Mix Rescue: Strong Arrangements

Photo: Chris Boland www.chrisboland.com

The foundation of a good, engaging mix is a strong arrangement. But how far can you go to improve things at the mixing stage?

When SOS reader Trevor Piggott recently sent me over a mix he was struggling with, I could hear that the sonics lacked some clarity and punch compared with the Bob Clearmountain mix he was referencing against (the Simple Minds song ‘I Wish You Were Here’), and that the lead vocal wasn’t commanding the listener’s attention enough. There was, however, another more insidious malaise, because he’d also fallen into a trap that ensnares many project‑studio users: relying too heavily on repetition, especially of the copy‑paste variety.

It’s an easy thing to fall into, and often goes like this. First you create a four‑bar pattern with maybe drums, bass, and some chords, and then you quickly copy‑paste that so you can crack on with writing a song over it. By the time you’re done, you’re beginning to get a bit bored with the bare‑bones arrangement you’ve heard looped so many times, so you begin adding more parts to freshen up the pattern. But while each new layer does re‑enthuse you at first, its novelty inevitably declines with repetition as you work, eventually leaving you with an arrangement that, despite being saturated with musical parts, leaves you with a niggling sense at mixdown that something’s still missing — no matter what processing or effects gizmos you try.

In this article, I’d like to share some of the practical arrangement and mixing techniques I typically use to address such issues, and show how I used them to rework Trevor’s production, upgrade the mix sonics, and reinvigorate his enthusiasm for the song.

Once Less, With Feeling

One of the first things I did with Trevor’s song was look for opportunities to shorten the structure. After all, if any kind of musical pattern gets staler the more you repeat it, why not simply reduce the number of repeats? As it happened, there was an eight‑bar instrumental section that was treading water between the second chorus and the onset of the guitar solo, so removing that was an easy win. But there were also two separate intro sections that delayed the arrival of the first vocal verse until the 30‑second mark. I chose to sort of fold those into each other, to get to the lyrics 10 seconds sooner. On a smaller scale, I also pruned out a few repeated sections on a per‑track basis, muting the bass guitar during the introduction, progressively weeding out more backing‑vocal layers for the earlier choruses, and removing the chorus piano hook from the guitar solo section (where the backing track was plenty busy already).

One easy way to differentiate the sections in your arrangement and provide more of a sense of build‑up through your song is to reserve some sonic layers for later in the timeline — as you can see Mike doing here with some of the chorus backing‑vocal tracks in his remix.

Simple cuts like that will only take you so far, though. Another more useful strategy is to modify some of the repetitions so they sustain the listener’s interest better. If you think about it, exact repetition is actually quite an unnatural thing in real‑life music‑making, because human musicians never really play the same thing twice — indeed, that’s an inherent part of the magic of the live gig experience! So in this case, when I realised that the main piano riff comprised three repetitions of the same rhythm, I decided to edit the second one into a slightly different shape. Similarly, the chorus’ main lead‑vocal melody comprised a pair of identical lines which I was able to transform into a slightly more interesting ‘call and response’ pairing by shifting the final note of the second line to a higher pitch. A melodic clean guitar riff that looped through the pre‑choruses and choruses was edited so that the pre‑chorus version became much sparser and simpler, which meant that the chorus iteration then felt like a musical development, instead of a straight copy. And there were multiple instances where I was able to mute individual instruments for a moment just to kind of remind the listener of their presence — most notably the drum kit just before verse one and during the first half of verse two, and the bass guitar just before the final choruses.

Difficulties with the click track during the tracking sessions had led the band to build their original drum track from short loops of the drummer’s playing, but this robbed the song of both short‑term musical variations and long‑term performance dynamics. To remedy this, Mike first reconstituted and synchronised the original continuous drum track using detailed edits, and then enhanced the section differentiation by thinning the track count at strategic moments.

And, of course, if you go to the trouble of capturing live performances, it’s a shame to squander their humanity by subsequently looping sections of them — which is what the band had felt obliged to do here with their live drum tracks, because of difficulties maintaining the song’s swung groove against the click during recording. Fortunately, they’d archived the original live take, so I was able to re‑import that into my mix session and use editing to deal with its timing issues directly instead — this meant I could retain all of the player’s nice little musical accents and pattern refinements. And, while I was at it, I decided to pare back the drum kit texture during the first 40 seconds of the song by muting the room mics and tom close mics and low‑pass filtering the kick‑drum and snare close mics, thereby giving more scope for the drums to build up through the timeline.

Stealth Layers & Arrangement Build‑up

Where regular editing methods don’t provide enough scope for introducing variation, a great alternative can be adding ‘stealth’ layers to supplement the existing parts. I used this tactic for the echo‑y triplet guitar chords that underpin most of this song, EQ’ing a stock Jazz Guitar patch from NI Kontakt to get a similar sound, then programming two different upper layers to subtly differentiate the part’s verse, pre‑chorus and chorus voicings. (I also reused the region‑specific low‑pass filtering dodge I’d tried on the drums, restricting the guitar’s upper spectrum early in the song to improve the long‑term dynamics.) The programmed bass part benefited from some layering too, with an added sub‑bass synth lending the line extra power for the later choruses and during the second half of the guitar solo.

One way of apparently introducing variation into a repeating part is to subtly layer a similar‑sounding MIDI instrument alongside. In this remix, for example, Mike used a Jazz Guitar patch from Native Instruments’ Kontakt to extend the upper harmony voicing of the song’s main echoey guitar loop, to suit different sections of the arrangement.

You have to be careful when adding extra parts at mixdown like this, because few musicians like to feel that you’re tampering with the essential musical material. So it’s wise to keep such contributions in the ‘subtle to subliminal’ range. But if you feel that a section of the arrangement needs something a bit more ostentatious, you can reduce the risk of a negative response if you create that new element by remodelling some existing recorded track, perhaps from a different section of the song. For example, I’d jettisoned a piano special‑effect track while whittling down the song’s introduction, but later I was able to use this as a ‘new’ atmospheric element to differentiate the second verse and pre‑chorus from the first.

This trick helped with the song’s second chorus too. You see, the third chorus had been bolstered with heavily distorted electric guitars, but there was no real sonic progression between choruses one and two. Luckily, Trevor had recorded DI signals for those parts, so I could generate less heavily driven versions of the guitars during the second chorus, bridging the ‘energy gap’ between the first and third. Similarly, moving an iteration of a clean guitar riff from the first part of the solo to the start of chorus five not only generated an arrangement ‘lift’ for the second half of the solo, but also for the second of the final choruses.

Mixing For Clarity

If you ask me, it’s hardly worth doing any real mixing until you’ve adequately addressed arrangement issues like these. What’s the point in getting bogged down in plug‑in settings before you’re able to judge sounds within their final context? So it was only once I’d resolved my repetition concerns that I turned my attention towards my first main mixing goal: achieving ‘clarity’, in other words making sure all the layered parts could be heard without the overall mix tonality becoming woolly or bloated.

There are lots of fancy ways to attack this but it’s important not to neglect simpler tools. Straightforward filtering is a great workhorse, for example, and it had an important role to play in this mix. High‑pass filters on the drum overheads, room mics, electric guitars, piano and effect returns really helped keep the low end clutter‑free for the kick drum and bass guitar, while low‑pass filtering helped remove abrasive upper‑spectrum masking frequencies from the distorted guitars, as well as pushing some of the harmony vocals and clean‑guitar riffs more into the background behind the (more musically important) lead vocal and solo guitar parts.

Another family of ‘clarity enhancement’ techniques essentially involves moving energy from overpopulated bits of mix real estate to more sparsely occupied regions. As with a lot of project‑studio multitracks, this project had an overabundance of lower midrange, a good chunk of which was coming from the bass guitar. So I cut that firmly around 140Hz, but then compensated for that loss of energy by adding more true low end (from the added sub synth layer) and boosting the midrange around 1.5kHz. To put it another way, I traded some low midrange (that the mix had too much of) for more sub bass and midrange (where the mix had greater headroom available). Likewise, there were two different tracks (the piano atmospherics in the second verse/pre‑chorus and the clean guitar riff first heard underneath the second half of the guitar solo) where I deliberately mixed in some octave‑upwards pitch‑shifting to move their frequency emphasis further up the spectrum and hence reduce their reliance on the low mids.

This principle of shifting between areas of the mix can apply to the stereo image too, and I used a few different methods to clear space at the centre of the stereo image for the most important arrangement elements (ie. the kick, bass, snare, and lead vocal). For mono tracks, panning is usually the first port of call, and I did hard‑pan the most heavily distorted rhythm guitars in this mix. But for stereo channels (such as the Hammond organ and the clean guitar riff subgroup) the simplest way of clearing the centre is to reduce the level of the stereo signal’s Middle component using Mid‑Sides (M‑S) processing. Occasionally, you may need frequency‑selective control, and I did here: I wanted to thin out just the lower spectrum at the centre of the echo‑y guitar’s stereo image, and in such cases an M‑S equaliser will likely be more suitable.

Don’t use the same widening tactic for everything, because every widening effect has its own potentially undesirable side‑effects.

To clear the central image of an instrument, an alternative to cutting its Mid component is to find a way to boost the Sides, and then turn the whole track down — either way you’re shifting the Mid‑Sides balance in favour of the Sides component. There are lots of ways to do this, from specialist insert processors (such as the freeware Polyverse Wider plug‑in I applied to Trevor’s chorus piano hook) to send effects (such as the widescreen modulated delay/reverb send effect I used to spread the clean guitar riff and piano atmospherics). My main advice here is that you don’t use the same widening tactic for everything, because every widening effect has its own potentially undesirable side‑effects (perhaps it makes the timbre chorus‑y in mono, or it distances instruments from the listener), and I think it pays not to compound them.

It may seem counterintuitive, but the imbalance between an instrument’s different notes can also cause a loss of mix clarity. For example, the echo‑y guitar part in this mix featured one overplayed note that overwhelmed the mix before any of the instrument’s other notes were coming through clearly. With MIDI parts, some swift reprogramming can remedy this, but with audio I often find that surgically notching out some of the offending note’s harmonics can achieve a similar result, making it possible to raise the instrument’s fader to a point where the rest of the notes become more audible. You can implement this kind of processing with most regular digital EQ designs, but I use it often enough myself that I appreciate those few EQ plug‑ins that offer a special ‘multi‑notch’ filter type that you tune to the target note’s fundamental frequency and then it automatically cuts a specified number of that note’s harmonics into the bargain — Voxengo’s Gliss EQ, for example, or Melda’s freeware MEqualiser.

If there’s one note in a musical part that dominates over the others, that may prevent you from fading that channel up enough to hear the other notes with sufficient clarity. One solution to this problem is to use a specialised multi‑notch equaliser (such as the Voxengo GlissEQ plug‑in Mike applied in his remix) to rebalance the errant note so that all the notes can come through more clearly.

Again, it’s easy to get carried away with more complicated processing hacks like these and forget that probably the best all‑purpose clarity‑boosting trick is simply riding the channel fader! It might seem a bit low‑tech to just push up a track’s fader when it’s worth hearing and pull it down again when other tracks are more important, but it’s a tried and true technique — and I used it all over this remix, especially on the guitar and piano.

Powering Up The Drums

Another important part of this remix was trying to increase the power and punch of the drums, which is something I’m often asked for tips about. To state the blindingly obvious, the first thing to do is turn them up! The rub is that fading up any instrument in a mix often shines an unwelcome spotlight on performance inconsistencies that need sorting out. For example, I realised from comparing Trevor’s mix against the Simple Minds reference that the kick and snare needed greater prominence in the balance, but neither of those drums felt consistent enough in level that I could simply push up the faders — some hits always ended up feeling too strong or not strong enough. So one of the first things I did was limit them both. Limiting might seem an extreme choice, but I wasn’t triggering more than 4dB of gain reduction and I set a fairly long 450ms release time, so the processing was really just levelling out the performance hit‑to‑hit rather than reshaping the drum envelopes.

Sharpening the attack of your drums can also help them cut through more, and most DAWs now have transient processors that are great for this. In this case, I used one built into my Reaper DAW for the kick drum. In some situations, though, the frequency‑selective control provided by some third‑party plug‑ins can be handy. When I applied full‑band transient processing to the snare in this mix, it seemed to add more of a sharp ‘edge’ than a meaty ‘punch’ (if you’ll forgive the wine‑tasting terms...), so I was happy that I could turn to iZotope Neutron’s multiband transient processor to boost the drum’s sub‑200Hz region independently.

Upper‑spectrum EQ boosts are frequently used to emphasise the forwardness and aggression of drums (I boosted 9dB on the kick drum, for instance), but it’s important to realise that EQ can also undesirably soften your drum transients. The reason is that most normal EQ designs don’t just boost and cut frequency regions. They also delay (or ‘phase‑shift’) some of them as a side‑effect, which can effectively ‘smear’ well‑defined transients, making them sound less punchy. In practice, phase‑shift side‑effects won’t cause appreciable transient‑smearing problems most of the time, as long as you keep your EQ moves fairly moderate. But there’s one common EQ move that you need to be wary of in this respect: high‑pass filtering. This is something people often do almost as a reflex in small studios, to avoid problems with subsonic rubbish that doesn’t show up on typical two‑way nearfield monitors. But with drums, high‑pass filtering can significantly reduce the solidity and power of the instrument’s attack on account of the phase‑shift, even if the frequency response effects seem pretty minimal. I deliberately didn’t high‑pass filter the snare for this reason.

The subjective power of drums isn’t just about their front‑end ‘spike’, though, because there are other ways you can contribute to the illusion. Adding some dense room reverb is a time‑honoured approach, and hearing this in the Simple Minds mix encouraged me to try something similar on Trevor’s snare. With any heavier reverb like this, it’s important to realise that the effect will to some extent be perceived as part of the overall snare tone, so do make sure you spend enough time auditioning different reverb patches, and don’t be afraid to use those effects assertively to get a combined timbre that really fits the mix. In my case, I spent a good 15 minutes trying different reverb plug‑ins and presets before settling on a blend of two different custom‑tweaked room patches. Even then, I didn’t leave those reverbs static throughout the song, and automated their return levels to ‘inflate’ the snare sound more during the higher‑energy sections of the song.

Emphasising the sustain of your drums can also suggest to the listener that the kit sounds more powerful, and compression is the natural choice for this. The danger is that traditional compression carries the risk of counter‑productively blunting your drum transients. This is one of the classic scenarios where parallel compression (mixing compressed and uncompressed drum kit sounds together) really comes into its own, because the transients of the uncompressed signal will always reach the mix bus, irrespective of how sadistically you drive the compressor on the parallel path. In this remix, for example, I cranked up some serious gain reduction using an aggressive fast‑attack, fast‑release compression setting based on the ‘all buttons’ mode of a UREI 1176 limiter. This all but flattened out the transients of the compressed signal, but the uncompressed channel’s transients still arrived safely at the mix bus. The result: more sustain, without any loss of attack.

Leading From The Front

My final key requirement for this mix was that the lead vocal should be right up front where it would demand the listener’s attention for the lyrics. As with the kick and snare, that meant both increasing its overall level and controlling its balance more stringently, to avoid it ever overwhelming the backing track and undermining the band’s sense of size. A single layer of dynamics control is rarely enough to walk this mixing tightrope, so I first used a limiter to catch the loudest peaks, then followed it with slower‑acting 3:1 ratio compression, to squeeze the overall dynamic range by 3‑4 dB.

Increasing the density and consistency of a vocal signal’s upper spectrum can help bring it to the front of the mix, and Mike achieved that for the verse vocal in this mix using a combination of analogue‑modelled saturation (to generate additional harmonic content) and fast‑acting compression acting on the frequency range above 5kHz.

Beyond basic dynamics processing, there are a number of things you can do to bring vocals to the front. Anything that increases the number of harmonics in the upper spectrum usually helps, for example. Not only does it brighten the tone, but that sense of brightness makes the part less susceptible to frequency masking. In this case, I set up a dedicated parallel distortion channel for this and, for instance, mixed in the distortion’s added harmonics to supplement the singer’s natural high frequencies during the verses. I also used a multiband compressor for all the song’s lead vocals, to further brighten the frequency region above 5kHz without overemphasising the already bright‑sounding noise consonants and breaths.

If you’re using any vocal reverb effects, there’s always a danger they’ll drag your singer backwards in the depth perspective, and judging by Trevor’s mix and choice of reference material, I knew I’d need to take precautions against this, because vocal reverb was definitely on the menu! So here’s what I did:

1. I fed some of the reverb from a tempo‑sync’ed feedback delay effect. This makes it easier to lengthen the apparent reverb tails without as much of a sense that you’re washing out the whole mix with cavernous reverb.

Here you can see three things Mike did to the vocal reverb in his remix to keep the lead singer sounding up‑front: de‑essing the reverb send to avoid consonant sounds ‘splashing’ in the effect tail; adding 50ms of pre‑delay to the reverb itself; and assertively EQ’ing the reverb return to distance the effect and reduce its midrange content.

2. I made sure that all the reverbs I used had at least 10ms of pre‑delay. This helps separate the dry vocal signal from the reverb, weakening the distancing effect.

3. I de‑essed the reverb sends, to avoid splashy consonants in the reverb return from pushing in front of the dry vocal signal.

4. I used EQ to cut regions from the vocal reverb returns wherever they seemed to muddy the vocal tone, become too audible in the midrange, or sound too bright by comparison with the dry signal. A useful tip is to temporarily turn the reverb up 3‑4 dB too loud while EQ’ing, because that makes it easier to hear which specific frequencies need cutting.

Besides the vocals, the other lead part was a guitar solo, which I also wanted to have at the front of the mix. Initially, though, the recording presented two difficulties. First, the distortion felt like it had been driven a bit too hard, such that I didn’t feel I could fade up the more musically important pitched information without my ears being fatigued by aggressive upper‑midrange hash. And when I used upper‑midrange EQ cuts to tackle the harshness, the guitar ended up feeling distant compared with the cymbals and other guitars in the arrangement. Second, a long modulated echo effect had been baked into the guitar recording, so I couldn’t adjust those effects independently.

Fortunately, Trevor had recorded a DI signal for this part too, which gave me some room to manoeuvre. (It’s never a bad idea to capture a ‘safety’ DI signal when you record electric guitars, even if you never actually use it — it can really save your bacon at mixdown if you misjudge the amp or stompbox settings.) Re‑amping the DI with a less heavily‑driven sound allowed me to reintroduce some less abrasive‑sounding high frequencies into the mix. This already helped pull the guitar sound forward. Then, because this re‑amped signal included no effects, I could adjust the wet/dry mix of the echo effect according to how I mixed the re‑amp with Trevor’s original guitar part. I could have just ditched the original guitar effects entirely and created totally new ones for the re‑amp track, but the echo had a certain character that I felt might prove tricky to recreate from scratch — I risked throwing the baby out with the bathwater!

Music Before Mix

In this month’s remix, I’ve showcased lots of different mixing techniques for enhancing clarity, beefing up your drum sound, and bringing lead lines closer to the listener, but none of those will do you much good if your listener loses interest in the music. So it pays to think twice every time you’re tempted to just copy and paste.