The Making of GunGirl 2 OST: pt. 6

Mastering, and preparing audio for in-game use is the usually last of the most important steps to preparing a soundtrack.  Without the proper polish, even the best music can sound dull, and besides, there certainly has got to be SOME reason people get excited about “remastered” albums.

When I finish putting together a track for a game (or any track, really), I’ve got a well-mixed version that I wouldn’t mind listening to, but I often have to crank up the volume or play with it a little bit to be able to jam rightfully.  Additionally, if I sent this track to a game developer, they probably wouldn’t be very happy with how the track looped in the game (assuming it to be used in a looping manner, of course).  So how do you fix this?  Well, have a listen and follow the jump to learn more!

Right at the beginning of the track “Oblivion,” you should hear a pretty obvious soundscape preceding what feels like the first beat of the song, combined with a crack of thunder.  However, you won’t quite hear this same start in the game – instead, the song starts right up on that first beat.  For even the simplest loop to sound correct you have to align your seams very carefully, a trick which is pretty easy to pull off.

When I’ve got my song finished in its premastered form, I bounce the entire thing to a separate WAVE (or sometimes to be efficient I bounce it to a track within my DAW that I put on solo and apply my effects chain to).  Either way you work is fine, but you should have your premastered uncompressed audio, including about 4-5 seconds of trailing audio (or whatever it takes to capture your ‘reverb trail’ – the trail of sound that comes naturally after your last notes that include delays and reverbs, or natural sample releases like you’ll find in EWQLSO).

To make an efficient loopseam, you should make sure you stick pretty close to your bar-grid and split the audio at the first beat of the measure that follows your last measure of music.  Then, to create your in-game loop, push the start of your reverb tail, or the second half of the split audio, to the very beginning of your audio.  You’ll end up with layered sound of the beginning of your track and the reverb trail.  Now, in many cases, you’ll find that when playing your song from the beginning the reverb trail is actually a noticeable sonic entity and is really quite gross.  This is where you may find yourself tampering with the time that you cutoff – sometimes I have to add some ambient effects or change last music at the end to get a proper loop, but either way you do it you have to be extremely careful and pay close attention to your loop.

Marsh Hell (D2) is an early example of “Oblivion” that demonstrates how it appears in-game, with a more sudden and abrupt ending – you’ll find that it loops seamlessly.  The way I create the soundtrack versions (yes they are separate versions) of the music is by manually looping and finding a good fadeout point.  This can be tedious, and don’t say I didn’t warn you.

As for mastering your music to give it more of an ‘edge,’ there really is no single bit of advice that can be given.  Different genres of music call for severely different approaches, but I’ll go through a quick runthrough of my mastering chain for many of the GunGirl 2 tracks.

I start with some slight EQ, to add a little sparkle, boost the bass if needed, and turn down the muddy regions in the 200hz area – however, when I say ‘slight,’ I mean it in the same way I said I applied effects earlier – I rarely boost or cut more than 1dB, but stay closer to .5dB on average.  Really, all EQing should be done per track before this point.  A high-pass filter at 20hz isn’t a terrible idea, and cutting everything about 25,000hz isn’t an awful idea either.  (Purists will tell you 20khz is the high-end to cut at, but I also like to have a little elbow room)

I follow with dynamic multiband compression in subtle amounts – I don’t like to overcompress and using multiband compression can often skew your previously sparkly mix – I usually find a preset I like as a good start and toy from there, and once I’m happy, again, I crank down the mix % of the compressor.

Stereo imaging and harmonic exciters are also in play, but as I’ve also said before, generous panning and EQ before the mastering stage let’s me avoid cranking those too high.  What I -do- recommend is removing all stereo width below your 80hz subbass frequency, to really give tight kick and bass sounds.

The last step is my favorite step, and I admittedly probably abuse this a little: The limiter / loudness maximizer.  It’s smart to cap your mix around -0.3dB or -0.2dB, but never, EVER, EVER set your margin to 0.0dB as that will register as clipping to a lot of editors (like Audacity).  And like I said already, ELBOW ROOM IS GOOD.  You can set your threshold to whatever you like, and it will vary per mix, but make sure you end up with a consistent sound across your album.  Make sure your using as transparent a limiter as possible, because you don’t want to know it’s there – I prefer Izotope oZone for pretty much all of these steps, but there are plenty of free ones that can give you a decent sound.

Last but not least is dithering, and I pay careful attention to letting my mastering suite do it instead of my DAW (which means I export with NO dithering from the DAW and dither in my mastering chain).

There are more than plenty of good resources on mastering on the web, especially if you’re going to be using oZone, so I encourage you to read around in multiple places before you get your feet too wet, but then again, most learning is done by doing.

Next time, I’ll be giving away all my secrets to distribution, and tell you how it’s possible to make a little money off of even a freeware soundtrack.

About the author