Despite paying a studio to record and mix their album, Tommy Allen’s band weren’t keen on the result. Could our engineer do any better?
When reader Tommy Allen first contacted me, his frustration was as understandable as it was unmistakable. Despite financing five days tracking and four days mixing in a local professional studio, he felt that the mixes of his blues-rock band’s fourth album left a lot to be desired. Listening to one of the songs, ‘My Father Never Loved Me’, and comparing it with favourite releases from Tommy’s record collection (names like Keb Mo, Anders Osborne, Cry Of Love, Jeff Healey, Scott Henderson and Stevie Ray Vaughan), I couldn’t help but agree. Despite excellent performances from all the musicians involved, the mix as a whole seemed rather boxy and narrow, with a sluggish kick, a rather characterless snare, overbearing cymbals, and a guitar texture that was quite stodgy and distant-sounding.
In response to this disappointment, Tommy had decided to invest serious effort into learning to mix for himself, so that he could personally remix the raw multitracks more to his own taste — but this was easier said than done. For a start, it turned out he’d overcooked the reverb on his amp during the tracking session, and had to rerecord the main rhythm parts to remedy this. However, even then, the ramifications of some of the original tracking engineer’s miking decisions proved tricky to deal with, and he was also unsure what kind of editing and phase-alignment measures might be necessary to handle spill between the different mics. So I suggested that he send me over the multitracks, to see whether I could get a more appropriate mixed result within my Cockos Reaper DAW system, and thereby provide him with something of a template for the remainder of his remixing efforts.
This month’s featured song is from the blues-rock band Trafficker, comprising Tommy Allen (guitar, vocals), Emil Engstrom (bass), and Damon Clarridge (drums). Founded in 2001, the band have toured extensively, and have already released several well-reviewed CDs of their own original material.
Although the raw tracks numbered more than 30, the arrangement was actually pretty straightforward: 10 channels of drum mics; a miked bass amp and its DI signal; a main double-tracked rhythm guitar and an associated harmony layer, all dual-miked; a couple of further guitar tracks (including a solo), which had been both multi-amped and multi-miked on the original tracking session; a stereo Hammond-organ keyboard patch; a couple of tracks of shakers; and a single lead vocal track. For rock music like this, I’ll usually start work on the drums, almost always focusing on the overheads at the outset, as they tend to have the biggest effect upon the overall drum sound.
Immediately, I encountered one of the problems Tommy had been struggling with: whether on account of mic positioning or gung-ho analogue processing during tracking, the cymbals and hi-hat were sounding pretty abrasive, especially in the 4-5 kHz zone, such that it seemed impossible to balance the cymbals at any sensible level in the mix without drastically masking the guitars and lead vocal, or making the whole mix tonality fatiguing on the ear. Swingeing high mid-range EQ cuts didn’t seem particularly effective in banishing the sense of harshness, and undesirably clouded the timbre of the snare ambience into the bargain. The room mic was of little help either as it was thrashy and overly forward-sounding, as well as being mono, where Tommy’s references all clearly implied that he was looking for a decent stereo spread.
I’ll admit that the prospects seemed a bit bleak at this point, but a slightly unorthodox workaround eventually presented itself. What the engineer hadn’t done, thank heavens, was gate out any spill from the four tom-mic tracks and, as luck would have it, the mid-range cymbal tone was significantly smoother through those. So I hatched a plan to use the tom mics, not the overheads, for the bulk of the cymbal sound.
First, I balanced the tom mics to create a passable ‘fake overheads’ signal, using narrow EQ notches to reduce some overenthusiastic pitched rack-tom resonances, some high-pass filtering to combat proximity-effect bass boost on the two floor toms, and a 6dB cut at 11kHz for each rack tom, to reduce a touch of edginess. I then set up a heavily compressed and EQ’d parallel channel, fed from all four tom mics, to emphasise the cymbal spill elements further. Although the resulting stereo picture was, as you might imagine, rather vague, the cymbals tone became a lot more mixable as a result, so I figured it was a price worth paying. I did still use the ‘real’ overheads, but in more of a supporting role, tailoring them with high-pass filtering and mid-range EQ cuts to reintroduce some of their high-frequency ‘air’ to the sound — after all, the tom mics were all dynamic models, so that wasn’t really their forté.
Although the cymbals were now sounding more palatable, and the toms were nicely upfront, the kick and snare close mics were both absent from the mix, so those instruments still sounded more distant and diffuse than you’d expect for this style. With the kick, this was easily remedied using the two kick tracks. A mic inside the drum had provided a nice tight sound that I EQ’d only a little, adding a few decibels of sub-80Hz low end and some beater emphasis with a broadband 1.5dB peak at 1kHz. The second ‘mic’ was actually a loudspeaker driver that had been wired up to an XLR plug. This is a trick I see turning up on increasingly large numbers of project-studio drum recordings, but I have to say I’m not crazy about it myself, because it frequently seems to produce unusable levels of resonant sub-bass and little else besides. Fortunately, though, this multitrack bucked that particular trend, and the speaker-cone signal actually complemented the inside-kick mic signal rather well, once I’d inverted its polarity and refined the phase-match with a sample-delay plug-in to arrive at the punchiest combination.
The snare drum wasn’t quite as straightforward. Although I only had to dip a couple of less attractive pitched resonances with narrow peaking filters to achieve a fairly respectable over-snare close-mic sound, that’s only one of the ingredients you normally need to put together a rock drum mix. The basic problem is that an over-snare close mic picks up just a few aspects of the drum’s character — most commonly a strong attack transient and a dose of lower mid-range weight (often enhanced by the mic’s proximity effect). In light of this, many engineers will multi-mic the snare to flesh out its sound, perhaps with a mic underneath the drum to add more snare-wire rattle or a mic at the side to capture extra shell tone. Here, however, there was only the single close mic, which left the instrument sounding fairly one-dimensional.
Another tactic people often use is to place the overheads and room mics so that they capture plenty of snare ambience, taking advantage of the more naturally balanced snare timbre those more distant mic positions are able to pick up. Unfortunately, I’d jettisoned the room mics completely in the process of combating this multitrack’s aggressive cymbal capture, and had given the remaining snare ambience in the overheads and tom mics a timbre that was, while far from unpleasant, nonetheless fairly understated. As such, the moment I faded up the distorted electric guitars, they immediately masked all the snare-drum’s character, so that it sounded unflatteringly like someone attacking the Michelin Man with a paint-ball gun! (Funnily enough, this same issue doesn’t seem to concern kick-drum sounds nearly as much, because people have grown more used to hearing the equally incongruous-sounding convention of an exclusively close-miked kick drum within an ambient rock-kit mix.)
In short, then, I didn’t feel the multitracks actually contained enough raw material to produce a suitably rich snare sound — so I turned to sample-triggering to bail me out. Slate’s Trigger plug-in did the honours on this occasion, providing plenty of upper-spectrum complexity and sustain from its Bonham-inspired LedSnare instrument. It’s very rare to find a sample that fits your mix perfectly, no matter how big your library of instruments, so I always make an effort to sculpt any triggered sample specifically to fit my mix. In this case, for instance, I deliberately removed the triggered sample’s transient by lengthening the amplitude envelope’s attack setting to 26ms, seeing as I already had plenty of punch coming from the close mic. I also notched out a couple of dissonant lower mid-range pitched components, as well as pulling the 3-6 kHz region down by 3dB.
While this definitely thickened up the snare sound, the snare still sounded a little dislocated from the kit, given the comparatively low level of snare ambience. To address this, I loaded Samplicity’s freeware Bricasti sample collection into Christian Knufinke’s trusty SIR2 convolution plug-in and made a beeline for the ‘Clear Ambience’ impulse response (one of my all-time favourites). As usual, I shortened the impulse considerably to focus on its acoustic character without introducing any obvious ‘reverb tail’, and also narrowed the return’s stereo spread — somehow this helps it feel like it’s contained ‘within the kit’, more like natural recorded ambience, rather than enveloping the kit and giving itself away as an artificial addition.
A dusting of overall room ambience (from another Bricasti impulse response, ‘Studio A’) and a touch of HF ‘air’ boost from SonEQ Pro’s 16kHz shelf completed the kit sound, and I also drove it quite hard into Cytomic’s The Glue SSL-style bus compressor, registering 4-5 dB of gain reduction at a 4:1 ratio to generate a bit of level pumping — something that’s usually beneficial for most rock styles. The downside of heavier bus-compression settings like this is that they can dull the attack ‘edge’ of the kick and snare, so I used Reaper’s routing options to feed a little of those sounds past the compression in order to mitigate that.
Compared with the drums, the guitars were a walk in the park. On many projects I find myself having to play around with phase adjustment to get the best out of the guitar-amp multi-mic rigs, but here most things already seemed optimised, and once I’d discarded one of the amps used on the original session (which Tommy had already told me he didn’t really like at the time), there was very little EQ required.
As is par for the course with close mic placements, most of the tracks needed some high-pass filtering to counteract the proximity effect, but otherwise the only other significant EQ task was divvying up the available low mid-range mix real-estate between the different guitar layers and the bass guitar. From an aesthetic standpoint, I also applied a little 820Hz boost with Stillwell Audio’s Vibe EQ, as that’s a frequency that often helps guitars to sound crunchier without scratchiness, and applied 2-3 dB of peak limiting from the same developer’s The Rocket compressor, as well as using its Impetus control to add some assertive-sounding Urei 1176-style FET distortion components.
The only part where I drew in a bit more processing was the solo guitar, which seemed a little soft-sounding, and lacked the grit to push through the rest of the drums/guitars barrage, even with The Rocket’s Impetus control at maximum. This wasn’t a simple mix-balance issue, though, so my normal conservative EQ approach went out the window as I set about remoulding the essential character of the timbre.
Pushing a Sonimus Britson Channel well into the red created some useful extra mid-range harmonics, while Stillwell Audio’s 1973 and Vibe EQ plug-ins massaged the frequency response, most notably with a 4dB boost at 6kHz. That’s a frequency at which most guitar amps actually put out very little information, but bear in mind that real-world 6kHz peaking filters also affect the spectrum well below their nominal centre frequency. Finally, I also applied a bit of dynamic-range reduction just to the lower mid-range, using Toneboosters TB_Flx in order to retain as much warmth in the tone as possible (a specific request of Tommy’s), whilst also preventing some of the fuller-sounding licks from muddying the mix as a whole.
The bass sound was also very respectable on its own terms, once I’d found a good phase relationship between the mic and DI signals, but its powerful spectral contributions in the 80-150 Hz region put it into direct conflict with the low end of the electric guitars. Had this been a pop track, I might have cleared more low end from the guitars, but here I decided I didn’t want to sacrifice any more of their subjective weight, so instead scooped a big EQ peak out of the bass sound at 135Hz, and applied SoundRadix Surfer EQ 2’s pitch-tracking EQ to boost the line’s fundamental frequencies instead, thereby simultaneously clearing room for the guitars and extending the apparent low-end reach of the mix as a whole.
Furthermore, I added some distortion to the bass too, courtesy of Melda’s MAmp plug-in, band-limiting those extra harmonics with high-pass and low-pass filters from Reaper’s ReaEQ. My thinking there was that it could enhance the mid-range ‘growl’ of the guitar texture as a whole, as well as making the bass cohere more convincingly with the electric guitars and emphasising the more melodic aspects of the riffing (a big part of the appeal, musically) in the true mid-range — an important consideration for small-speaker mix translation.
Some musical genres are practically defined by intricate effects work, but with rock music I think it’s often a case of ‘less is more’. The more you can get the mic signals to fill the speakers on their own, the more powerful the band tends to sound. Indeed, most people who approach me for advice because they’re unhappy with their rock mixes usually seem to have over-egged the effects from a mistaken belief that it’s ‘the done thing’. Granted, if you need effects for a specific technical reason, then by all means add them. After all, that’s exactly what I did here with the two drum reverbs, and also when I added a simple quarter-note tempo delay to some of the guitar riff layers to enhance their sustain. But otherwise you might as well reserve the effects, mostly, for more creative ends.
On this project, for instance, I took a cue from some of Tommy’s reference tracks, and applied a bit of crusty-sounding spring reverb from Overloud’s SpringAge plug-in to the main guitars, not as any type of generic blending effect, but as a way of imbuing them with more of a rough-hewn, rootsy character.
For the guitar solo, I went for a simple eighth-note feedback delay using The Interruptor’s classic 32-bit Tape Delay plug-in, the idea being to lend it an expansive ‘guitar hero’ attitude. But nowhere was my effects use more obvious than on the lead vocal, where I went heavy on that old standby, the crunchy slapback delay, using different delay times for the left and right channels (95ms and 115ms respectively) to give it a kind of Haas-effect width. One great thing about this effect is that it has, by nature, a built-in pre-delay, so you can pile it on for the sake of personality without unduly distancing the singer in the mix. But it has a further advantage which is perhaps less obvious. You see, when you’re trying to get a lead vocal to both compete against and blend with thrashy drums and overdriven guitars, it’s often very useful to add distortion to it.
In this mix, that’s exactly what I did, both subtly with The Rocket’s Impetus control and Sonimus’ Satson Channel, and much more drastically via another totally cranked instance of Satson Channel in a parallel-processing configuration. What you’re always fighting with under such circumstances, though, is that the distortion tends to make the singer’s lyrics less intelligible, so there’s a limit to how much you can hype the distortion character before the words become an indistinct mush. If you apply extra distortion to a slapback delay instead, however, you actually get most of the distortion’s subjective advantages in terms of rock-and-roll attitude, but with less collateral damage to the lyric intelligibility. Who cares if the delay return signal has smeared words? If anything it just makes it less likely to confuse the dry signal’s lyrical message.
Tying up the loose ends to finish up this mix wasn’t too tricky, by virtue of the simplicity and effectiveness of the song’s arrangement, so there wasn’t the same need for detailed automation of the kind I sometimes indulge in on more chart-oriented projects. As you’d expect, the tom fills needed riding a little, given that their mic levels had been set according to the ‘fake overheads’ role of their spill signals, but that didn’t take long. I did a few rides on the bass and guitars to bring out the more interesting details of the riffs from time to time, as well as to refine the balance between the different guitar layers, but that was hardly very involved either, and even the lead-vocal rides were less in-depth than usual, on account of the guitar texture thinning out for the vocal verses.
About the most leftfield thing I did, in fact, was automating a high-pass filter on the main rhythm guitar part, so that the verse ‘chugs’ could have more low-end weight than the riff sections. As a final touch towards the end of the mix, I used Toneboosters TB_ReelBus tape emulation in my master bus to smooth the high mids a touch, as well as adding a few spots of EQ from Fabfilter’s Pro Q to bring the overall tonality closer to Tommy’s commercial references.
In this month’s remix, I’ve demonstrated how a bit of lateral thinking can sometimes help you work around drum-tracking problems, as long as you keep an open mind and are willing to take the track labelling with a pinch of salt! (For another example of this, check out SOS April 2008’s Mix Rescue, in which the snare drum sounded better on the kick-drum mic than on its own close mic.) But I hope I’ve also been able to show that reverb and delay effects don’t have to be applied to everything in your mix. Indeed, applying them more sparingly not only helps keep the mix sounding clear and upfront, but also increases the subjective impact of the effects you do actually end up using.
Tommy Allen: “Loving it! The guitars sound great and the energy from the drums is amazing. I like the vocals being dirtied up too, and the ‘wiggly’ delay/reverb on them. There’s lots of other reverb things going on within the mix which I like, although I’m unsure how to explain what I’m hearing! Enjoying the movement of the mix too, and the shakers sound really cool. I see what you mean when you said the previous engineer’s mix was boxy, because yours sounds more open. I also listened on all three speaker systems I have here and I’m surprised how deep the kick and bass come across on my Avantone Mixcube — love it! And on my mini hi-fi speakers it packs a punch too. Thank you for the time spent on this, Mike!”
Besides the core drums, bass, guitar and vocal tracks, there were also a couple of other little production ‘sprinkles’ that needed fitting in. The first was the pair of shaker tracks, each played with a different instrument. These felt a bit too bright for me. It’s a misconception, I think, that hand percussion has to be very toppy to cut through the mix, because I find that meatier-sounding percussion timbres are in fact a lot easier to deal with — you can fade them up far enough to increase the groove’s rhythmic energy without their high frequencies pulling them distractingly far forward in the depth perspective.
So, as I often do, I rolled these percussion tracks off progressively above about 2kHz using a very gentle low-pass filter, and also applied a few decibels of fast limiting from The Rocket to contain the odd overzealous onset peak, which otherwise poked out of the balance unduly. In addition, I routed the percussion parts through the drums bus compressor and added a little of the kit’s overall reverb to them as well, and that further glued their sound in with the drums as a whole.
The other mix sweetener was a Hammond organ pad — a classic choice for this kind of music. However, this was one area where I felt the arrangement could have been a little more developed, because this pad appeared pretty much the same in each of the three verses, losing a small opportunity for creating some longer-term build-up.
Simply muting it from the first verse seemed like an obvious improvement in this respect. While considering this, though, it also occurred to me that a Hammond organ might also provide an arrangement ‘lift’ for the final section of the chorus too, but there was nothing provided for this in the raw multitrack.
So I programmed something using NI’s Kontakt in conjunction with FX Point Audio’s Spinner LE rotary-speaker plug-in, and that seemed to do the trick. Although adding extra parts to any production at the mix stage isn’t something I ever take lightly, in this instance I figured I was probably on safe ground as long as my new part didn’t draw too much attention to itself or deviate too far from the spirit of the supplied verse-section part.