This month we sort out timing issues, replace drum beats and rescue some old-fashioned guitar parts.
The band Exhibit-1 consists of Christopher Hopkins on guitar and vocals and his wife Rachel Hopkins on bass. They recorded the track that's up for a Mix Rescue at Shabbey Road Studios in Caerphilly in 2006, with Chris Jones filling in on drums and Al Steele engineering. The backing tracks were recorded live to a click with Chris on acoustic and guide vocal, Rachel on bass and Chris on drums in a separate booth. Chris then took a rough mix home and dubbed on the additional instruments using his Fostex VF160, with a Boss GT8 providing guitar effects.
Chris then took the WAV files back to Al Steele, who loaded them into Pro Tools and mixed them, in line with Chris' suggestions. The track ended up on the band's self-financed debut CD, It's Better To Travel Than To Arrive, released last spring. Chris said: "We only had a certain amount of money to record and mix the album, so we rushed a bit on a few of the tracks, and 'Days Like These' possibly suffered the most. We give huge credit to Al Steele for engineering and mixing our stuff. Had we had a little more time we could've made it better, but he did an amazing job considering the limited time he had to do it. This is why I sent 'Days Like These' to Sound On Sound. A fresh set of ears always helps."
On loading the files into Logic and listening to the original MP3, I thought that 'Days Like These' was actually pretty well mixed, so I decided to ask Chris what it was that he wasn't happy about. He told me he'd left out a chorused guitar part because he didn't like the chorus effect — it sounded a bit dated — and there was also a rhythmically-chopped synth part that he'd left out but now wished he'd used somehow. As regards his choice of sounds, he wasn't too happy with the way the snare drum turned out, as it wasn't crisp enough, and he'd noticed a snare beat missing altogether about a third of the way through the song. The band recorded the track against a click, presumably to make editing easier, but Chris had noticed what he describes as 'a slight lurch in timing' at the end of the first chorus, and he wondered if there was anything we could do about that. He also felt that a line from the chorus was sung better first time around, so he wanted to try pasting this into the second chorus. Chris admitted to liking the production of bands such as the Kaiser Chiefs, Manic Street Preachers and Franz Ferdinand, but was also realistic enough to concede that this song might not fit the sound of those genres.
The song follows a more or less conventional guitar-band format but with a couple of additional supporting keyboard parts and a number of guitar parts (and overdubs), including an acoustic rhythm guitar, a chorused electric guitar, a fuzz guitar, one main electric guitar line and two further electric guitar riffs that come in at strategic places throughout the song. The acoustic drum kit was recorded using stereo overheads and close mics, including a separate hi-hat mic. The main vocal was underpinned by a second part, sung an octave below, and a harmony part that comes in during two specific sections.
I do these remixes in Logic, as it is the software I feel most comfortable using, so you'll see it mentioned plenty of times in this article. But the techniques used are valid for any multitrack DAW system, and most can be done using hardware if that's your preferred method of working.
My first job was to sort out the drum sound. The original kick sound lacked both weight and definition, probably due to mic choice, mic placement or drum tuning. This would require some fairly heavy-handed EQ on the kick, with a touch of Logic 's Exciter to bring out the beater slap. The final EQ setting comprised a 75Hz boost, a 182Hz dip to reduce boxiness and a 3.5kHz boost to bring out the slap, followed by further enhancement courtesy of the Exciter plug-in. To beef up the snare, I used Logic 's Pitch Shifter to add a little sub-octave presence, on the 'Drums' setting (to avoid time delay or flamming issues), after first gating the track to get rid of as much spill as possible. The Exciter plug-in was again used to brighten up the sound from the snares, and Logic 's Limiter was put at the end of the chain to even up the beats and to prevent the channel overloading on loud hits. I also fed the snare's aux send to the UAD1 Plate 140 reverb I'd set up for the vocals.
As expected, all three tom tracks had to be gated to kill the constant sympathetic ringing that is picked up by the close mics, but other than a bit of plate reverb, nothing else needed doing to them. Rolling off the really low bass-end cleaned up the overhead mics (12dB per octave at 220Hz) and I also put in a 250Hz notch to kill some room boxiness. A hint of broad 12kHz boost (just 1dB) was added to enhance the cymbals and we were done. Additional tambourine and shaker parts were balanced and given a little of the plate reverb, but nothing else needed doing to them.
Next came the bass guitar. This sounded a bit bottom-heavy in the mix, so I thinned it out a bit by rolling off steeply below 50Hz and above 7.7kHz. Some compression was used to even up the level, and a hint of 240Hz EQ dialled in to bring out the mid-range character of the sound. Heard on its own, the 200-300Hz range of the bass guitar can sound a bit boxy, but once the mix is playing you'll often find that boosting in this region makes the part more audible.
We've posted Jon's and Paul's mixes and some before and after MP3 files of individual tracks on the SOS web site at: www.soundonsound.com/sos/aug07/articles/mixrescueaudio.htm.
Next, I turned my attention to the guitars. I don't know what processing or pedal effects had already been applied, as all the guitar tracks for this project were supplied fully processed.The main vocals had been well recorded and didn't need much processing. In addition to some gentle compression, Noveltech's Vocal Enhancer was used to give them a little more presence in the mix.
To stop the chorused guitar (mentioned earlier) from sounding too dated, I put it through a rotary speaker plug-in (set to a slow speed) and then applied a broad 4kHz boost to make up for the top-end loss caused by the rotary emulation. This produced a really nice 'liquid' guitar sound that sat well under the mix and disguised the chorus pedal quite effectively. I was also a little worried by the brittle tone of the acoustic guitar part, so I used the SSL Duende Channel Strip plug-in to apply compression and EQ. A little mid cut, combined with top boost, added the necessary zing, which I fine-tuned when the rest of the track was playing.
The Fuzz Guitar track sounded almost like a DI'd fuzz box to me, with a lot of brash overtones that just got in the way, so I fed that through the Line 6 Gearbox amp-modelling plug-in (Mississippi Criminal Amp model with 4 x 12 speaker), which produced a much more believable amped sound that didn't take up too much 'frequency space' in the mix. Guitar 1 was a much more rounded distorted guitar sound that worked fine just as it was, so I left it alone.
Riff 1 turned out to be a very 'chimey' overdrive sound, and all I did here was apply some low cut (24dB per octave at 124Hz) and some high cut (24dB per octave at 10kHz), augmented by a gentle 1200Hz boost. This cleaned up the sound and added a bit more sparkle, while filtering out unwanted high-end grit.
Riff 2 was a much more gritty, distorted sound, and in the mid section of the song it included some very Frippy-sounding harmonics that had to be dropped in level by a few dBs to stop them dominating the song. To sweeten this sound and to add focus, I again used high- and low-cut filtering to home in on the mid-range, but then fed the signal through Logic 's own overdrive plug-in, set fairly mildly to add a bit more edge. Without this processing the sound seemed to have a lot of low end and a lot of high-end fizz, but once the rest of the mix was up it got lost too easily.
Synth 1 appeared to be a fairly basic sawtooth type of sound, chopped using a square-wave tremolo to form a rhythmic element. This part sits under the song most of the time but becomes more exposed right at the end, and I panned it from right to left as the last note faded. A second synth played a string-like high line that I felt sounded a bit too confrontational most of the time, so I used the UAD1 Roland Dimension D plug-in (a kind of super-deluxe chorus) to soften it a little, and then automated its levels so that it was most noticeable during the intro to the song, before the vocals kicked in.
The vocals were pretty well recorded, and there was no obvious processing on the tracks I received (other than a little 'nip-and-tuck' courtesy of Melodyne, to tighten up the pitching). A second vocal line doubles the first on certain phrases but is an octave lower, giving the whole thing a bit of a 10cc flavour, with a further, more conventional harmony line during the high section following the main chorus. As an experiment, I copied the harmony line to a new track and used the Antares A-Vox Choir plug-in to simulate four voices from the one part, then added this back under the dry version to thicken it slightly, without allowing the effect to become too obvious. Used sparingly it worked nicely, so I left it in.
As an effect, I also took a couple of short phrases from the harmony section and then treated them to a 100 percent wet reverse reverb (Space Designer with the IR reversed), sliding the parts forward in time so that the end of the reversed part coincided with the start of the un-effected vocal part. I'd originally wanted to use this reverse effect to bring in the chorus as well, but Chris felt it detracted from the impact of the chorus —he'd experimented himself with trying to find a sound effect that would fit in at this point but he'd never found anything he was happy with. Aside from a little compression to even up the already well-recorded vocals, I also used the Noveltech Vocal Enhancer to add a bit more presence to them, to help them cut through what is actually quite a busy mix. Reverb came from the UAD1 Plate 140 (a great emulation of an EMT plate), with a two-second decay time that kept the vocals from sounding too wet. With so many guitar parts, the mix could easily have got very cluttered, but fortunately Chris' arrangement avoided too many things playing at the same time.
Before starting the mix, I decided to fix the missing snare-drum beat by copying one from elsewhere. When doing this kind of edit, remember to copy the overhead mics as well as the close snare mic, as overhead mics make a big difference to the sound of the snare. It is important to copy them together, so that you don't end up with slight timing differences. With the missing beat corrected, I grouped all the drum tracks — so that if any drum part was selected for editing they would all be selected. Again, this prevents the situation where you end up accidentally moving one part slightly out of time with the others. The timing issue that Chris had spotted was easily resolved by moving the snare-drum beat that came immediately after the break. This entailed slicing through the drum parts either side of it and then nudging the section ahead in time until it sounded right. It didn't need to move far to give the desired improvement.
Once I'd set up what I thought was a good drum balance, I routed all the drum parts, including the shaker and tambourine, to their own subgroup, so the track could be adjusted using a single fader. Noveltech's Character was used very sparingly to add a bit of air and definition to the overall kit sound, then Logic's Space Designer was placed over the whole kit, with a half-second drum-room impulse response loaded up. This gave the whole kit a nice, lively sound, without making the reverb too obvious. The only send effect I used was the UAD1 Plate 140 two-second reverb, but I also passed the whole mix through my Drawmer 2476 Masterflow mastering processor to add some very low ratio, three-band compression and some gentle, three-band tube emulation. A Waves L2 limiter took care of the final output, though I tried to avoid going more than a dB or so into limiting on the peaks, as I don't like trading off loudness against quality.
Most of the balancing could be set up and left alone, though I did need to automate the chorus guitar track level to pull it out of the way during the vocal passages, and I needed to dip the level of the high synth part after the intro. I also took the opportunity to mute a few sections where there was guitar noise but nobody was playing. Once the sounds could be heard in the context of the full mix, I did some final juggling of levels and revised some EQ settings.
The end result was a very big-sounding mix, but still with good separation between the various parts. The drums now sounded punchier, and the guitar parts much more solid. The quirky, Fripp-like guitar line in the centre section really worked to give the song a more 'arty' edge and the vocals came over as strong and confident. There is always a limit to what you can do with already-processed guitar parts but, after spending a couple of hours on this mix, I was pretty pleased with the result.
Chris said: "Paul's preliminary mix was 80 percent great, but I felt the reverb on the vocals wasn't what I had envisaged. Paul sent the second mix shortly after receiving my comments and it is fantastic. I love the sparkly bite the snare has now and, of course, the worst of the timing glitches have been fixed. My first reaction on hearing the opening drums was "Wow!", and I also love the sound on the guitars — less 'fat'. The part at the end of the first chorus, going into the verse (which used to cause an awful lurch), has now been magically restored and the space created by EQ'ing the guitars works a treat. I love the end."