Our engineer demonstrates that there’s no point showing off your production chops if they don’t serve the song!
Just because you can play something, it doesn’t necessarily mean you should — I often find myself offering this advice to musicians in the studio. It takes a long time to become good on an instrument, and it’s very natural to want to showcase these skills, but good, experienced players know when to keep things simple. They’ll instinctively deliver what’s right for a song, and when they choose to break out into something a bit more technical or ‘clever’, the contrast just adds to the impact. Of course, this lesson — that we must approach songs with a certain selflessness — applies every bit as much to mixing songs as it does to playing them.
In our modern world of knob-twiddling and mouse-clicking we can easily become so immersed in the process that we lose sight of what we’re trying to achieve. We (rightly) invest time in learning, but we can be bombarded from many different directions with new tips and techniques, and advice on how to get things sounding a certain way. Add in the bewildering quantity of plug-ins and other tools now available to us, and it’s not surprising that it can feel harder than it should to present a song in the best way possible!
At the heart of a good mix is the balance in levels between the various elements involved. This sounds so obvious, but it’s amazing how easily we can misjudge it, and how often you can find yourself contemplating some other process when a simple nudge of the fader is what’s required. If you’ve had to do a lot of work to rescue or reinvent a drum sound, for example, and you’re pleased with what you’ve achieved, it’s easy to leave it a little bit louder than perhaps it should be — you’re rightly proud of what you’ve done. Such decisions are not easy to get right, and it’s with this in mind that I want to discuss my mix of the track ‘Lovely Day’ by Gavin Chapell-Bates. It wasn’t until quite late in the mix process that I realised I’d done precisely what I’ve described above, and what I’ve told so many artists not to do! After puzzling for some time about why things weren’t quite working, it was when I simply dropped the level of the carefully-crafted rhythm section that the song suddenly presented itself in the right way.
Gavin Chappell-Bates, pictured above, is an acoustic folk-rock singer-songwriter from Cambridge, England. Gavin’s music has featured on Tom Robinson’s show on BBC 6 Music and Janice Long’s show on BBC Radio 2. He regularly appears on BBC Introducing, as well as other local and national stations. Gavin is currently recording his second album, the first single from which is called ‘Lovely Day’ and was released on Friday 18th August 2017.
In many ways this was quite a straightforward mix: I had drums, bass, two acoustic guitars, lead and group vocals to contend with. Gavin performs live with just his singing and his acoustic guitar, and these needed to form the heart of the mix. The drums were playing a skiffle-style beat, and had been played with regular drum sticks rather than brushes or rods, so I could see that these would need careful management if I was to prevent them from becoming ‘too much’ for the song. The bass guitar bounced along with the drums, and the second acoustic guitar, which came in and out of the track, had a chimey, almost mandolin-like sound. A large ‘stacked’ backing vocal sound provided a big part of the vibe and character of the song, and the all-important lead vocal was supported by two double takes, intended to be placed quite low in the mix (if you could hear them, they were probably too loud). There wouldn’t be a huge amount of ‘space’ to play with mix-wise, then, but the song is done and dusted in under three minutes and is intended to reflect the title of the song, by being energetic and positive throughout.
My initial game plan was sound. I created a rough static balance of the song before getting into any serious plug-in action. I then spent a little time working on the main acoustic guitar and vocal, my thinking being that the rest of the instruments should fit around these key players. My main decision with the acoustic guitar was which of the two mic options to choose — it had been captured via both a small-diaphragm condenser and a ribbon mic. Each sound had its qualities, and I preferred the slightly brighter condenser, though I kept the ribbon track on standby in case this brightness became too much in the context of the mix. Recently, I’ve enjoyed applying a bit of parallel compression on acoustic guitar — I hate being able to ‘hear’ the compression working on many acoustic instruments, and this tactic tends to bring up the little details in a more natural-sounding way.
I do like to firm things up in the context of a full band song, however, so as well as the parallel compression, I inserted Softube’s new Tape plug-in, which does an excellent job of removing a bit of the ‘sterility’ of ultra-clean digital recordings. This plug-in also has a control to add or remove some top end, and I played around until I found a combination that seemed to work alongside the vocal. I wanted this guitar to be quite bright and forward — but without sounding harsh.
As well as this initial processing, I used a few elements of my Pro Tools mix template, which mainly comprises a chain of plug-ins on the master bus and a few aux channels with effects already set up, so I don’t have to interrupt my creative flow to set them up when mixing. On the master bus, I had another instance of the Tape plug-in and an SSL 4k bus-compressor emulation, but by far the most significant plug-in here was an EQ.
I’ve been trying out a few different plug-ins in this role lately, and in this song I was using Kush Audio’s A-Designs Hammer emulation to apply a generous top-end shelving lift from 15kHz upwards, as well as a little nudge at 60Hz and a decibel or so cut at 250Hz. The idea, particularly with the high-shelf boost, stems from the hardware world: you use one high-quality EQ to add a high-frequency lift on the master bus, so that you don’t need your lesser EQs to do so much work on individual instruments. In the software world you have multiple instances available, of course, but the best analogue-modelling EQ plug-ins can be relatively resource intensive and if you know you’re going to need to add a lot of high-end, you can find it easier to do so from the outset with a single plug-in like this. You do have to be a little careful, though, and listen to the effect — there aren’t that many sources in a typical band mix that have a lot useful information going on towards the very upper reaches of our hearing, and if you start adding generous ‘air’ boosts to vocals and drum overheads, things can become overly bright quite quickly. However, it’s also worth noting that the effect of a high-shelving EQ boost extends much lower down the spectrum than the chosen turnover frequency.
I found myself straying outside of the box for some of the vocal mixing, which is unusual these days — I now tend to use outboard gear more while recording than mixing. Having a new toy was part of my motivation; I’d recently got myself an old Roland RE-201 Space Echo that I was keen to try and use. I used the excellent Radial EXTC re-amp box to get things quickly out of my DAW and back, and this device has the bonus of having a wet/dry control, which was perfect for this application. As I’d already come out of the box, I also took the opportunity to use my hardware 1176 compressor to give the vocal quite a generous squeeze with a medium attack and fast release — this made the vocal sound nicely upfront — before the vocal went into the Space Echo, when the fun of searching for a cool vocal effect began.
Remember what I was talking about at the beginning of this article? Well, what I like to call ‘New Toy Syndrome’ is another manifestation of the same thing. Just because it’s there, it doesn’t mean we have to use it! I was having a great deal of fun playing with the RE-201, though, and in the process found lots of settings that were completely inappropriate for the song! I actually really liked the effect of just running the vocal through the unit without any echo at all, as it seemed to balance out some side effects of the heavy compression. In the end, I recorded two versions back into the DAW: one with no echo, but running through the tape on the RE-201, and one with just a touch of a slap echo. In both cases, this was a parallel effect, with the Radial unit set to about 60 percent wet.
Back in the box, I felt like I was winning, as the lead vocal and acoustic guitar now had a nice, up-front modern sound that I could begin to build the track around. I tucked the two lead-vocal doubles around 20 degrees to each side of the lead vocal, and quite low in the mix. Their job was to just add a little support and weight, rather than be an obvious feature. The group backing vocals required closer attention: I had to arrange eight different layers of harmonies to create a composite part, which needed to feel big without taking up all the space every time they came in. To achieve this, I went quickly through the layers doing some quite drastic EQ cuts to remove what I deemed not essential for a particular layer. This seemed to make them all knit together, with each part still being audible. It felt like a bigger sound than it actually was, which is what I wanted.
For instance, for the higher harmony, I high-pass filtered up to around 300Hz and removed some top end from the lower parts. I also treated the parts as a whole with some fairly heavy compression and an EQ removing some low mid-range, and the parts were carefully panned across the stereo field. Effects-wise at this point things were very simple. I had two quite plain-sounding reverbs, one at around 1s and a longer one at 2.3s. I used different combinations of these on the various vocal parts and throughout the mix.
I mentioned earlier that my main concern with the drums was that the drummer had played with full sticks, when there was an argument that brushes or rods might have worked rather better. The drum sound was good though, and I especially liked the stereo room mic that was available. The kick drum needed a bit of attention first; this needed to be the driving force for the whole rhythm section but it was a little inconsistent dynamically. Sound Radix’s Drum Leveler plug-in is perfect for this kind of problem, and it allowed me to pull up and level out some of the quieter hits more quickly and more effectively than with regular compression.
Once the close-mic kick channel was feeling more like a solid pulse, I headed to that stereo room mic to see if I could make it a significant part of the drum sound — I find that drum room-mic options can often sound really impressive in isolation but fail to work in the mix. I spent a few minutes listening to just the kick and close snare mics and the room mics, playing around with a few different processing options for the room channel and auditioning the different possible options in terms of polarity. Happily, I managed to get the majority of the drum sound with just these three channels, and I just needed to blend in a touch of the overheads to fill out the top end.
When writing this article and going back to check what I’d actually done to the room mics, I was mildly shocked at just how much EQ I’d applied! A 15dB boost at 60Hz, a 10dB cut at around 350Hz and a shelf removing most of the very top end. It sounded good, though, so apparently I hadn’t given it a second thought at the time.
I found myself doing a few things I don’t typically do on a drum mix, and this became something of a theme. I always have all my drums routed to a subgroup, where I’ll apply a few bits of processing, but I often don’t like to compress the kit as whole, as it can sound too ‘pumpy’. But I ended up doing that here, and to good effect. With Waves’ Kramer PIE compressor, I needed only a decibel or two of gain reduction to make the kit feel like it was ‘bouncing’ a bit more. There wasn’t a huge amount of other processing going on with the drums, but I spent quite a lot of time playing with the balance and trying to judge how much weight they should have for the song.
By compressing the room mics a bit, and bringing up the kick and snare, they could easily be made really ‘big’, but that would have been wrong for this setting. Thinking back to what I said earlier about what was right for the song, I actually found myself deliberately making the drums sound smaller. This was mainly achieved by playing with the faders, but I also carefully removed some lower mids on various individual channels and on the drum group as a whole.
The bass guitar ended up being the last instrument I looked at in detail, and what and how much of a role it needed to play required careful contemplation. The drums had quite a generous low end now, which felt right due to the nature of the driving kick drum part. So I ended up thinning the bass out a bit by cutting a little at around 60Hz. This can be quite a difficult thing to do if you listen to the bass in solo — it can sound strange in isolation — but in the context of the whole mix, and with this being quite a busy bass part, it felt right that the drums held the very bottom of the mix and that the bass made way for them.
Something else was bothering me about the bass guitar though, and I couldn’t quite put my finger on what. On a hunch, I loaded it into Celemony Melodyne to see what was happening with the tuning... and sure enough, after about the first 30 seconds of the song, the bass went further and further out of tune until by the end of the song it was almost a semitone flat! I fixed this with Melodyne and then checked very carefully to see if it was adding anything unpleasant in terms of artifacts.
I felt like I’d given every part of the mix a bit of care and attention and, as a whole, it felt broadly in shape. But it didn’t feel finished, and there were still a few issues that were troubling me. By far the most important of these was the lead vocal, and it seemed to be quite tricky to get it sitting in just the right spot.
I tried quite a few different permutations with the vocal level, and listened to the mix on various speakers. It either felt too loud or too quiet, and I wanted to know why! I discovered that the reason lay in the mid-range, as it quite often does; there was a resonance in Gavin’s voice that was making it feel too loud when the part was placed where it needed to be, level-wise. I realised I’d subconsciously tried to address this earlier in the mix by making a few very sharp EQ cuts at 1-2 kHz and with some broad de-essing.
I’d just downloaded a new plug-in to try at this point, and it proved to be very useful indeed for this problem. Soothe by Oeksound is a dynamic equaliser plug-in that can automatically notch out unpleasant resonant frequencies. Guiding the plug-in to focus on the area around 1-4 kHz, it seemed to do a fantastic job of smoothing out the vocal sound and giving it a more balanced tone. Great stuff! I suppose it seem a little like cheating, but you must still assess the result, and I find it refreshing that there are genuinely new tools like this available alongside the seemingly endless supply of classic analogue gear emulations.
Gavin had provided a few reference tracks during the mixing stage, and he was keen for the track to feel bright and lively to suit the aesthetic and message of the song. I do think it really important to check your work against other productions, but it’s crucial that you make sure the tracks you use have something vaguely in common with what you’re working on! One of the references we were using sounded great, but the arrangement was vastly different to ours. I could still tell, though, that it had an overall ‘brightness’ that was lacking from my mix.
Gavin was attending the mix session at this point, and we had a nice dynamic going where we toyed with the pros and cons of making the whole mix a bit brighter. It felt like the track could handle a bit more of a top end boost on the mix bus, except that the acoustic guitar became a bit harsh. I went back to the acoustic track itself and removed the extra brightness I’d added, and we found what we thought was a nice sweet spot for the overall mix.
I was pleased with how I’d made the individual parts of the track sound and also how they knitted together generally. It sounded... OK. But not great. It all felt a bit too chaotic, and whenever I heard the song with something like fresh ears it took me a little while to get used to it again — which was not a great sign! As an experiment, I started listening to the song afresh with just the guitar and vocals playing, and then adding parts back in. It struck me very soon that the source of my troubles lay in the rhythm section and, without giving it a second thought, I turned the drums, bass, and percussion all down by 3dB. It worked. All of that processing and close attention... and this simple nudge down in level made the song more about the vocal and guitars, which is exactly what was needed!
Putting myself in the artist’s shoes helped me think more clearly towards the end of this mix. I’d played with new toys, experimented with different techniques, and patted myself on the back for how I managed to make the kick drum work with the room mics. Does the average listener care about any of this stuff? Nope. Of course they don’t. Great sonics do help a song do what it’s supposed to do, but we have to find a way of stepping outside our world of technical tweakery to focus on what the artist is trying to present. Gavin and I had got together to finish this mix after a little break, and I’d naturally gone past that point of wanting to go in and look at things in detail. We sat and talked about music, different artists and just listened. We then were able to go in and make some broad changes which I think helped improve the finished product enormously.