This month the team travel to Wales to mix vocals, acoustic guitars, and a troublesome hurdy-gurdy!
There's no denying that we get less snow in England than we used to, but whenever it does snow, you can bet that Hugh and I will be out in some remote corner of the country doing a Studio SOS visit. This time is was to visit Paul 'Spyder' Fitzpatrick, who'd relocated his studio from his native Manchester to a mobile home outside his friend's farmhouse, on top of an 850-foot hill just south of the Snowdonia National Park in Wales. The climb to the top of the hill was via a winding, snow-covered dirt track, with a steep slope offering slithery oblivion on one side, and innumerable kamikaze sheep on the other! We needed GPS navigation to find the place, as the nearest village had a name that looked like a bad hand of Scrabble with all the vowels removed, but Paul welcomed us with the obligatory Hobnobs and hot coffee while we chatted to find out what kinds of problems he was having.
Paul's studio, based around a Digidesign Digi 001 interface and Pro Tools software, is set up at one end of the lounge of his residential caravan. A corridor runs down one side of the structure to the lounge area and, as Paul has his studio set up along the wall where the corridor enters the room, it is offset to one side of the room rather then being symmetrical. Ergonomically, this seemed the best place for it, so we decided to leave it there. Paul was a little worried that his passive Celestion F1 monitors might not be accurate enough for monitoring, and the lack of acoustic treatment was affecting both the quality of the monitoring and also his ability to record vocals in the same room without suffering coloration from the wall and window reflections. The light curtains fitted over the windows weren't really adequate to control the acoustics, so Paul's only real concession to acoustic treatment had been to place some of the furniture cushions behind the monitors.
Although fairly simple, the studio includes a surprising amount of outboard gear, as Paul mixes through a Behringer 8000A analogue mixer rather than within Pro Tools. Partly this is because he likes this way of working, but it's also because his 400MHz Mac G4 stalls under the load if he uses more than a small number of processing plug-ins. To expand the number of physical outputs from the Digi 001 Paul had added a Fostex VC8 expander, so that he can have 16 independent feeds going into the mixer from Pro Tools.
For effects, Paul has an old Alesis MIDIverb unit that he likes for its crunchy reverb sound when overdriven, and he also has a TC Electronic M*One XL dual-engine reverb for more conventional applications. Other effects come from Line 6 delay and filter units, and he has an original Line 6 Pod for recording guitar, though he prefers to mic up his old Watkins Dominator amplifier. Processing comes from an Alesis 3630 compressor and a Behringer quad gate, while any synthesized sounds triggered over MIDI come from a Korg NS5R sound module. As ever, the first task was to play some test music and see what the monitoring system sounded like.
The little Celestions were being powered from a Cambridge A1 hi-fi amp, and despite their size they sounded tight and detailed with a sensibly flat tonal balance, albeit without any real deep bass. However, the imaging was being compromised by reflections from around the room and by scattering from the computer VDU and the effects rack between the speakers. The speakers themselves were propped up on cardboard boxes rather than being on proper stands.
There was little we could do about this on the day other than adjust the speaker heights and angles by fitting Auralex Mo Pads beneath them, but, as Paul is clearly a very practical person, he can easily build a more substantial shelf from MDF or chipboard and then use solid blocks or purpose-built MDF boxes to give the speakers the height they need. In an ideal world, the speakers should also be slightly forward of the computer monitor to eliminate reflections from that source, but with the present setup, this wasn't practical, so we settled for getting the monitors as far forward as we could.
To avoid damaging the walls, we propped up the Auralex foam we'd taken with us just to demonstrate the effect it had, and Paul agreed to find a way to make the fixings more secure while still being removable. A pelmet runs around much of the room, so our old friend the picture-rail hook could be used here. For the side walls and ceiling, Paul planned to use Velcro pads, which seemed as good a solution as any. The arrangement we arrived at was one panel either side of the monitoring position at head height, one fixed to the ceiling above the mixer and three suspended vertically over the window that covered most of the back wall of the room. By hanging these on the pelmet, there would be an air gap of three to four inches behind the foam that would increase its effectiveness at low frequencies. My suggestion was that any vocals could be recorded with the singer standing with their back to these panels to cut down on reflections. Paul has a Rode NT1A as his main vocal mic, so setting this to a cardioid pattern should give good results in this situation.
After making these adjustments to the room and the monitor positions, we played our test tracks again and found the imaging to be very noticeably improved, and the Mo Pads had evened out the bass end, as well as reducing some of the resonances of the cardboard boxes and shelf, although the speakers still lacked deep bass. When the budget allows, Paul plans to replace his speakers with active monitors (or at least add a suitable subwoofer), but in the meantime, he'll need to avoid the temptation to EQ his bass sounds to make them sound larger than life on these speakers, as that will result in bass-heavy mixes when played on full-range systems elsewhere. Constant referencing to well-balanced commercial tracks while mixing would help avoid this problem, effectively resetting the aural memory of how the mix should sound on those speakers.
One further problem we identified was caused by Paul's use of a low, lightweight, folding chair. This placed his head almost exactly midway between the floor and ceiling, and we noticed that the perceived bass level dropped even further in this position. Raising the seat position by as little as 12 inches avoided this, so a new seat would be a wise investment.
With the monitoring improved, we could now get down to helping Paul with his mix problems, as he'd been working on a track for a female singer and was having trouble interpreting her artistic demands. The track was essentially a piece for vocal and acoustic guitar, but it was underpinned with a hurdy-gurdy drone and augmented by some subtle strings later in the song. There was no bass and no formal drum part, though Paul had used some sparse percussion courtesy of Spectrasonics' Stylus to drive the track. His problem was that the singer wanted the recording to have a free, outdoor feel to it, with an element of distance about the hurdy-gurdy part, but Paul's reverbs sounded too heavy and room-like. He was also unsure about his use of EQ and compression, especially in the light of his distrust of his monitoring system.
We noticed that to play his Pro Tools mix, Paul had to press the Record Pause button on his CD recorder, as he'd wired everything up to monitor via this piece of equipment. He was also using the mixer's main outputs rather than the control-room outputs, which meant he couldn't solo tracks on the mixer and also couldn't control the volume other than via his power amplifier. He'd done this because he suspected that the main outputs were quieter than the monitor outputs, but we decided to try rewiring the system more logically anyway. This entailed connecting the control-room outputs to the power amplifier in the usual way, then connecting the CD recorder as an external two-track machine, fed from the mixer's main outputs and returned via the two-track return jacks. This didn't seem unduly noisy and in any event the CD recorder could still be fed from the mixer's main stereo outputs, so any noise audible on the monitors due to a noisy control-room output stage won't be recorded to the final master. This arrangement would also make it easy to compare reference CDs to his mixes as he builds them up, simply by flipping the two-track return switch.
While the mixer has dedicated aux returns, Paul had instead decided to use the desk 'dub' style, where spare channels are used as aux returns. This has the advantage of greater controllability, and also enables other effects to be added to the effects returns via the channel aux sends. The TC Electronic reverb can be configured in a number of ways. For the current song, Paul was using it in dual-engine mode, where each engine provides a different reverb effect and is fed in mono from one of the two inputs. The two reverbs then each generate a stereo output and both stereo outs are summed at the unit's stereo output jacks.
It is also possible to use the unit as two mono-in, mono-out reverbs, but in most instances stereo works best, so I panned the two outputs hard left and right rather than leaving them in mono as Paul had. Each of the two reverb inputs was hardwired to a post-fade send on the mixer, so essentially it behaved as two more or less independent reverb processors, but with their outputs coming back on a single pair of faders. The relative balance of the two stereo reverbs could be controlled with a knob on the TC Electronic box itself, or just by adjusting the aux send level to each input.
He'd also used the same 'channels as returns' setup for his Alesis reverb and for his Line 6 delay and filter. Strictly speaking, a filter is an insert effect rather than a send effect, as you don't normally want to mix in any of the dry sound, but if this is fed from a pre-fade send rather than from the more usual post-fade send, you can arrange to have a 'wet only' effect by turning the send up and the channel fader down.
Once we'd got Paul's mix up in Pro Tools, I started by listening to the acoustic-guitar track. This had been supplied to him as a WAV file from the artist, so he had no control over how this was recorded — it sounded DI'd to us. She wanted a warm sound, but Paul's EQ treatment had made it sound too woolly at the low end, so I re-EQ'd it using some lower mid-range cut to tame the boxiness that so often creeps in. As usual, we located the problem frequency by creating a narrow boost and then sweeping it through the mid-range until it picked up the boxy honk we wanted to tone down. To get rid of any really low end that wasn't doing anything constructive for the sound, I set the lowest EQ band as a low-cut filter and swept it up to around 80Hz.
Paul had also used some very heavy compression with threshold settings of around -25dB, so there was a lot of gain reduction making the guitar seem unnecessarily squashed. This was remedied by setting the threshold higher so that there was only 6-8dB of gain reduction. Because Paul's computer can't run more than one Realverb reverb plug-in at a time without waving the white flag, we decided to use a sparing amount of whatever reverb we set up for the vocals just to take the edge off the guitar, and as I wanted to experiment using the software reverb on the hurdy-gurdy, this meant using the TC Electronic unit for the vocals and guitar.
Paul had over compressed the vocals, which not only makes the sound appear squashed but also brings up room-acoustic problems in the quieter passages and at the ends of phrases where the level tails off. The EQ he'd used also made the vocals sound rather thin, so I zeroed the EQ to make a fresh start and pushed the compression threshold up to reduce the amount of gain reduction. On the whole I felt the basic vocal sound was good, but there was a nasal honk in the 1kHz region that needed addressing, so I located this using the usual sweep method, then applied a few decibels of cut at that frequency. A hint of 'air' EQ at around 12kHz added a little sparkle and that was pretty much it. The biggest challenge was to find that light and airy reverb that would give the sound an outdoor feel.
In the end I used a combination of two reverbs from the TC Electronic unit: an ambience patch set up on send one and a shortish plate on send two. By combining these, the vocal retained an intimate feel without sounding too dry. Although the vocal pitching was pretty good, there were still a few slightly suspect parts where notes tailed off, so I decided to use a light touch of Antares Auto-Tune to fix this. Paul hadn't really got to grips with Auto-Tune, so I showed him that the main thing is to set a key that matches the vocal part and then to set the correction speed slow enough to get rid of the clichéd 'yodel-effect' that we've all come to know and love — and eventually hate! I generally set the tracking speed to fast, just to check the scale I have picked is correct for the song, and then I back it off to halfway or even a little lower so that the pitch-correction would only be applied to sustained notes. This worked perfectly, so Auto-Tune was applied to the main vocal and also to a double-tracked version that Paul wanted to bring in on the choruses. As the artist had fairly rigid ideas about where any additional backing vocals and string synth parts should come in, I left these for Paul to arrange with her.
Paul Fitzpatrick: "I got some invaluable advice from Paul and Hugh, including EQ'ing tips for acoustic guitar, vocal-treatment tricks, reverb settings, and some pointers on sorting out the hurdy-gurdy. This helped on the mix I was working on, making the acoustic guitar and hurdy-gurdy sit better together — quite a task in such an unusual studio. The introduction of Auralex foam was pretty amazing, and stopped the room sounding like the inside of a big shoe box! I am now busy with picture hooks and MDF, and on Paul White's advice I am now looking to purchase a subwoofer before I get back to mixing again. I have a couple of remixes to work on for other artists and I look forward to now being more confident in the final result. So, thanks to Paul and Hugh for braving the snowy mountains — I hope you enjoyed the day as much as I did!"
That left the hurdy-gurdy, which is essentially a bagpipe substitute for asthmatics! Strings are used instead of pipes, and these are forced to drone by turning a handle, but the audible effect is much the same! The dry recording sounded too abrasive and upfront, so rather than trying to beat this into submission using EQ, I thought I'd try using the Realverb plug-in to create a short reverb with a restrained high end that could be used to create a more realistic impression of distance. This particular reverb also has a distance parameter that aids the impression. Once processed, the hurdy-gurdy sat nicely under the guitar, lending a Celtic feel, but without making its presence too obvious. Paul planned to use automation to bring this up in the intro and between verses, but to drop it back when the vocal was present.
Although it wasn't our intention to finish the track, we got the bare bones of it sounding fairly sweet, and Paul saved our version so that he could come back to it and continue work later. He also saved our TC Electronic M*One and Realverb reverb patches for future use before we once more braved the snow-covered roads to return home.