This month's mixing challenge was to retain the energy and excitement of a live recording, while adding a healthy dose of vintage character.
Singer/songwriter Clive John is the man behind the Johnny Cash Roadshow, which must be a strong contender for the most popular Johnny Cash tribute show in the UK; he's constantly on tour and covers all the hits in truly authentic style. Clive sings and plays acoustic guitar and is backed by drums, bass (upright and electric) and electric guitar, in addition to a full quota of 'Carter Sisters'. However, he's also had some success as a songwriter in his own right, and occasionally slips one of his own songs into the show.
A few months ago, Clive recorded a complete show, capturing a multitrack feed from the front-of-house console on an Alesis HD24 hard disk recorder. The idea was to produce a live DVD and album, but as that project included only Johnny Cash songs, he was left wondering what to do with the multitracks of his own song, 'Spirit'.
I suggested to Clive that it would be a good track for me to tackle as part of our Mix Rescue series, because live recordings often present rather different challenges from those made in the safe confines of a studio. For one thing, there's invariably a lot of spill into the vocal mics from the backline and monitors, and when, as in this instance, an acoustic guitar is DI'd using a pickup system, the tonality is usually not up to the standard you'd expect from a nicely miked guitar in an acoustically sympathetic room. You also have to remember that the live-sound engineer might not have found it necessary to mic up everything you'd need in a recording, as in many venues the mixer is there to balance the PA with the sound of drums and amps used on stage. In short, then, I felt there should be some interesting mixing challenges.
Initial Assessment
On examining the individual tracks, I found that the main vocal mic had indeed picked up a significant amount of spill. The acoustic guitar was clean enough, but had that slightly 'quacky' piezo tonality, and while the electric bass DI was also clean, its sound didn't really cut through the mix. There were also a few noticeable timing errors, albeit fairly small ones, as well as some fluctuation in tempo as the song progressed. It seemed that the electric guitar amp had been miked rather than DI'd, but it came over as quite thin and 'peaky' sounding, and I could tell that this would also require some tweaking.
Drums are always an interesting challenge. In this case, the kit appeared to have been recorded using a pair of overheads, two tom mics, a snare mic and a kick mic. One of the overheads seemed to have picked up a lot of snare and hi-hat, so I suspect that this had been placed closer to those kit elements than would be normal in a recording studio. The kick, snare and tom mics displayed reasonable separation, but the sound of the kick drum was rather disappointing, and the snare came over as being quite harsh. That's not untypical for this sort of live recording, but clearly some work would need to be done here!
Too Much Polish?
With any live show recording, a decision has to be made on just how far you're to take the polishing process, as too much tinkering can strip away the live feel, leaving the track sounding a bit unnatural. Timing correction is a case in point: on the one hand, you don't want there to be obvious mistakes, but on the other, you don't want to introduce a robotic, artificial feel, and correcting sources with lots of spill can be problematic from a technical point of view. With Clive's recordings, it wasn't practical, or even desirable, in my view, to address the general tempo fluctuations, but I did tighten up a couple of individual hits, just to keep the kick and bass together during exposed sections. This was done using the simple method of slicing the track into regions and then lining up the bass-guitar notes with the kick drum, using the waveforms in the main Logic Arrange page (and then checking with my ears!).
There's a section at the start of the song where the bass first comes in playing a few bars of reinforcing notes to underpin the acoustic guitar. These notes were recorded at radically different levels, and I felt that they should ideally have been consistent, so I isolated the notes and tweaked their gain until they matched up. This also brought up a little background noise in places, so I applied a fade to the end of each note to make it less obtrusive.
Having made those slight adjustments, the next step was to isolate and either mute or remove any track sections where nothing was playing. For me, this is a key part of mix housekeeping, and it tends to get done before starting any creative work. This was particularly important in the case of the lead vocal; the level of spill from the backline was very significant, and this meant that whenever Clive stopped singing, any compression that I applied to the part would otherwise pull up that spill. Using an automated gate plug-in was out of the question, because the sheer level of spill would have led to mis-triggering. Level automation would have been another way of handling the pauses between vocal lines, but I prefer to do this using a gain plug-in rather than the channel fader, and such things aren't easily visible from the mixer or Arrange page views, so, as with the bass notes earlier, I opted to slice the track into regions. To stop transitions into and out of 'spilly' vocal sections sounding too abrupt, I left half a second or so at either end and applied fades in and out.
Pitch Perfect
Clive's vocal pitching is generally very good, but it wavered a little on the entry to the repeating refrain towards the end of the song, so I inserted an instance of Logic's pitch-correction plug-in and automated its bypass switch to bring it in only where needed. You need to be particularly careful when pitch-correcting a vocal that has audible spill, as the spill also gets pitch-shifted, but in this case I decided we could get away with it. As usual, the pitch correction was set only to those scale notes needed for the part, which for this section was E-minor.
The backing-vocal tracks, which don't come in until the repeating refrain at the end of the song, were 'cleaned up' in a similar way to the lead: even on tracks where the amount of spill seems low, it can really help the clarity of a mix to reduce it further, especially in exposed sections of the song. As this arrangement starts with solo acoustic guitar plus vocals, I took the precaution of muting all non-contributing tracks until their instruments made an entrance.
Another essential stage of my mix housekeeping routine is the subgrouping of things that are best controlled via a single fader, such as the drum kit and the backing vocals, and that's what I did here. There were three male and three female backing vocals, but the pitching on two of the male vocal parts, supposedly singing in unison, was way off and beyond any kind of electronic redemption. I simply muted the offending parts and then resorted to trickery to fatten up the remaining good male backing vocal. In order to preserve some sense of mystery and anticipation, I'll not let on exactly what that trickery was until a little later!
Drums
Sorting out the drums took quite a while, as I wasn't sure how the mics had been set up at the gig and I wasn't happy with either the kick or snare sounds. Experiments suggested that the kick could be EQ'ed to give it more punch, but as I didn't think it was ever going to be great, I decided to replace it using a sampled kick from Toontracks' EZ Drummer Nashville kit. There was enough separation to allow Logic's Pitch to Score function to extract the kick hits reliably, once I'd adjusted the threshold setting, after which the MIDI notes (which tended to vary as the software tried to establish a pitch) were all transformed to a fixed note and then moved to C1 to trigger the replacement kick. I also adjusted the MIDI dynamic range to achieve the more consistent level I felt the song required.
The snare drum came over strongly in both the snare close mic and the overhead mics, but the close mic had a brash, rather unattractive quality to it, almost as though the snare had been miked only from beneath. Heavy mid-range EQ cut improved it, but again I found that it was easier just to replace it. The snare from the same Nashville kit matched the snare sound as picked up in the overhead mics reasonably well. I did mix in a little of the original snare to keep the bright attack, but only as much as I might with an under-snare mic. Neither tom track presented any particular problems that EQ couldn't deal with by boosting the fundamental pitch and notching out some unwanted ringing, so other than silencing these tracks when no toms were playing, little needed doing to them.
Electric, Acoustic & Bass Guitars
A simple fix for the brittle electric-guitar sound was to insert an amp-modelling plug-in and use only its speaker emulation section.By trying a few different virtual cabinet and mic setups in this way, I was able to achieve a sound that still rang out nicely but sounded fuller and less nasal. Cabinet emulators come in handy for all kinds of fix-up jobs where basic EQ doesn't seem to do the trick, and a 2x12 cabinet with dynamic mic emulation worked fine this time around.
I'm not certain whether the bass signal came from the back of the amp, or straight from the guitar via a DI, but it sounded too clean to me, without enough mid-range presence to sit well in the track. My fix here was to use a UAD Dbx 160 compressor emulation, which I adjusted to give around 10dB of gain reduction on the louder notes to even it up. I fed the resulting signal through Softube's Bass Amp Room plug-in, selecting the 4x12 model. That's a really simple little plug-in but it sounds great, and it added the required mid-range character, although once compression had been applied, some of the notes were sustaining rather too much. SPL's Transient Designer provided a convenient solution: one adjustment was all it took to dial back the decay time to create a tighter bass sound. Not only did this create a more authentic and punchy sound, but it also cleaned up some fret noise that had been audible during the note decay.
That left the acoustic guitar, which was a little mid-heavy. After trying straight EQ, I managed to get a good subjective result using multi-band compression, hitting the lows and mid-range a little harder than the highs. The compressor band levels could also be adjusted to add a little high-end zing, and to scoop back the mid-range slightly. Logic's own EQ was also used to cut out unnecessary lows and to add some air at 7kHz and above. While the result still didn't match that of a studio recording, it did at least come over as well balanced and supportive. Only one edit was necessary, where a slightly choked note during the intro was replaced using a copy of the same note taken from the next time that phrase repeated.
Vintage Vibe
With all the basic tidying and broad tone-shaping jobs done, I could move on to the more artistic aspects of mixing — and it's incredible just how much easier these become if you've paid attention to all the little housekeeping tasks I've described above!
My main aims were to retain the live feel, and to achieve a retro sound that fitted in with the Johnny Cash vibe Clive captures during his shows. The song itself is about the human spirit, so I thought making the three female backing vocals sound a bit 'church-ish' by adding extra reverb would work well. To this end, I set up PSP's new SpringBox spring reverb on Logic's first aux send, with the diffusion set fully up to reduce the 'twanginess'. I wasn't sure this would do what I wanted until I tried it, but it sounded perfect for the song and made a nice change from the more usual plate or generic synthetic reverb. The backing vocals were mixed via a subgroup, where they were compressed as a single element.
To achieve the vintage vocal sound, I used a little compression, courtesy of UA's LA2A plug-in, combined with some very gentle EQ, just to cut the lows and to add some air above 5kHz. During the intro, there was a suggestion of popping at the start of the word 'past' so I automated a low-cut filter, bringing its frequency up to 275Hz only during the offending plosive. That's a simple but really effective trick in such circumstances. An instance of Waves' Vocal Rider plug-in was used to iron out most other fluctuations that were caused by mic distance variations during performance, while Logic's PlatinumVerb was set up to provide a burst of early reflections, delayed by 94ms to create the slap-back effect that typifies this style of music. A little spring reverb via the aux send I'd set up previously completed the picture. Some simple level automation was needed to make the vocal level more suitable in the stripped-back parts of the song, but nothing forensic was needed.
Chorus Tricks
Mixing isn't a linear process, and you often have to go and revisit things you thought you'd perfected earlier, tweaking until you're happy that things sound right. As this mix progressed, I noticed some very slight tuning issues with the electric guitar in places, but rather than reaching for a pitch-correction plug-in, I opted to use a more organic-sounding chorus effect. I did this by copying the track, delaying it very slightly and then inserting a pitch-correction plug-in on the copy, setting it to track fairly slowly. This introduced just enough non-cyclic pitch variation to add a gentle shimmer to the sound. I mixed the treated copy track lower in level than the dry track, which ensured that the result wouldn't sound obviously processed.
This brings me back to the backing vocal thread I left hanging earlier... because I used the same trick on the one good male backing vocal part to add a double-tracking type of effect. In this case, though, I used an equal balance of the two tracks. Panning the two versions left and right added a sense of stereo width. Again, level automation was used to balance the guitar against the vocals during the verses and to bring it out for those sections that feature instrumental interludes.
Thundering Applause
The song's ending segues into a thunder sound effect, which had been picked up quite clearly by the main vocal mic, and I faded this mic's signal beneath the audience applause mic to help it blend in.
All that was left then was to fine-tune the balance between the various song elements, and to decide what, if any, final bus processing would be required. I'm a firm believer that balancing and bus processing need to be performed at the same time, simply because any overall dynamics processing on the bus will affect the subjective mix balance. Having auditioned a few of my preferred processors, I ended up using PSP's Vintage Warmer on the bus, set to achieve about 3dB of gain reduction, and the signal then went through an instance of Slate Digital's Virtual Tape Machines plug-in, which I set to 16-track, 15ips, and used the input gain to push the virtual meter needle 3 or 4 dB into the red. This smoothed out the raw edges of the sound to a useful degree, and helped to glue the parts together. Finally, I placed a UA Precision Limiter at the end of the master-bus signal chain, just to trim a dB or two from the loudest peaks. Job done. Or so I thought!
Studio Sound
I was already pretty happy with the mix, and was about to call it a day when Universal Audio announced that their new Ocean Way plug-in (see the review in last month's SOS) had become available. As I was now familiar with this mix's intimate details, it was the perfect candidate for testing out what this plug-in could do. I used around 30 percent of its drum room setting, with both mid and far mics mixed in, just before the mix hit the limiter — and to my ears it created a very convincing liveness, with an appropriate sense of space between the listener and the performers. I've included mixes with and without it (see the 'Audio Examples' box) so you can hear the results yourself.
Ultimately, this is a live recording, and I've deliberately tried to retain a little of that 'rough around the edges' character that distinguishes many such performances. I think it all came together pretty well, successfully capturing the vibe of being there, especially in the version processed via the Ocean Way plug-in.
Audio Examples
We've placed a number of audio examples on the SOS web site so that you can hear for yourself the difference Paul made to this mix, and better understand the effects and processes he has described in this Mix Rescue.
Reader Reaction
Clive John: "I really like what you've done with the mix — it's like it has been approached with a fresh pair of ears, and it has a more contemporary feel to it. I also quite like the Ocean Way room effect you put on it, but preferably dialled back a bit, so it isn't quite so pronounced. The electric guitar has a nice effect on it, and the replacement drums definitely sound better than the originals. Those layered backing vocals at the end of the song sound pretty epic, too. I can't wait to put it up on my web site.”