STEVE LEVINE: Recording 'So Far Away'

Sound On Sound's May 2000 SAE/MPG competition offered readers the chance to win a recording session in a London studio with a top producer. Tom Flint talked to the lucky winners about the recording of their demo — and to producer Steve Levine, who took on the job of transforming it into a potential chart‑topper.

Encouraged by the excellent response to our first demo competition, which took place last year, SAE (School Of Audio Engineering) and the Music Producers Guild got together to sponsor a second contest in May. This time, the studio of choice for recording and production was SAE's flagship facility in London. The MPG were involved in the judging process and one of their members, producer Steve Levine — known for his work with The Honeyz and Culture Club — took on the difficult task of producing a release‑quality version of the winning song in just three days.

To get to this point, of course, we had to pick a winner — and as with the last competition, we were surprised by the sheer diversity of ideas, styles, productions and performances we received. After the first stage of judging, a few of the best demos were sent to the MPG where a team of experienced producers and engineers selected the track they deemed the most promising. The unanimous choice was a track called 'So Far Away', written and produced by Austin Maguire, with vocals and lyrics written and sung by collaborator Clare Canty.

Austin studied classical music at the University of Ulster, before moving to England to take a degree in Popular Music and Recording at University College Salford. Since graduating, Austin has been busy writing soundtracks for multimedia installations whilst taking on the occasional piece of TV work. Clare also studied Popular Music and Recording in Salford and now finds work as a singer, teacher and songwriter. "I do jazz gigs performing Ella Fitzgerald‑type stuff, and I've had a few pop hits with a band called Lucid," explains Clare. "I didn't know Austin at college but I knew his face, and we ended up being next‑door neighbours at one point. He was always busy so it took us years to start collaborating."

Starting Out

Well before Clare's involvement, 'So Far Away' was a compositional idea intended only for Austin's own album project. "It was one of the very first ones I wrote for the album about a year and a half ago," explains Austin. "It started out as an experimental rap track in 3/4 called 'Wick Wack', with poetic rap vocals laid over a commercially structured melody. I wanted it to be based around live‑sounding instruments such as piano, bass and kit, all set against a background of ethnically orientated percussion. The rap worked really well, but I moved on to different things and I didn't want to waste what was a perfectly good commercial song structure and chord sequence, so it was at that point I began working with Clare. When Clare came on board the structure stayed the same, so it didn't take a lot of reworking."

For the recording and production of 'So Far Away' Austin used his home setup, built around a Yamaha 02R mixer and an Apple Macintosh 7600/132 fitted with a Korg 1212 I/O card. Sequencing and hard disk recording was done using Steinberg's Cubase VST software, and effects and processing were provided by a Digitech Studio Quad and the built‑in effects and dynamics on the 02R. The sound sources were Austin's Akai S3200 sampler and his Roland JV1080 module fitted with Classical and Bass & Drums expansion cards.

The song evolved from a Roland JV1080 piano part recorded as MIDI into Cubase. This start point set out the verse, bridge and chorus structure of the track. Drums and percussion were soon added. Austin: "All the drums were synthetic but I tried to make them sound live, so if it ever needed to go live the song would cross over without any major hassle. They were a mixed bag of samples from individual hits to bits of loops. At that time I was cutting things up in the Akai S3200 a lot."

Effective Sounds

One of the most distinctive features of the demo track is its prominent use of sampled sound effects, including distant shouts, electronic sweeping tones, and several random crackling and beeping parts. Austin explains why he chose to use them. "I was trying to create an atmosphere which would complement my sound and give the music a degree of originality, so it was a case of creating a background bed using quite a few different sounds. Exactly where they came from I've no idea, they're just different bits and pieces I've collected over the years."

Keen to use approximations of live instruments, Austin added a guitar line using a sample off the Akai S3200. "I was doing soundtracks that were more classically orientated and that was the guitar sample I always used to use when I was writing," explains Austin. "The problem was that I always wanted to use session bass and acoustic guitar for this track, but processor speed and lack of RAM on the Mac prevented me from doing so. When I heard we were going down to the SAE studio, I knew the vocals were to be rerecorded so I was able to delete them from the track and use the freed processing power to replace the synthetic bass and guitar with session parts played by Dave Williams."

Also key to the feel of the demo was a string part used in the song's choruses, which was a sound from the JV1080's Classical expansion board. The track's bass sound also came from the JV1080, this time using a preset from the Bass & Drums expansion board.

Austin's next step was the addition and subsequent removal of the original rap vocal, although the backing recorded up to that point remained the same when Clare came to replace the vocals. Clare explains how she composed for the pre‑written backing. "Sometimes when you're singing you've got several different styles that you do but you can't marry them all together in just one style, so I was wondering what to do. I had been doing dance style with Lucid, but I didn't really like doing the big diva thing so I thought I'd try to do one of my jazz voices and chill it out. Even when I hear finished things on the radio I'm always hearing other harmonies and melodies, so I tend to work things out pretty quick. The hardest thing is getting words. I always write melodies first, then try to get a storyline in my head. For a lot of stuff I get inspiration from old folk tales. I grew up in Scotland, my mum's side of the family is Irish and my dad's side is Lithuanian, so I grew up with a rich culture of different folk tunes and poems."

Inspired by Clare's vocal, Austin decided to add the prominent flute part, most noticeable in the song's intro and choruses. "The folk‑story lyrics and voice it gave the song a Celtic feel, so I asked a friend of mine called Brian Finnigan to record a bit of his flute to pick up the Gaelic theme. That was just recorded through a Shure SM58, as a necessity rather than choice. For vocals we used Clare's Beyerdynamic jazz‑gigging mic, recorded via the 02R."

Preferring to add his effects and processing after recording, Austin still had a few adjustments to make to the sounds during the mix. "Given the processing power available on my computer, I have to steer clear from using any effects or EQ in VST, so I used the built‑in effects on the Yamaha 02R. On a lot of the tracks I used larger Hall‑type reverb; the compression on the vocals and the flute is off the 02R too.

"For vocal reverb I used a plate reverb from the Digitech Studio Quad. The vocodery effect was Auto‑Tune software, overdriven. I had several versions with that effect on, and we were having discussions about whether that would date or pigeonhole it, which we didn't want to do. We ended up keeping it because it wasn't used for long."

Once satisfied with the arrangement and mix, Austin mastered the track onto CD using IK Multimedia T‑Racks mastering software.

From Salford To Sae

Aware of the extremely tough task of producing a finished track in three days, SAE's Martin Berneburg took charge of the session to ensure that the production could be completed in time. Also on hand was SAE lecturer Xander Snell, who helped set up the project and aided in the use of the studio.

The Neve studio at SAE was chosen for the session. Steve Levine likes to bring his own rack of favourite equipment — including an Apogee AD8000 A‑D converter, Millennia Media Origin voice channel and Tascam MX2424 hard disk recorder — so this had to be temporarily incorporated into the setup too. In addition, a complete Cubase VST PC system with an Echo Mona I/O card (kindly supplied by Steinberg) was set up in parallel with one of SAE's Mac‑based Pro Tools systems running Emagic's Logic Audio as a front end. FX Rentals supplied an Akai S3200 with Jaz and Zip drives for Austin's samples to run from. Audio material was tracked to the Tascam, while the Pro Tools system was used for comping and editing vocals.

Austin explains how things were set up: "The first day was spent recreating the track in the studio with Martin and the SAE crew, and didn't involve Steve or his engineers. All the files were transferred from Cubase to Logic Audio; we exported and converted all the audio into a Logic‑compatible format. I brought several Zip disks containing all of the samples in Akai S3200 standard format and the expansion chips from my JV1080 which have all the relevant patches. I took the live bass down with me in the form of a Cubase VST AIFF file. I also emailed the MIDI file to Steve so he could get a general overview of the track layout and arrangement. Originally, I'd had to compromise on my outputs and have several different instruments coming through the same outputs — so here, to increase the number of audio outs and give each one its own channel for the mix, all of the samples were split between an Akai S3200 and two of SAE's Emu E5000 samplers."

Weekend Workshop

With everything working smoothly after the initial setup day, Steve Levine and his programmer Darius Zickus and mix engineer Mark McGuire arrived the following day to begin work. From the demo, Steve had identified a few key areas in which he thought the song's production could be improved, and so the session began with an initial discussion on the creative direction in which the track should go. Steve explains his first reaction to the track: "I thought it was a bit odd at first, because the sound effects were louder than I thought was appropriate — I couldn't properly hear any of the rhythm. I found myself being incredibly distracted by these effects and I thought there was a song screaming to come out of it. It sounded to me as though Clare had a potentially fantastic voice, but I thought it was all a bit blurred.

"The very first thing I wanted to do was look at the rhythm side. Having spoken to Austin I realised he used Cubase and he said he'd been wanting to compare it with Logic for a while, which is my preferred sequencer. So I asked him to email me a MIDI file. I had a look and could see it would be reasonably straightforward to transfer it over to Logic. He had some audio tracks but most was still MIDI at that point, so we started with Austin's MIDI. Darius, my programmer, and I made a conscious effort to work on the drums. As soon as we did, there was an amazing turnaround — it really did lift the song.

"The song is in 3/4 which creates both a negative and a positive: it's positive because it's unusual and gets your attention, but negative in that it is much harder to pin down the rhythm, because people still try to count it in 4/4. So it was very important to anchor the kick. There are two bass drums, one doing the main beats, and the other doing little offset beats. There are also two hi‑hats. The first is a straight threes pattern, which glues the basic block together and plays all the way through. A second hi‑hat joins it in the chorus. We also put a shuffly jazz‑feel ride cymbal sound on the other side of the stereo field, so when the chorus kicks in that gives a really nice lift.

"Because the song has a jazz feel, I wanted it to have a much better and stronger brush sound. The brushes give it an additional rhythm because not only are they playing the anchor beats, they're also providing a separate track with some sweeps and whooshes. Again that's one of my own samples. A lot of articulations are really important with that type of sample, so you can find the one that feels right. When I recorded them originally I sampled them in stereo, with a top mic to the left channel and a lower mic to the right, so you play them back via two mono channels in the desk and can adjust how much snare buzz you have versus the hit sound. It's incredibly effective.

"There were a couple of congas and daraboka percussion loops I left in, but I tidied them up as sample edits. They were really noisy so we filtered them, but the result was really good because they filled a better space tonally. They were panned three‑quarters on each side to create a little bit more of a stereo space."

Frills & Thrills

When it came to the other parts, Austin had already replaced some of the original MIDI parts with liverecordings. "Since the demo was made Austin had put the real bass on which was quite a lot better, and just a couple of little nips and tucks on that made it really good," explains Steve. "For example, there were a few clicks due to the inadequacies of his recording situation and he said those would have been impossible to tidy up in Cubase as easily as they were in Logic. I also used BIAS Peak for really fine, detailed editing.

"Austin also recorded an acoustic guitar, which we changed a fair bit because it had been under‑recorded. It was quite funky and dynamic but parts of the dead notes were incredibly loud in comparison to the chords. By increasing the gain on certain parts and pulling other bits down before compressing, we managed to improve the overall performance of that quite a lot and made it sound more musical."

The major bone of contention during the session was Austin's unusual sound effects. Steve explains how he adapted them for the new recording. "There are three or four tracks of sound effects. I think I took a couple out and EQ'd them slightly differently so they fitted in the track a bit more and obviously nowhere near as loud. They still give it the same quality as the original, but they're not the focus.

"We also improved the piano part dramatically. Firstly we took the programmed piano away and then Darius played a new piano which you can hear on the final version. That's one of my samples from the Emu E4XT Ultrasampler. It's quite a large 64Mb sample, which sounds very natural.

"The synth part was a JV1080 basic patch called 'Pulse Key', with the effects edited very slightly. I changed that because I didn't think the bridge lifted enough, and the piano he originally played didn't cut through because it was rhythmically very similar to the guitar, so you had two things fighting for the same space. Darius also played an organ patch in the chorus called 'Augmented'. That was one of those things where you flip through the sounds, loop a little bit, and think 'That sounds pretty good.'

"The string pad was taken from Austin's original part, but Darius changed some of the voicing. It's still in exactly the same place in the song but has some additional octaves, and has been split into two sounds. They're almost panned to extremes, with the lower range on the left and higher on right."

Vocal Workouts

Perhaps the most significant change to the song was the rerecording of the lead vocal. Clare describes her experience of the process: "Steve gave me a free reign. I quite like breath growls, sibilance and dirt on a track and I think a lot of the time pop, especially R&B, is too clean. They use so many tracks just for the main vocal and they're very smooth — I'm a bit bored of that. A lot of other people I've worked with cut out the breaths and they don't want any of my Scottish accent. Considering it was three days, I was absolutely amazed what we got."

Steve was also pleased with the resulting vocal recording. "Even if I do say so myself, I think the results are pretty spectacular considering the hours we had. A lot of that is due to the fact Clare is a good singer, so the vocals didn't take more than two hours maximum. We did nine takes in total, plus two tracks of experimental vocals for the samples. It was probably an hour and half for comping, which is reasonably fast — then we were up and running with the next thing."

To record the vocal Steve used a newly acquired BLUE Mouse valve mic, coupled with his much‑loved Millennia Media Origin voice channel. "I just wanted to try something new with the mic," explains Steve. "I think the design of the mic head is quite inviting to singers, so they get quite close to it without being intimidated, and I like a close vocal sound. The mic goes directly into a Millennia Origin voice channel which is absolutely fantastic. They're not cheap but you get a valve signal path, a solid‑state signal path, and a phenomenal DI box for all the synths when recording to hard disk. The vocal sound from about 5 or 6kHz upwards is very good and very smooth. The compression on the Millennia is quite subtle, so the more extreme compression is done with the Waves Renaissance compressor plug‑in. I also used the Waves De‑esser plug‑in, which allowed me to get a very bright vocal sound without excessive sibilance.

"I used the Millennium in the full tube range with the transformer in because I'm a big fan of transformers — they give you a larger‑than‑life sound, and it protects you from extraneous noise, hum and RF. I used a fair amount of top boost at around 16k and a little bit of mid on a really wide Q. Then I had a little bit of low‑frequency cut to get rid of any room crap."

Despite the extensive sonic tweaking that went on, the end of the rerecorded 'So Far Away' was the only part of the composition to receive a structural change. "It was one of those accidents that happen when you copy a file and dragged it a little bit to far!" admits Steve. "We thought it sounded quite nice as an ending so we worked it around, added the little percussion sound, reverb, and dub echo on the end of the vocals."

At The Mix

Time being limited, Steve ensured that mixing was ongoing throughout recording, with help from Darius programming and Mark working SAE's Neve desk. Steve: "As soon as I'd finished the vocal comp we were working on the sound, so by the time the last thing went to the Tascam multitrack we had a fairly close picture. The actual mix took no more than an hour and a half. Although there is Flying Faders automation on the SAE Neve, the mix wasn't automated at all. Having the lead vocals and backing coming out of the computer I was quickly able to balance the harmony using Hyperdraw in Logic, so Mark literally had the fader coming in at unity gain. The 'oooo' pads were coming live from the Emu, with Hyperdraw controlling the MIDI volume. Mark only had to add a little more reverb and extra compression on the vocal.

"All the other keyboards and drum parts were printed with a reasonable amount of EQ. For the acoustic guitar we did use more EQ and more compression on the desk, because the more compression we put on the better it sounded. We thinned it on the board as well.

"We mixed it down onto Yamaha D24. As a multitrack it suffers badly against the competition, and it's certainly not as good as the Tascam, but I've been using it as a mastering machine. It's very flexible, because you have the choice of whatever I/O card you want to install, including using the analogue card with four tracks at 24‑bit/96kHz if you want. In this case, though, we could come out of the back of my AD8000 via the RCA digital connection into the D24's RCA digital connection at 24‑bit/44.1kHz. Also, having eight tracks, you can also have an a capella mix on tracks three and four and the backing track on five and six, so if there's a situation where you think the lead vocal is not loud enough you can balance between the different versions and tweak the levels a little."

Completing a project within such a tight timeframe can be a recipe for disaster, so the question had to be asked: what did Steve and Austin think of the result?

Steve: "I think it's really good now, and the vocal performance is great. The lyrics are really good and that passed me by on the demo. I don't know how good it would be if we had another week on it. And, of course, a couple of other dance remixes would open a few doors."

Austin: "I picked up a lot from Steve — the way he set up the relationship between the vocals and the rest of the mix as a whole. He recorded them in one fell swoop, whereas I recorded them in sections, which then presented problems with levels. The changes to the rhythm track gave Clare a better timing base to sing around, too.

"Steve's use of panning was also interesting; that's something I don't really experiment with enough. And he's put a nice commercial edge on the song, whilst retaining those jazz and folk influences. Given Steve's track record, I trusted him from the start and I'm well happy with the outcome."