state my general distaste for loud mixes. Take (We Are Renegades) W.A.R., for example. An okay album in terms of lyrical content and the performance of the artists and so on but I can barely listen to more than three tracks in a row before my ears start to hurt. The beats are just crushed to death and the vocals sound way too shouty with way too much in the presence region (1.5khz to 3kHz). To me, and this may be personal preference, it just sounds like an unpleasant and unmusical mix.
So when looking to make your music loud I would very strongly suggest that you listen to how the sound is changing as you do it. It may be getting punchier with heavier bass but is the tonal balance actually pleasant? You can, and should, strike a balance between the RMS level and the tonal balance – it’s because they can do this well that the great mastering engineers get paid so damn much.
Aside from levels, you will also need to monitor the stereo image. As per usual, use your ears. However, home studio setups have notoriously bad stereo imaging (near-field monitors and small rooms = poor directional mixing) so don’t be scared to rely on your goniometer. Apart from stereo widening, there is little that you can do to introduce stereo range to a stereo mix (this is something you should have worked out during your mix). If you do add stereo spread make sure it does not affect anything below approximately 300 – 350Hz (higher if you can swing it). There is a lot of energy in the low end which will stand out like dogs balls if it is panned all over the place. Also, apply stereo widening lightly. You can be heavy handed with it during the mixing stage but while mastering it should rarely have more than 8 – 10% mix.
Technically, you will want to achieve an even Left/Right balance but this isn’t always as easy as it seems. The goniometer will visually represent the energy distribution between left and right channels. You want stuff happening on the left and the right but you want the average energy to be pretty central. Of course, it is more important to have it sound central than actually be central but sound like it’s to the left. However, you loudness seekers will be able to get a couple more dB out of your master if the energy is distributed evenly between left and right channels.
Part 6: Addressing The Whole Project
As I said earlier, you should attempt to ensure that you have developed comparable tone, levels and “feel” across all songs in a project during your mastering sessions. The most important of these qualities is “feel”. Songs are, and should be, very different. They create different feels by emphasising different qualities within a composition – some songs rely on dynamic range, others rely on pronounced melody, others rely on a wall of sound, others rely on complex tonal relationships between various elements or simple separation of distinct elements etc. etc. If you are mastering an LP, for example, you will be torn between two different necessities:
1) To create a consistency of sound across the masters for the entire album and avoid certain songs jumping out or sounding like they don’t fit.
2) To maintain the essence of the song by mastering it in a way that suits that particular song.
The only way to ensure this is to experiment and revise and modify your masters as you go. But there can be some logic to this.
Specialised mastering programs (such as WaveLab or Soundtrack Pro) are designed to take numerous stereo mixes, put them in order and apply separate effects to them. These programs really only offer the advantages of streamlining and some key features (such as the ability to burn at a Redbook standard and generate DDP data), however, any decent DAW can be used for audio mastering (though you will need specialised burning software to generate Redbook standard CDs).
Part 6: The Basic Mastering Process
1) Listen to your full album from start to finish and decide on the correct order of songs. Ask these questions:
a. Do the songs seem to move together well?
b. Are there any dissonant movements between songs e.g. chord movements that sound odd?
c. Will any of the songs need to be blended/overlapped? If so, where should the second track start from if some one should want to skip to that song on their cd player?
d. How long are the gaps that you want between songs?
2) Listen to each song and make notes:
a. Do you imagine this to be a loud song or a soft song?
b. What is the defining element of this song? Singing? Rapping? The piano melody? The kick drum?
c. Is there any thing in particular that jumps out at you as needing to be addressed.
NOTE: Listening objectively is a skill. What you really need is a real arsehole sitting beside you who will say, “Did you really think that splash cymbal sounded good or is this some kind of joke?” When we’re making/mixing music we tend to listen to the idea of what we’re making rather than the actual sound of the product itself.
3) Set your peak levels. Peak levels should be the same or, if the peak levels differ, there should be a good reason why (e.g. the low level one doesn’t have drums).
4) Apply compression and meter.
a. Each track will be compressed differently. Do NOT use a one-size-fits-all approach.
b. The aim of compression at this point is to “glue” a mix together NOT make it loud – you will be limiting later to achieve this if that is what you want.
c. Listen to the tonal characteristics of the song and the relationship between instruments in the mix. Experiment and choose the compression that makes the mix sound the most tonally balanced.
d. Just for good measure let me stress – you MUST differentiate between “well mixed” and “mixed loud”. If you compress at this point to make things louder then you will end up with a worse master rather than a better one. Assuming that your mix is already good, the aim of this stage is to increase the sense that it is “mixed well”.
5) EQ that shit.
a. Be gentle. Mastering EQ is a light touch. A slight boost (1dB or so) with a wide Q (from 80hz to 400hz, for example) on the top and bottom end is often all that is needed.
b. Don’t be scared to give a fairly heavy boost above 17.5kHz or 18kHz. You won’t immediately hear it but, if you do start to hear the boost, pull it back a dB or two. This will add “air” and a sense of space to the recording – something that is easily lost in digital recording.
c. Again, this is not the time to become obsessed with loudness – you are simply sweetening the overall tone. If you find yourself EQing to address issues related to specific instruments, samples or parts of the song then go back to the mix and correct it there. This stage is not where you fix mistakes.
6) Excite that shit.
a. Ask yourself – is this necessary? If the answer is “no” then skip this step entirely. If you feel that your mix is really nice, and your instruments are glued, and the tones are sweet but the track still sounds a bit flat THEN you should try some excitement.
b. Ask yourself – what makes this track flat? Does it have no sense of space to it? Is it really soulless in the bottom end? Does it seem to come from everywhere and nowhere? Choose an appropriate plugin to deal with the issue.
i. Airless top end – this may be helped by adding an exciter plugin which begins operating around 12kHz to 18kHz (experiment and listen). Take it to the point where it becomes noticeable then drop it back a fraction. Additionally, try LIGHTLY applying a stereo widener above 15kHz or so.
ii. Soulless low-end – try a saturator of some kind but be careful. These can actually suck low-end out of your mix if you are too heavy handed. The “soul” or “warmth” of a mixes low end will sit between 60hz and 400hz but beware of the 100hz and 300hz region (indiscriminately bashing away in these areas will impact on punch and low-end space or roominess respectively).
iii. No convincing sense of stereo field – if this is not a mix issue (which it often is) then it may require the LIGHT application of a stereo reverb and/or a stereo widening too. The emphasis here is LIGHT. Take it too the point of audibility and then drop it back a notch. If you aren’t sure whether it’s too much or not enough then listen back in headphones – any heavy handed stereo manipulation will become immediately apparent.
7) Limit that shit:
a. Look at your notes – is it meant to be loud? Spacious? Subtle? Remember those notes as you limit the mix.
b. Generally, your limiter will allow you to adjust at least the Input Level, Gain and Max or Output Level (the brick wall). So:
i. Set your output level or -0.1dB (or lower if it’s a quiet track)
ii. Increase your Input Level to bring the mix up to the maximum volume before clipping (watch the input meter, not the output meter which will not go above -0.1dB)
iii. Increase the gain until it sounds good and as loud as you want.
c. But what if it’s not loud enough? Well, this is usually caused by decisions you made in the mix and mastering EQ stages. Reconsider the decisions you made about bass instruments and frequencies in particular. Your mix should not sound like a master – if you’ve boosted hell out of the bass during the mix it will just turn to mud or distorted nonsense during the master before you can achieve the loudness that you want.
d. Alternatively, reconsider wanting it loud at all and jump on the dynamic range bandwagon.
8) Check the master on many different systems/formats.
a. This is where you need to think about the dominant listening environment. If the track will be mostly listened to on iPod headphones then listen to it a lot through iPod headphones as well as home stereos, car stereos and club/DJ PA systems. If it’s meant to be listened to online, listen to it through laptops and shitty PC speakers.
b. Make notes on anything that you notice which you don’t like.
c. Go back and gently adjust your mastering settings to address any issues and find a nice balance across all systems.
Part 7: The Technical Side
After a master is created, the next step is to send it off for reproduction. If this is just an internet release, then you need to do little more than bounce it to MP3. If you’re going to have it stream from the internet (through SoundCloud, for example) have a listen to some similar tracks on the site. Do certain types of masters sound better than others? Will you need to adjust your master? With streaming services it is often better to err on the side of louder but beware of distortion. Streaming will lower the bitrate and potentially introduce more noise – loudness can overcome this. However, distortion will also be extremely noticeable, partly because people will be listening on systems that emphasise the presence range of music (2kHz to 3kHz).
If you are looking to release a CD, you’re going to need a Redbook burning program or software that can produce either Redbook burned CDs or DVD audio with a DPP file. A Redbook CD is simply one that complies with Redbook standards and, thus, can be played on any ordinary audio CD system. Non-Redbook CDs will not be as compatible. Burning a CD to Redbook standard will allow the CD pressing plant to simply duplicate the whole CD as you burned it. This is great for limited run releases, however, it is prone to greater error rates than other methods and it takes longer to produce – but it will also be cheaper because it requires no extra work from the CD plant.
For longer runs of CDs (2500 or more) you will be better served to burn your mastered songs to DVD as uncompressed audio (.WAV or .AIFF files will do) and supply a DDP file with them. Basically, a DDP file (or more accurately, set of DDP files) is a bunch of data that is supplied with raw audio that includes information on starts points, endpoints, playtime, cross-overs, pauses and track metadata. I’d love to tell you more about this but, to be completely honest, it’s way beyond my pay grade so, if you go down this track, then you’re best to get a mastering engineer to help you or go to http://www.gearslutz.com and politely harass the experts there.