Mix Rescue: Rio Callahan

This month, as well as sorting out a few sonic gremlins, with edits, EQ and other tricks, we aim to draw out more of the 'moments' in a track.

Glasgow rockers Rio Callahan, the subject of this month's Mix Rescue.

As much as we might lament the decline of big‑budget recording sessions, one of the beautiful things about the modern‑day music business is that musicians and bands alike have the ability to experiment with the production process in their own time, at home and without huge expense. Certainly, this approach seems to have worked pretty well for Glasgow‑based band Rio Callahan, the subject of this month's Mix Rescue. This pop/rock six‑piece released their latest work, The Average EP, in mid 2011, having recorded and mixed the whole project in their own lock-up in the East End of Glasgow.

The down side of this DIY approach, though, is that it can be one hell of a job separating your roles as band member and engineer, so that you can give your all in both. Being in a band myself, I understand that this is often a very tricky balancing act, so I've got bags of respect to the bassist Cameron for taking on that challenge. While pleased with the result, though, the band didn't feel that they'd quite got the very best out of all of their songs, and after talking to them I offered to remix the EP's lead track, 'Re-ignite'.

After a listening critically to the track a few times, it appeared to be pretty well balanced, and certainly very listenable, but I could tell that there were a fair few areas in which I could make improvements. It seemed to me that the mix was indeed suffering from the 'band member as engineer' issue, the result being that the approach was more a case of getting a recording of a good live performance, as they'd deliver it at a gig, than of making the most of the recording and production process. There's nothing wrong with that approach per se, but if you plan on releasing something, it will be compared with more polished productions, and although, as I said, the original mix sounded well balanced, that balance felt rather static: I didn't feel as drawn into the song as I could be.

Prep School

My goal, then, would be to make the mix sound bigger, more polished and more engaging. To do that, I'd need to work on the individual tracks, bringing a little more space and clarity to the sound, but most of all I felt that I'd need to try to tease out and highlight some of the 'moments' in the song, which just didn't come through strongly enough.

The multitrack files comprised over thirty separate tracks, including drums, bass, guitars, pianos, Hammond organ, trumpets, and both lead and backing vocals. I dropped the files into Pro Tools and played the song a few times, creating a rough level-and-pan balance as I went along. This approach allows some time to get into the feel of the song, and helps me figure out in general terms how best to go about the mix: I won't necessarily work the same way for every record, because each song is different. I also checked all the files for phase‑cancellation issues, to nip such problems in the bud, and as none of the files were named in any meaningful way, I labelled them as I went along. This sort of up‑front housekeeping makes it so much easier to focus on the job of mixing!

With this prep‑work done, I'll often start a mix by working on the star element (usually, but not always, the lead vocal), but in this case I adopted a different strategy. As I'd noted some issues with the drum sounds, I decided I'd start with them, just to give the track a good, firm spine. In fact, in the event, I left much of the work on the vocals until near the end of the mix — although that's less shocking than it might at first seem, as the original vocals already formed part of the rough balance, and I was thus able to consider how other parts interacted with them.

Just For Kicks

An audible flam on the kick drum, as shown on this waveform, detracted from the power of the song's groove, and thus needed fixing!To resolve the flamming kick issue, Gordon found a good, solid kick beat from later in the track and used Pro Tools' Sound Replacer to substitute it for the offending sounds.As you can see in the screenshot, most kick-drum hits had their own 'after-hit'. This could, perhaps, have been down to a ripped front drum‑skin flapping after each beat, possibly the drummer's technique, or maybe the action on his kick pedal. Whatever it was, it interfered with the rhythmic feel of the piece.

I could have tackled this in a couple of different ways. In some tracks, you can use a 'strip silence' function to isolate each hit, and then shorten the events for each kick hit (or attempt to do the same with a gate). In my experience, though, a better option is often to replace the problem hits with more solid‑sounding samples — the trick being to pick samples that don't change the overall character of the kit.

In this instance, I managed to grab a suitably solid-sounding hit, with no 'after-hit', from later in the song and save that as a sample. That's about as unobtrusive a sample as I could have hoped for, and it allowed me to replace all of the offending hits with the good one, using Pro Tools' Sound Replacer. This brought a much greater sense of solidity to the drums. Replacing kick-drum parts that have been played with busier foot‑patterns can often sound a little unnatural, but fortunately the pattern wasn't too busy here: most were single hits, so I was able to pop in the sample without compromising the feel. If you want to learn more about drum‑replacement and reinforcement techniques using samples, there's an in‑depth guide on the SOS web site at /sos/mar11/articles/cutting‑edge‑drums.htm).

Snare Drum

With the kick now working as it should, I turned my attention to the snare, which sounded super‑dead. I appreciate that there are times when a snare needs to be heavily damped to suit the vibe of a certain track, but in this case it sounded to me as though the snare had been choked just a little too much. Another issue was the high level of bleed in the snare track from the hi‑hats. There's nothing wrong with working with bleed if you know what you're doing — it has played a part in the sound of countless great records — but it wasn't helpful here.

This is an issue that's obviously best tackled when recording: it could be down to poor mic choice or placement, the choice or placement of the hi‑hat, or to the drummer's playing, and correcting any of these issues at source will give the best results. Still, there are tricks you can deploy in post‑production. Luckily, the hi‑hats (being played in relation to the snare) were on the off-beat, and this allowed me to duck the hats on the snare track without affecting the sound of the snare drum itself. A traditional threshold‑based noise gate or side‑chain compressor wouldn't have given me the consistency I needed, due to the varying velocity on the downbeats, so I opted to draw in automation manually. This may seem like a tedious and labour-intensive approach, but it's one that affords you far more control.

Room For Improvement

I love using a room mic when recording acoustic drums, because when you crush that bad boy it can sound amazing! Unfortunately, there's a certain amount of randomness to whether or not it will work well, and on this track I found myself working with a mono room mic that sounded OK on its own but was so distant that it sounded as though it had been placed a mile down the street! Well, maybe not quite a mile, but as you can see in the screenshot, the time difference between the signals from the overhead and room mics was pretty large, and the result was that bringing up the room-mic fader made it sounded like there was a crazy snare flam. There's no point in completely time‑aligning the room-mic signal with that of any other kit part, of course — because that would defeat the object of having a room mic in the first place — so to bring things back in line, I simply nudged the room mic waveform back (to the left) until I was happy with the sound in relation to rest of the kit.A room mic had been used when recording the drums, and although it was a useful sound, it was much too far back! To counter this, the room‑mic track was manually nudged nearer in time to the rest of the kit, with the results judged by ear.

A cool little trick I like to use, and which put in an appearance on this mix, is to grab a single snare hit from the room-mic track, then crop it to size and save it as a sample. Then I can use a close-mic snare track to create hit points that can be used to trigger this new sample. By doing this, I'm able to achieve that big room sound on the snare, while managing to avoid all the other wishy‑washy cymbal work that you'd end up with if you simply turned up the room-mic track. It's simple, but it can work really well, and means that you don't have to spend an age auditioning your sample library to find the perfect snare sound to layer in for the same effect.

Drum Bus

With the key elements of the kit now firmly under control, I sent all of the drum tracks to an Aux channel in Pro Tools to create a drum bus. On this channel, I placed an instance of the Waves SSL G‑series Bus Compressor plug‑in, followed by Waves' L2 limiter. The latter is a plug‑in that can sound over-harsh when it's really crushing, but I mostly use it for subtle dynamic control, and in that role it works perfectly for me. You could achieve a similar effect with many other limiters, though: they all do a similar job to some extent, though they do all have slightly different characters, and it's worth experimenting with a few different ones so you can develop a feel for which tool you want to reach for first.

Firm Bottom

Personally, I think it's hard to go wrong with DI'ing the bass, as the band had done here — and there are some excellent modelling plug-ins if you decide later that you want an amped sound. But as well as this being a nice clean recording, the part had been well played, which meant I didn't need to replace or fix anything, only to apply a touch of compression and EQ. I started by rolling off the top from around 6kHz, much as a speaker would do, and used a narrow dip around 300Hz to notch out a nasty 'honk'. I also ended up dipping around 62Hz by a few dB, but not because of anything wrong with the bass sound. Rather, this was to allow the bottom end of the kick to really own that area of the spectrum. Overall, then, there wasn't any crazy tonal sculpting going here, just some good playing and a few simple EQ processes. Last in the chain was another L2, just to give everything a bit of a squeeze, so that the bass and drums together now provided a good, solid foundation for the rest of the track.

It's important to go through each part and listen for clicks, pops and other offending sounds, which can become a real problem when you start compression and limiting. On the piano part, there' was an audible spike, which was treated by cutting and pasting the same note from elsewhere in the track.Despite all this talk of compression and limiting, and while I do use compression quite a lot, I try not to be too aggressive with it. I generally get much better results aiming for just enough gain reduction to bring a part's dynamics under control. In my experience, you tend to end up with a much more spacious‑sounding and engaging mix that way — and if you really do want to crush everything and raise the level of the mix, you can do so on the master bus or, later, during some sort of mastering process.

Piano To DI For

The piano sound, which was a stereo DI'd electric piano, was a case in point. Having EQ'd this part to sound a little more dark and moody, and less bright and happy — and thus more in keeping with the track — it required only very gentle compression to bring a few wayward velocities under control.

A little corrective work was required on this part as well. At the start of the song, where the piano is more exposed, there was an audible spike in the sound, which could have been a power surge during recording, or perhaps a bad edit when the session was being prepped for a mix (don't forget to crossfade when doing edits like this!). I tackled this problem by grabbing a copy of the same chord from later in the track and using this to replace the one with the spike. It's easy to fix problems like this, but you really do need to listen out for them!

Guitars

It sounded as though there were two different electric guitars used to record this song. One was distinctly Strat-ish to my ears, and unfortunately its tracks all sounded a bit murky and dull (which isn't normally the description you would give when listening to a Strat!), whereas the other, whatever it was, had a bit more 'beef'.

My guess is that the limitations of the Strat recording were due to the mic choice and/or placement. Whatever the cause, I EQ'd the Strat parts more aggressively than I did the others, looking to brighten the sound a bit. I started by rolling off the bottom end by about 100Hz, rolling off the top at about 5.5kHz and then dipping at around 300Hz with quite a narrow Q. To compensate for the dullness, I found myself boosting in the 3-5 kHz region, as this seemed to give the Strat a much more appropriate place in the overall mix. It did feel at times as though I was fighting the Strat rather than working with it, but I did manage to get it to a good place eventually. Once again, I applied a gentle touch of compression on both tracks, and at the end I panned each to one side, about 45 percent of the way towards the left and right extremes.

Trumpets & Organs

I thought the three trumpet parts sounded awesome: they were really clear and tuneful, and they suited the theme and mood of the song to a tee. So little was required from me here. I used some compression, just to control the more strident notes, and I filtered out some of the low end to make space for other sounds in the mix. As the mix evolved, I also boosted the trumpets a little in the upper mids, just to pull them forward. There were three trumpets, and I panned two of them slightly, one to each side, and left the other in the centre, which created a nice, balanced spread. This sounded good, to my ears, so I did little more there other than ride the levels a little.

The Hammond organ appeared only once in the track, and I felt that I needed to make something of a feature of it. To do this, rather than simply bringing the Hammond part in loud from the off, I pulled the guitars down a bit, letting them take a back seat while the listener was directed to the new sound. Peeling back the layers in a mix like this, whether with level automation or subtractive EQ, can really work wonders, as you're always left with plenty more room to develop the track, whereas chunky EQ and level boosts often leave you without any more space in which to build. This tweak made for a nice change, and it was then a simple matter of tucking the part into the mix, yet again using some soft compression, this time combined with a bass roll‑off and a small high‑shelf boost (about 3dB), just to make sure the organ came to the fore.

Vocal Processing & FX

The song begins with four layers of vocals singing the same lyrics in harmony but, while I understood what the band were trying to achieve, the effect was undermined by something of a plosive circus! Layering vocals can create a wonderful sound, but you really need to focus on both the timing and levels of harder sounds, particularly stacked plosives like this. To address this, I manually nudged the three backing vocal tracks to lock more tightly to the lead vocal, and manually dipped the plosives a good four or five decibels on each of the backing vocal tracks. This all helped to smooth the flow of the backing vocals.Getting the timing of backing vocal parts right can be critically important to the sound of your mix. As well as dealing with the sound of multitracked plosives and sibilance, Gordon tightened the timing of the parts manually, using the lead vocal as a reference.

My plug‑in chain for these vocals was the SSL 4000 channel and a de‑esser. I normally manually dip all the 'essing' for each vocal, but I also like to have a de‑esser plug-in there as back up. Each vocal was then sent to a vocal group bus on which I'd placed another L2.

In the original mix, the vocals at the start of the track seemed a wee bit dry, considering that there were only vocals and piano at that point, and I decided to go for a wetter sound, using two reverbs and a delay, each on their own stereo Aux track. For the lead vocal, I used a short, dark vocal plate that stayed the same throughout the whole track. In addition, I used a longer reverb (around 3.5 seconds) and automated its level, boosting it at the more ambient moments of the track, just to make things feel a little richer. For the delay, I used a short slap‑back setting and again automated with a fader ride, for the same reasons. It's hard to say how you approach automation like this, as it's really quite an intuitive thing: you just know when it feels right.

Another key thing to mention regarding vocal processing is that I often put a frequency analyser plug-in on the lead vocal track, to help me identify the part's dominant frequencies. I'm then in a better position to be able to 'scoop' three or four decibels of those frequencies out of any of the other tracks that seem to be obscuring or competing with the vocal. Again, this subtractive approach to EQ on other parts — peeling back one part to reveal another — often seems to work wonders, and this case was no exception. A bonus is that you avoid the temptation to boost bits of everything until the whole mix sounds like mush!

Magic Moments

Gordon made plenty of use of level automation to maintain a sense of movement and momentum in the track.

It's all very well talking about how I treated the individual elements of the mix, but you'll recall that I mentioned at the beginning of the article that I wanted to tease out some 'moments'. This is a term I use to describe all those little things that make a mix seem so much more detailed and polished than a simple but accurate recording of a band's live performance. There are a hundred ways you can approach this, but here I worked on a couple of things in particular.

The first 'moment' is, inevitably, the intro. In the original mix, the soft vocal intro lacked any ambience, and it really needed to shine. I've already discussed how I'd nudged and comped all the vocals, and made them more ambient using two reverbs and a delay, and for me this combined with the dark sound of the piano to instantly put the song in a new place. It also made the next vocal passage (which is more 'sung out') seem bigger and much more present. I then used automation to pull the piano down by a few dB when the vocals started, nudging them back up again when the vocals finished, just so that the piano didn't intrude on the vocals, but still came to the fore when there was space for it.

When it came to the single trumpet/brass solo in the middle, I played with a few ideas, and in the end decided to mute all of the drum mics except the mono room mic. The room mic was heavily compressed and sounded huge, and I chose to pan it hard left for that section. Immediately, this gave the bass and trumpet some time: another 'moment' in the bag. As it kicked back in to the next section, I pushed the levels up on all of the drum mics and switched the panning of the room mic gradually to centre again, which gave the next section a nice lift.

The 'down' vocal section after the trumpet solo was another good opportunity to take the mix somewhere new. This time, I decided I'd go for a filtered lead and backing vocal sound, courtesy of the trusty old telephone EQ effect. It can seem a cliché at times, but I strongly suspected it would suit this song, and to my satisfaction it instantly gave the track a greater sense of intimacy. Needless to say, where the drum fill came in to lead the listener to the song's outro I took the telephone filter off. The result was that the next section really kicked in, full of life and sounding huge, to create a fitting end to the track.

Finalising

In this screen you can see just how much of the processing attention was given to the vocal parts (just look at the number of insert processors relative to those on the other tracks); and you can see an instance of Waves' modelled SSL bus compressor that was used to give the mix a squeeze. Note that although the ratio is a pretty hefty 4:1, the attack setting is gentle and the compressor was only bringing gain reduction of up to 4dB.

I'm no mastering engineer, but with the mix pretty much done, I did try applying some pseudo-mastering treatments on the bus. Another instance of Waves SSL G‑series bus compressor was set to yield about ‑4dB of gain reduction on a 4:1 setting. I'd placed a multi‑band EQ in the next insert slot, but in the event didn't use it. Finally, I had another instance of L2 with the ceiling at ‑0.2dBFS, which would both catch any stray transients and avoid problems with clipping, as well as taking care of dithering to 16‑bit when bouncing down.

Before actually bouncing down any track, I skim through the session just to double-check that there are no red lights on any of the channels or plug‑ins that warrant closer inspection of the sound. I always keep an eye (and ear!) on this throughout the mix, and have confidence in knowing that I have limiters in place to shave off any stray peaks, but it's well worth checking anyway.

Overview

I have a lot of respect for Rio Callahan, and for anyone else who takes it upon themselves to record and mix their own music (I'm sure this is how most of us start out). They'd certainly worked on their material and performances for this EP: I really like this and the other tracks, and singer Ross Ferguson's vocals in particular. Unfortunately, though, many of the source recordings could have been better, and this had made the mixing process for the band a little more challenging than should have been the case. This, in turn, probably meant that the original track had been robbed of some of the impact it has when played live. As you can see, then, I spent a lot of time on this track cleaning, tidying and prepping things for the mix — whereas the actual mixing process was pretty plain sailing.

I was really happy with how my mix turned out, and I hope the band can take away some pointers from this article and use them to their advantage in future recording sessions — whether it be about taking the time to get the sounds right at source, how to fix things when they go wrong, or how to add that final, polished dimension to the recorded sounds. It is a lot of work, for sure, and it certainly takes a lot of time to figure out all the little things that help make mixes 'pop', but it's well worth the effort!  

Gordon McNeil is a producer/mixer working from Gorbals Sound, a newly built music complex in Glasgow. He's currently producing new albums for the band GoGoBot and singer-songwriter Maeve O'Boyle.

www.gorbalssound.co.uk