Cats On The Beach: We go back to basics and turn up the heat on an indie-rock production.
With almost endless processing power at our disposal, it’s easy to forget that not all mix problems can be solved by throwing CPU cycles at them. And when you do find yourself piling up plug-in after plug-in, it’s often worth stepping back and asking yourself whether you’ve identified the issue correctly — and, if so, whether there might be a simpler and more sonically transparent way of dealing with it. After all, few mix problems are really new. How might an engineer have tackled the same problem 40 years ago, without today’s digital magic to call on?
This month’s Mix Rescue could be seen as a case in point. Cats On The Beach are an indie-rock band from London. Having built up a fanbase through gigging, they’d used crowdfunding to pay for the recording of a six-track EP, which had been tracked and mixed at a well-appointed London studio. However, they emerged a little unhappy with the way the song ‘Flashing Swords’ had come out. Guitarist Chris Timmins used the SOS Forum to canvass opinion, saying “For us the whole song is good up until the two-minute mark where, in our minds, it needs to open out into some kind of atmospheric climax. It’s just sounding flat, to be honest.”
Pinpointing the origin of a mix problem sometimes requires a bit of lateral thinking, because the symptoms don’t always have an obvious connection with the source. In this case, there was no doubting that at the point where the track needed to take wings and fly, it remained stubbornly earthbound. The question was: why?
Listening to the track as a whole, I began to wonder whether this specific problem might be resolved by tackling some more general issues with the mix. The instrumental balance seemed askew to my ears, with the vocals and drums too loud, and the close mics too dominant within the drum sound. The guitars, by contrast, were disappearing into the background, and the vocal sound was abrasive to say the least, with fierce and distracting sibilance. The mix as a whole also sounded slightly ‘scooped’ and rather over-processed, almost as though it had been fed through a resonant filter. Finally, several forum posters commented on Rob Marenghi’s vocal performance, which emphasised passion and commitment at the expense of accurate pitching.
Chris sent over the files as a Pro Tools session, and opening it up revealed hundreds of plug-ins, including a great number of vintage-style compressors, console emulations and the like. I had none of these installed on my laptop, so couldn’t inspect the settings, but I was able to open the instance of iZotope’s Ozone that occupied the master bus, and this confirmed some of my suspicions about the over-processed sound. The equaliser was doing much more work than a master EQ should have to, and of particular note was a narrow peaking boost at about 3.5kHz. It’s almost never a good idea to use narrow-band peaking EQ, especially on the master bus, and I’m pretty sure this was responsible for the ‘resonant filter’ effect I was hearing.
Happily, auditioning the raw tracks showed that the studio had done a fine job on the tracking side of things. As well as setting up two snare mics, plus close mics on the three toms and inside the kick drum, the engineer had taken a belt-and-braces approach to capturing the sound of Liam Williamson’s kit as a whole. There was a conventional stereo overhead track, recorded with a pair of AKG C414s; but the drums had also been captured using the classic three-mic ‘Glyn Johns’ drum-miking technique with one mic above the snare, a second above the floor tom and a third out in front of the kick drum. The first two were recorded with ribbon mics, and had the typical mid-rangey ‘thwapp’ that this often generates.
Chris and Rob’s guitar cabinets had been miked front and back, to good effect, while Nick Harris’s bass had been simultaneously miked and DI’d, and both paths sounded pretty nice. The only place where the engineer’s mic choices had let him or her down was, unfortunately, on the lead vocal. Even with no processing applied, this was spitty and harsh. Sometimes, no matter how many expensive capacitor mics you have in your studio, a cheap dynamic or a ribbon is a better option.
Because you have to start somewhere, I started with the drums. A brief inspection of the session made clear that the original engineer had gone to town with plug-ins here, applying not only EQ, but compression, transient shaping, saturation and bass enhancement into the bargain. A duplicate snare track had been used to trigger a sample, and whereas I’d have expected the conventional overheads and the Glyn Johns mics to be an either/or choice, both arrays seemed to be in use, albeit at a very low level. Given that the source tracks seemed to be in decent shape, I was hopeful that it might be possible to get a good drum sound together by focusing more on the basics.
When confronted with a multi-miked kit, the first thing to do is usually to check for any phase issues between tracks. In this case, there were no major problems, but I decided that one of the snare mics fitted in better with its polarity flipped, while the kick drum mics benefited from a delay of 96 samples or so.
The next question was whether to follow the original mix engineer’s example and use both the overheads and the Glyn Johns mics, or whether to commit to only one array. There was a certain tonal logic to combining them, since the latter’s chunky mid-range tone complemented the C414s’ airy brightness, but on the down side, they presented different and incompatible stereo images. Once I raised the other faders, it was clear that the close mics would contribute more than enough mid-range in any case, so I muted the ribbons.
I did, however, retain the ‘front of kick’ Glyn Johns mic; at the right level, it added some welcome thickness and punch to the kick-drum sound. Unfortunately, at the same level, it also introduced objectionable amounts of gritty snare and cymbal spill. My solution for this was to use a compressor with the side-chain filters set to focus on the upper mid-range, so that kick beats passed relatively unscathed while spill was attenuated.
Some drummers have a tendency to attack the cymbals and hi-hat a bit too vigorously, which can make it hard to get the drums themselves loud enough in the mix. I suspect that the original mix engineer felt this was the case here, which would explain the decision to employ a snare sample and to minimise the contribution of the ‘whole kit’ mics. If so, however, I think the mix engineer was partly dealing with a problem of his or her own making. In the Pro Tools session, compressor plug-ins had been placed across many of the individual drum tracks, including both sets of overheads, and these would likely have made any imbalance in the kit sound much worse.
Employing a lighter touch with the dynamic processing let me bring the overheads up a lot before the brass-bothering got out of control, and this in turn helped to generate a more natural and open drum sound. In my final mix, in fact, I didn’t use any conventional compression on the drums at all. To even out a few rogue kick-drum beats without bringing up spill, I turned to Sound Radix’s clever and transparent Drum Leveler, while FabFilter’s Pro-MB multi-band processor was applied to the snare bus and the drum master bus to shape the drum tone rather than for dynamic control. The only full-band compressor in action was FabFilter’s Pro-C, again strapped across the drum master bus, but used fairly subtly as a parallel processor.
The aforementioned tonal shaping, meanwhile, was mostly a matter of controlling a couple of common problem areas. I wanted to retain enough ‘thump’ in the 200Hz region to give the drums plenty of weight, and sufficient ‘snap’ between 1 and 2 kHz to ensure they punched through; but at the same time, I didn’t want the drum mix to become muddy or brittle-sounding. Getting this right involved a fair number of EQ plug-ins scattered across the drum tracks, but a relatively small amount of equalisation in total, as each band was applying only a couple of decibels of boost or cut.
I made up for this relatively abstemious approach to drum processing by using no fewer than three different reverbs on the kit. It’s a habit of mine to set up a short ambience patch on almost every mix, as it can add substance and interest without really making things sound wet. This was fed mainly from the overheads, but also the close mics to an extent. I also find that snares and toms usually benefit from their own dedicated reverb. This is partly because the reverb patches that suit drum close mics often don’t work well on other sources, and partly because I like to return the snare reverb to the drum bus so that it gets processed along with the mics, almost as though it was a room-mic track. And in this particular case, I felt that the impulse response I’d chosen in HOFA’s IQ-Reverb was only getting me 80 percent of the way to the ideal snare sound, so I set up a second snare reverb using Exponential Audio’s R2. Indulgent, but it seemed to work.
Turning next to the bass, I’d have been happy to use just the miked cab from a tonal point of view, but it was quite noisy, and in the first and last section of the song there was drum spill onto the mic (the middle section was obviously a later drop-in and was clean). Together, the noise and spill made it hard to get the bass loud enough in the mix if I used only the mic, so I decided on a mic/DI blend. I used the Pro Tools TimeAdjuster plug-in to put the DI in perfect sync with the miked signal, and used the Bomb Factory SansAmp plug-in to add some grit to the otherwise very clean DI’d track. I then bused these to a single Aux channel where they received the attentions of a compressor and equaliser.
There were four guitar parts, each miked from front and rear. Experimenting with the relative polarity of these tracks yielded some useful options: with both tracks set to the same polarity, the sound was quite thick and meaty, while flipping the rear mic yielded a thinner, brighter sound. I ended up taking the former approach for the rhythm tracks, though they then required some low-mid cut to keep the thickness under control, while I preferred the leaner, pointier sound for the lead overdubs that came in later in the song. In fact, I ended up using SoundToys’ Filter Freak to make one of them even more lean and pointy.
Two other supporting instruments were used in the song: a Rhodes piano and a brief MIDI organ part. The DI’d Rhodes part sounded rather plain without further ornament, so I ran it through Scuffham Amps’ S-Gear amp simulator, with some tremolo and delay.
As I worked through the song, I began to feel that the lack of impact in the last two song sections wasn’t solely a mix problem, but had its roots in a couple of decisions relating to the arrangement. At the key two-minute point, a prominent delay effect had been introduced into the rhythm guitar sound. Considered in isolation, it was a nice effect, but, as delay effects will, it tended to push the guitar away from the listener. The drumming, meanwhile, maintained a pretty steady kick-drum beat throughout the song until the very last section, which was played on snare alone. The kick drum dropping out deprived this last section of some of its drive and power, not to mention a significant chunk of bottom end.
After a brief, failed flirtation with cutting and pasting kick-drum beats from elsewhere in the song, I realised that neither of these issues could be addressed directly. Instead, I used the bass guitar to pick up the slack. Although the song hadn’t been tracked to a click, the band had kept excellent time — and, as is typical of bass lines, all of the phrases were repeated. By swapping over similar phrases, it was relatively easy to cobble together a ‘double-track’ for the DI’d bass part. I duplicated this, panned the two copies hard left and right, and ran them through a snarling, mid-rangey SansAmp setting. This meant that as the guitar receded into the background and the kick drum dropped out, the bass could become more aggressive and up-front to compensate. In the end, I even played up the spaciness of the middle section’s guitars by running them through a fairly psychedelic patch in SoundToys’ Crystallizer.
I also marked the changes between sections by radically rebalancing the three reverbs I was using for drums. In the sparse opening section, I kept all three fairly high in the mix. I then dropped out the ambience altogether for the second section, before reversing the arrangement in the last section, so that only the short ambience was used and not either of the snare reverbs.
As I’ve mentioned, the original mix sounded over-processed, and I suspect that the over-use of master bus processing contributed quite a lot to this. The Pro Tools session contained half a dozen plug-ins across the mix bus, including at least three compressors and a tape emulation, though I don’t know if these were all active. The mix engineer had also used Slate Digital’s Virtual Mix Rack on many tracks, along with a number of vintage emulation plug-ins from the likes of Universal Audio. And although I felt that the mix as a whole was better off without most of this, I could see there was a role for some analogue-style grit, warmth, saturation or colour.
I often don’t use mix-bus compression, but in this case I felt that it added something: not so much the mythical ‘glue’ that squashing the master bus is supposed to provide, but an overall harder and more ‘in your face’ feel. Configuring FabFilter’s Pro-C in Middle & Sides mode also helped to even out a few imbalances between centre-panned elements such as the drums, and the guitars, which were mostly hard-panned.
I also experimented with various saturation effects on the mix bus, and although I liked what they did to some elements, they also tended to undermine the work I’d done to tame the harsh vocal sound and aggressive cymbals. In the end, I set up an Aux track with FabFilter’s Saturn distortion plug-in set to the ‘Magic Mastering’ preset, then sent varying amounts of each instrument to this. Finally, I used another multi-band compressor across the mix bus to add brightness in a controlled fashion.
At this point, things were coming together nicely — up to a point. The more natural mix balance was a definite improvement, but even my additional fuzz bass parts hadn’t completely solved the problem that was originally bothering the band. While the song was no longer collapsing in on itself at the two-minute mark, it still wasn’t kicking up as many gears as it needed to.
As I mentioned at the start of the article, working in modern DAWs presents so many high-tech ways of tackling problems that it’s sometimes easy to overlook the traditional, simple solutions. And in this case, the solution couldn’t have been simpler: at the point where the song needed to get bigger, I just turned everything up.
This might sound like an absurdly crude thing to do in a modern mix context, but it’s a trick that has a long pedigree. I’ve read, for instance, that Queen producer Roy Thomas Baker would crank the master fader by as much as 8dB to help his choruses stand out. In this case, it worked a treat. In fact it worked so well that I used it again 40 seconds later, when the song stepped up another level in intensity. (To be specific, what I actually did was to automate the level of the ‘submaster’ group through which all the tracks were routed, so that this level change happened prior to my master-bus compression and limiting.)
Together, these two step-wise fader movements added a total of about 2.5dB gain by the end of the song; that may not seem a vast amount, but it was enough to ensure that the new song sections arrived with a bang.
As a mix engineer, I think you know you’re on the right track when the artist is able to give you very concrete requests for revisions, rather than saying “this just doesn’t quite work for me”, or “it needs to be more orange”. In this case, the band felt that my initial mix wasn’t as bright as they’d have liked; they also wanted more aggressive pitch-correction on the lead vocal, and they picked up on some detailed points concerning the vocal sound and level. All of these were straightforward to address, and after batting a few files back and forth, we quickly arrived at a mix that everyone was happy with.
There was only one element in the entire ‘Flashing Swords’ multitrack where I felt that an inappropriate mic choice had caused problems. Unfortunately, it was the most important element! In the original mix, Rob Marenghi’s lead vocal was way out front, which highlighted the harsh tone and exaggerated sibilance. However, simply removing the existing processing and lowering the vocal level didn’t really solve the problem, which was pretty much baked into the original recording. After several SOS Forum posters had commented on the vocal pitching, the band also agreed that some tuning was desirable, so my first step was to load the vocal track into Melodyne and set to work.The style of singing made it hard to get particularly transparent results, but I did my best. One thing I find with pitch-correction is that it’s often better to process just the sustained vowel sounds, as the consonants don’t have much of a pitched quality in any case, but often give rise to strange artifacts when corrected. My normal approach, as here, is to correct the entire audio part, then load it into a new Playlist on the original vocal track. Then I’ll create a ‘comp’ from the corrected and raw tracks, hoping to strike the best possible balance between accurate pitching and natural sound. I ended up going through this process twice on ‘Flashing Swords’, as the band felt that my first attempt at ‘Melodyning’ the vocal was too conservative.
There remained the question of how to make the vocal sound nicer. Simple de-essing helped to a certain extent, but didn’t really address the harsh tone, and it took a lot of experimentation to improve things in this department. The solution I eventually came up with involved a multi-band compressor set to squish frequencies above 3kHz or so, and a parallel vocal track where I used SoundToys’ Radiator plug-in to add warmth.
One thing I can never predict when working with bands is preferences regarding vocal effects. Some people are outraged by the merest suggestion of reverb, while others are not happy until the vocal is positively drowning in the stuff. In this case, I felt that keeping the vocal dry would have been a mistake: as the only lead element for much of the song, it had a lot to carry, and I hoped that some effects would help to cover up the iffy tone and pitch-correction artifacts. In this case, the band agreed, and in fact asked for the effects to be more prominent than in my original mix. In the end I used a longish plate reverb from Exponential Audio’s R2 and a slapback echo from SoundToys’ Echo Boy, plus a tempo-sync’ed delay that was automated to bring out certain words.
Backing vocals didn’t really make an appearance until the last section of the song, where I felt that the band’s ‘oh-ohs’ could have a more gang-style feeling to them. To create this, I duplicated the tracks and swapped the order of the ‘oh-ohs’ on the duplicates, in effect creating double-tracks from the existing material. The whole lot was then bused through SoundToys’ Radiator, to add warmth and brightness, and the same company’s MicroShift, for some additional stereo width.
Indie-rock four-piece Cats On The Beach describe themselves as “the right kind of mixed bag”. Their work has already attracted the attention of BBC Introducing, PR giants Playground, producer Paul Tipler (Idlewild, 80s Matchbox) and Sentric Publishing, as well as giving them a place under the wing of music charity Strummerville.
“It all seemed quite mono and flat,” say the band of the original mix. “The two final sections are supposed to lift the song to a climax, but in the original mix it seemed to run out of steam by the time it got there, and there was nowhere left to go. The snare was also a bit overpowering throughout and there were some mouth noises that were a bit too prominent.
“The new mix has addressed all of these issues. The guitars now have the width and space they need, the drums gel much more with the track as a whole and the climaxes work well. The new mix is exactly how we imagined the song to be. Thanks SOS!”
The original mix engineer supplied both a full mix and vocal, drum, bass, keys and guitar stems, so I’ve done the same. Surf to this month’s media page at www.sosm.ag/jun2015media to compare them.