Let me give you a few guidelines and ideas for your own mastering. Bare in mind, this is not your ultimate guide to professional mastering, instead this might be interesting if you want to master your own tracks to play them in DJ sets or to give them to a friend. Also, for artists performing sets in Ableton Live, I want to give you an idea about applying a little "mastering chain" to your live set output.
Mastering, The Philosophy
"Mastering should be hair, make–up and wardrobe; it should not be plastic surgery!" [Peter J. Moore]
Cleverly said by Peter J. Moore and what it means is that mastering is not supposed to completely change a mix, the appearance of a track, nor can it turn something unexciting, magically into something exciting. Instead it is the icing on the cake to make the a (hopefully) good mix be ready for the music market or a target group, your your audience at a gig. To put it simply, that includes equalizing overemphases and resonances, controlling peaks for a reasonable dynamic range, optimizing the stereo image, potentially applying color (if that is appropriate) and finding the right amount of loudness.
The Chain
Turning the low end mono
Technical EQing
Peak control
Musical EQing
Color
Loudness
1. Turning The Low End Mono
Your sub region is not recommended to be spread out in the stereo field. If you put it in Mono, you center your sub energy and focus it on the "middle", you "make room" on the sides and most importantly, you will avoid phasing issues or detect them if you have any. You will detect low end cancellations due to low end signals left and right being out of phase.
If you haven't done it in mixing, now is the last chance to check if your mix holds up when played back in mono. Switch the complete track to mono to check if nothing cancels out and gets lost through phase issues, then of course switch the whole track back to stereo. You might think nobody plays his music in mono nowadays but that's utterly wrong ! Laptops, Phones, portable systems, the famous kitchen radio are very often mono speaker systems.
2. Technical EQing
Technical EQing is the most important part of EQing in mastering. There can be several stages of EQing throughout a processing chain but this one is essential. So over time you will realize that I'm not a person to say "you must" or "you can never", but technical EQing is fundamentally important, should not be overlooked and is also, by its purpose, limited to cutting frequencies away. This step serves to get rid of resonances or overemphases. More about this and in more depth in the article EQing applied with logic and reason... This step's sole purpose is to clean up the audio and in mastering and you don't clean up your audio by adding boosts, you clean up by taking parts away.
FOR EXAMPLE: This can very well mean to put a 3db or even 6db cut on a high peak in the low end to achieve a cleaner low end or more coherency overall. That is just one example but the key is being precise and surgical. A rule of thumb: Better apply thin cuts.
I recommend to put an analyzer after your limiter (which you want to deactivate for the file export), put a high resolution to the spectrum to get a visual feedback, to see potentially critical peaks. ALWAYS USE YOUR EARS TO MAKE SENSE OF WHAT YOU SEE AND DO !!!
How else can you identify annoying parts ? What I do is mastering "into my limiter". I have my processing chain as a template that I load up for mastering. I might switch a few things around but the basic processing stays the same. So load up the template, load in a track for mastering and the first thing I do is making the limiter work so that the track gets louder and I get a perspective of how it sounds when it's loud. Now I might be able to hear better what bugs me and what I need to equalize. If not, you can still sweep through the spectrum with your Audition Mode on, that is your little headphone button in EQ8. Check it out if you don't know what this is.
Especially in your last step before publishing a song/track to an audience, you want to apply quality control and make sure it's perfect so you have to make sure nothing is annoyingly overpresent. But if you realize that you have to make radical changes, it might be better to go back to the mixing stage to fix your issue.
Ideally you would use an EQ that has a linear phase setting. If you listen very closely, when you do a low cut for example, you can sometimes hear how non-linear EQs slightly alter your sound.
3. Peak Control
Before the signals go into your limiter, normally your last device in the chain, or into any further processing (which is the way I see it) you definitely want it clean and you might want to control a few peaks, specifically transients which don't necessarily add to the musicality but which could use up gain in your upcoming devices or that could affect the behavior of the limiter in the end. I find that this is more relevant with non-computer music or rather music that has human attributes to it, for example drums and other played instruments tend to not be perfectly played. You will see it is very possible that you don't always need this or don't always want this but I want to mention it since it can be a legitimate step in a professional mastering chain.
Use a compressor with a fast attack time that also supports fast release times. That could be your standard VCA compressor with adjustable timing and i recommend you start out with a clean sounding plugin. Don't just slap a random emulation of whatever on there, that can go WRONG !
Aim for very a very light gain reduction of 1db to 2db max here, you don't big time compression and especially not any negative side effects from the this peak control. Also make sure that your compressor is fast enough to catch only those high peaking transients that you want to get rid of and that it recovers quick enough to not affect the musically necessary information.
Very very likely that Opto compressor won't be your best choice due it's slow reaction time and if you don't want to add coloration avoid FET or Tube devices/emulations too. I recommend to use a compressor that has an internal high pass in order to let the low end pass unaffectedly, otherwise your low end running through the compressor might lead it to do work unnecessarily hard and reduce gain when/where it shouldn't do work and should not reduce the gain. This is particularly important with FET and Tube compressor as they can absoutely mess up a low with their coloration.
4. Musical EQing
Musical EQing is the opposite of Technical EQing, it is your chance to support and enhance part of the frequency spectrum. Slightly enrich a low end or slightly give more presence or more air, that would be three very common things to happen in this step. Now bare in mind I'm talking about slight changes here. If you have to make drastic changes it might be time to go back to the drawing board and change the mix in the mixing stage. I don't want to put specific amounts here but a +6db or +9db boost is considered rather drastic in mastering...
Don't make boosts just out of principle, make sure they are necessary, reasonable and constructive, respectively non-destructive. A rule of thumb: Better apply wider boosts, contrary to thin cuts, they sound more natural or less identifyable.
Make sure to use an EQ that has a linear phase setting. If you listen very closely when you work on the low end, you can sometimes hear how a non-linear EQ can alter your sound.
Nowadays you can work in an m/s setup with many EQs which allows you work on "deficiencies" on the mid and the side separated. I only want to throw this into the mix so that you know it exists and that you can break your brain thinking about ways to use it :)
5. Color
If you want the signature touch of a specific device to color your audio, this is your last chance. I recommend to only take action if you know what you are doing ! Saturators, Tubes, Tape emulations they all work on your transients, your low end, your air band and could potentially make things worse. Signature devices or emulations could have been used during the production and mixing and very likely, if you master for a client, mastering might actually be the time where you want to work clean, especially if you are new to mastering. Sorry if that sounds disappointing and anti-climactic !!! I definitely want to encourage you to experiment with how your signals can be affected by different types of saturation/coloration and experiment with and without bass heavy signals. It's just that mastering (especially professional mastering) is not the time and the place for processes of which you are lacking essential experience.
6. Loudness
You have probably heard or read about the loudness and how especially popular music is nowadays well pressed together to allow less dynamic range but therefor overall more loudness. The thought behind it (louder is better) might be a crime against humanity. You don't have to participate in this race. If you master for yourself and try to aim for a gain reduction of only 1db, to 2db to give you a reasonable amount of RMS, that can be completely fine and sound absolutely good and "alive" and "breathing" depending on your music and how much room it had before it went into your processing chain.
On the other hand, a gain reduction alone doesn't say whether something is limited to hard or not, 4-5db could be reasonable if necessary. Always use your ears too. These things become easier to judge with more experience.
You don't have to drive your RMS meter - a meter which you will want to use after your limiter (turn off for file export) - to ever hotter levels. If you reach an RMS (let's say RMS is your average level opposed to your peak level) between -13db and -11db to maybe even -8db dependent on the music genre/style, that can sound good and still be fine - again - depending on your source material and how dense it came into mastering. But don't squash it more and more until your level meter is hardly moving at all, that won't be a pleasing listening experience for you and your audience. If you go hard enough, at some point you can start to hear distortion from the limiter, you are heavily alterting your transients and you run the risk of robbing your music of its live, its feel, its rhythmic potential...and so on.
Processing for a live set
This one is tricky and I actually don't like to give you a formula because I don't think this will work equally well for everybody.
You probably have thought about this before and it's not easy to come with a working solution. When you play your tracks in a set, you want them to have a coherent and consistent feel to them.
If you play complete stereo tracks in your live set...
...I'll be last person to judge that, we all start from somewhere. But what you can do is normalize them when you export them, this is a feature for the Ableton Live export and it means that your tracks' levels will be adjusted towards 0db automatically so that its highest peak will reach 0db. It can't go further because that would mean clipping, that red DANGER signal. I encourage you to mix your tracks similarly so that when you normalize your music during export and their level goes up as high as it can, they all reach a more or less similar loudness character naturally. Then you have a good base for an EQ-Limiter combo with either a fixated setup or a bit of freedom to react to what is playing.
You might have heard of normalization before and how it is controversial, but it's a nice tool when used correctly so pay close attention:
If you load all your normalized tracks into Live, turn down their sample gain by 9db...that is setting it to -9db in case you're confused. Not your channel fader but your sample gain please !!! By that gain change we make sure to have enough headroom for potential EQ boosts. I'm going to assume you will do only light EQing on your live tracks because you have mixed them already, correct ? Do they need EQing ? No ? GOOD, then by all means don't EQ them ! The obligatory low cut and high cuts might be useful anyways to get rid of unnecessary informations in regions that won't be transported by the speakers.
If you want, above is a simple rack that I've quickly made with two adjustable low cuts for mid and side, and three bands for the mid signals and two bands for the side signals with adjustable gain but fixated frequency, which you could adjust to your liking before you start playing, you just can't fast-access them on the fly. Granted, you will be able to optimize this rack, this is a quick and general setting that you can take a fews levels up.
Before you play your set, you have to try out what amount of limitation (remember it will be applied to all your tracks in the projects) will be the best compromise. Turn up the limiter gain and play a variety of different tracks of yours into the limiter, which acts as your brickwall that will not let your signals pass the ceiling that you have set. -0.3db will be alright for this purpose. Let's do just a tiny bit of math: we have lowered the gain of your tracks to -9db so turning the limiters gain up by 9db will be a place to start from. Play various tracks of yours to find the gain that make the most sense, could be that you will end up with a gain higher than the initially set 9db. You will want to aim for a gain reduction of 2db to 3db...or maybe more, it's really up to you but don't do too much, don't squash your tracks, you're live audience will thank you for it!
Questions ? Shoot me messages or comment here.