In music, 'arrangement' is the art of making a piece of music work as well as possible for a given context or set of instruments. In the classical music world, this can mean taking a piano score and distributing the notes and dynamics across the individual parts of a score for string quartet. Great attention is given to finessing the performance markings in the instruments' parts, as this helps the arranger maintain the same internal balance between musical elements as before. In electronic music production, however, the tweaking of musical parameters to make the elements of a piece balance and work well in a given context strays into territory that we call 'mixing'.
Fundamentals: Notes and Voicings
Before you load up a filter/ EQ insert on, say, a synth line, consider checking whether the highest and lowest notes of that synth line are crossing into the note ranges of other parts. If the synth goes down low, or your bass line comes up quite high, their paths may cross. This will overload the region of the frequency spectrum where the cross occurs, and could lead to clipping. Rather than dialling in a gain reduction at that part of the spectrum, it is often better to change the notes being played at that point so that their pitches are further apart. Where possible, keeping a 'buffer zone' of around seven semitones between the parts allows them each to be heard clearly, and will require much less drastic processing, giving you a more natural sounding mix.
Harmonics: EQ Pocketing
Even with your notes kept within discrete ranges, two or more parts in your mix may still sound as though they are fighting for space. In this case, some low-/ high-shelf or parametric/ peak EQ applied to one or more of the parts may be in order. This works because the harmonics of any brighter sounds occupy frequency space extending far above the fundamental frequencies (or pitches) of your instruments' notes. For example, if your bass line has a full and rich timbre, like a classic dubstep wobble bass, its harmonics may mask the synth line just above it. Here, you need to use some judgement: do you reduce the volume of the higher harmonics of the bass line with high-shelf reduction, or reduce the volume of the lower harmonics of the synth line' A compromise needs to be found, but it is important to bear in mind which of the parts is more important. In the case of dubstep, it is likely that the bass line is more important, so go ahead and carve out some more space for it by applying EQ to the synth line. Using EQ and filtering to make space for all the parts of your mix is referred to as 'EQ pocketing'.
If EQ pocketing requires filtering that is so drastic as to adversely affect the sounds of your instruments, spatialisation of the sounds can help to prevent them occupying the same space. Whereas filtering can help achieve separation between instruments in the frequency domain, spatialisation (or panning) separates them by moving the sounds apart in stereo space. Rock music recordings, where there may well be two guitars playing the same register, often have the two guitars panned quite hard left and right, for example. This prevents their notes and harmonics from crossing, and masking their sounds as a result.
With all of the separation between parts achieved by tweaking the arrangement, EQ pocketing and spatialisation, it can start to sound as though the elements of a mix do not quite gel together into one musical 'scene'. At this stage, it's a good idea to start bus compressing. Separating out our parts in the frequency domain allows us to hear them clearly, whilst gelling them together with dynamics stops them sounding disparate, giving us the best of both worlds and a much cleaner mix.
Bus compression involves routing tracks that need to gel to the same compressor, so that they are compressed together. This means changing the tracks' outputs to a bus channel (or turning up the aux send all the way, whilst turning the track's fader all the way down), and setting up a return track for the submix with our bus compressor as an insert. Candidate pairs of instruments usually include drums and bass, mid-range synths and stabs, or high percussion and hi hats. The compression should be kept as light as possible whilst still affecting all of the musical parts going into it. A good starting point would be a threshold of -7dB, with a ratio about 1.7:1. The attack and release times need to be set depending on the material, but attack times of about 25ms for drums + bass, or 10ms for synths tend to work well. Release times could be about 75ms for drums + bass, and 150ms for synths.
See our Logic: Bus Mixing tutorial for more information.
A favourite technique in house and techno music production, side chain compression uses the compressor to carve out a niche for a deep kick drum whilst keeping the rest of the mix loud enough. This means that every time the kick drum is heard, the bass and other instruments are ducked in volume momentarily. The 'pumping' sound that this gives to bass when used fairly aggressively can sound great, but may not work so well if instruments with more mid/ high frequency content are also going through the sidechain compressor.
To set it up, you need a compressor which has sidechain functionality, also sometimes called a 'key input'. The bass line or submix return track has the compressor set up as an insert, with the kick drum sent to the sidechain input. Now, the bass line is compressed according to how loud the kick drum track is. Depending on whether you are using the sidechain compressor for audible effect or for functional reasons, you may need to experiment with the compressor's settings. Audible pumping can require quite a high compression ration, about 4:1, whilst simple gelling or getting a little more headroom is usually about 2:1. The threshold will depend on how loud your kick drum is, but anything from -10 to -2dB should work.
Our sidechain compression and sidechain filtering tutorial looks at this in a little more detail.