You are here

Aquabats: Chaotic Recording

Interview | Band
Published December 1995

Aquabats are an unconventional trio dedicated to the creation and recording of unconventional improvised music — using unconventional and often improvised instruments — which they call Earth Music. Tim Turan tells the story.

No keyboards, no computers, no sequencers, no drum machines, no samplers, no guitars, no basses, nor any other popular music‑making device was used to make the music on the Aquabats Magiko CD. Natural ambiences, improvised acoustic instruments, found sounds, voices and a simple philosophy (nothing is written down) is what the band is all about. What started as a one‑off evening in January 1993, messing around with thumb pianos, roto‑toms and flute, ended as a one‑off evening in December 1993 launching our CD at a large party in Oxford. 'Earth Music' was born.

Tim Turan (35 — drums, voice, sound engineer), Frei Zinger (43 — flute, voice) and Sue Smith (45 — voice) make up the trio (everyone plays percussion). Each member has had over 20 years' experience of playing 'rehearsed band' music, including reggae, jazz, rock and pop. But even when bands improvise, they are still confined to the same format, timbres, loops and — most important — mentalities. Aquabats provides the perfect vehicle for realising our musical dreams, and the omission of normal 'band' instruments removes any pressure to compose and arrange in the traditional way.

When I invited Frei Zinger and Sue Smith up to my attic studio on January 11th 1993, the only intention was to make a noise and record it, and the only plan was to be there at 8pm. Sue arrived carrying a suitcase, inside which was what seemed like the entire known universe of hand percussion, plus a few other objects as yet unidentified by science. Frei Zinger held a flute in one hand and a bag in the other. Inside the bag were about 15 guitar effect pedals. I contributed roto‑toms and a collection of bells and kitchenware to the pool, and my children woke up next morning to find most of their noise‑making toys missing...


Though the instruments were unconventional, the setup for recording them was straightforward: two Audio Technica AT4033, two Calrec 1050C, two AKG C1000 and two Tandy PZM microphones were used. As they're all either back‑electret or condenser mics, they're well‑suited to accoustic (and quiet) instruments, and using them in pairs gave us four different stereo options. The mics were patched to an Aries 16:8:16 mixer, and then to a very old Tascam 38 multitrack with no alignment, no noise reduction, and no brakes, running with varispeed down 12% at 15ips (for around 40 minutes of recording time), and recording to Ampex 456 tape. A Symetrix SX202 mic amp took care of the pure flute, bypassing the desk altogether. Frei Zinger also used a bug mic on the flute. I printed the signals to tape extremely hot, so that the gain on the tape returns to the desk could be reduced — and consequently the tape hiss. An Alesis 3630 compressor prevented the hot signals from melting down entirely, and the ever‑wonderful Drawmer DS201 gate provided a useful degree of separation during duet or trio improvisations. All this was in one room, with up to eight live mics, and loudspeaker monitoring, dicing with feedback.

All the performances on the album are improvisations, and each track was started by who ever had the first idea. The first session produced the track 'Magiko': Sue and I played a duet on thumb pianos on tracks 1 and 2 of the Tascam, and another rather sick‑sounding thumb piano was added on track 3. Frei Zinger added a flute motif, miked with one of the Calrecs, on track 4. He then improvised to his improvisation, using the bug mic into an MTR Gain Brain, which brought the mic up to line level, then to the effects pedals and straight to track 5, effects and all. A huge 18‑inch roto‑tom with a hydraulic head (two drum skins sandwiching a layer of oil) and tuned to long‑distance elephant pitch (deeeep) was recorded to track 6. This drum was miked with an AKG C1000 two inches from the edge and about 2‑5mm from the surface of the skin. The proximity effect lifted the bass end significantly, and a little EQ boost at 40Hz and cut at 160Hz made it rumble without overload. I had only to lightly flick the drum with one finger to produce an enormous sound. Two separate trio improvisations of various small percussion and voice sounds were recorded to tracks 7 and 8 respectively, using the Calrec mics.

Once all eight tracks were used it seemed logical to call a halt. No track bouncing had been planned, and there was more than enough material to create a mix. Enter the Alesis Midiverb II and the Yamaha SPX90 II and SPX1000. Only reverbs were used on this track, preset 5 from the Alesis on the thumb pianos, and the SPX90 II handling the percussion and natural flute, with a plate reverb set to three seconds delay (high cut at 8kHz, plus 5dB cut at 120dB to reduce the effects of wind rumble from an enthusiastic flute performance — works better than a Rennie). The SPX1000 was the 'feature' reverb — a 30‑second Hall reverb setting, no filters or EQ, maximum density and diffusion, and no pre‑delay. The voices and big drum were fed different amounts, and the stereo return was panned equally left and right. The telephone at the end of the track was a spontaneous event, captured with the vocals (Hello Mick Moorhouse!); the high‑frequency component the phone adds to the reverb tail gave it a beautiful suspended quality.

During the mix it became obvious that the things we perceived as mistakes on individual performances were in fact crucial elements which had a profound effect when heard in the context of a mix — and anyway, earth music should be as random and naturally spontaneous as the earth itself. The whole lot was mixed to a Sony DTC1000ES and by midnight we had an 'audio thing', three minutes and six seconds of sonic stupidity, conceived, performed, recorded and mixed in four hours. This process was repeated once a week for 14 weeks to derive 14 tracks.

Audio Photography

All the original sounds on the album were captured using microphones (ie. no line inputs except the effects, which weren't original sounds). We were recording, storing and manipulating a real‑time image. Photography is a good analogy — like a lens, the mic frames and focuses the image, wide‑angle and telephoto lenses could be compared to cardoid and hypercardoid mics, the tape recorder is your camera body, and the tape is the film. A dictaphone would be an instamatic! The mixing process could be compared to developing the image.

A Sony Pro Walkman cassette recorder with an ECM 929LT stereo mic was used to record all our natural ambiences. For an outlay of around £350, this setup gives you instant high‑quality audio 'snapshots' on a tried and trusted medium that doesn't seem to mind sea water, rain, heat, cold, lightning, sand, children or, more recently, airport X‑ray baggage inspectors who couldn't handle what they were seeing on their screens as my holdall went through the machine (to them it was obviously a huge bomb, cunningly disguised as — a huge bomb).

The fourth recording session produced 'Thundermental'. I began recording thunderstorms 20 years ago, and on the night of August 30th 1992 I set up the Sony as usual to record an approaching storm. Thirty minutes later, a bolt of lightning at a distance of 30 metres took out my neighbour's garage. With the torrential rain as well, it was awesome — it had to go on the CD, and where better than at the beginning: The Big Bang.

So natural ambiences became part of the equation. The river Isis in Oxford city centre features on the track 'Flowtal'. The fabulous swirly phasing effect is a result of the 929 mic hanging by its own lead 1m above the water. Twirling around, the stereo mic received reflections from the stone walls of the bridge, including the direct sound from the river, and the mic movement caused phase cancellations and wild panning all over the place — no effects were used at all. You can also hear 'mud bubbles' — try this recipe: submerge a china teapot in a bowl of water, put your finger over the spout, and a steady stream of bubbles comes out of the steam hole. Adjust the timbre by altering the water depth, then drop the pitch by two octaves using a pitch shifter, or in our case the Revox PR99 MkIII.

The fabulous swirly phasing effect is a result of the 929 mic hanging above the water.

The natural sounds are transferred from cassette to the Revox quarter‑inch for editing, before recording to the Tascam 38. Lorries and a train were cut from a recording of birdsong, and most of the low end was removed with bass cut at 800Hz to remove the general city hum. The milkman was deleted from the thunderstorm with a razor blade, and a slight EQ cut at 3kHz cured the 'frying onions' effect produced by incessant rain on concrete. The low end was boosted by 4dB at 30 and 40Hz to try and achieve that cinematic rumble from the thunder. A 700Hz dip softened the sound of rocks being dragged by an aggressive November Atlantic on 'Aquabats', and a VC10 on its way to Brize Norton was cut from the river Isis in 'Flowtal'. The ambiences were mixed quite loud, because they were never intended to be an audio backdrop, but rather an integral part of the composition.

The Mixing Process

Mixing was made more complex by the nature of the instrumentation. This was not rock 'n' roll by any stretch of the imagination, and normal mixing practices simply did not apply. The range of frequencies in popular music is roughly 20Hz‑20kHz, and a competent engineer will know where to put everything — after all, the instruments are nearly always the same. The frequency range on a track like 'Bonsai Greeting' or 'Wokonwater' is mostly 1‑5kHz — ouch! Taming the harshness was a matter of placement rather than EQ or effects. Where sounds were placed in the stereo image was the most important factor, because once there was panoramic balance, the effects took care of themselves.

One effect that did not take care of itself was the DIY 3D process which features at the end of 'Zen Vox'. I've not yet heard any 3D system that convinced me that anything "went round my head", or anywhere else, for that matter. But I do love the effect of phase manipulation. The SPX1000 Flange B program allows you to adjust the phase of the modulation delay in 22.5 degree steps. One channel of the flange, 157.5 degrees out of phase, was sent to an Alesis MEQ230 graphic with the sliders set alternately to maximum cut and maximum gain (comb filtering). The other channel went to a DI box to utilise its phase button, and both channels then proceeded to the mixing desk. The sound being treated was a flute, recorded in mono with tons of distortion and feedback flange. On the desk at mixdown were five faders for the 3D flute, the mono track plus effects, the return from the graphic (panned left), the DI box (panned right), and the SPX90II set to a pre‑programmed slow equal left and right autopan. A completely hit and miss, trial and error session of pushing phase buttons and adjusting the flanger's mod delay in 1mS increments produced some startling results. We found that the EQ determined how 3D (or NO D) the image became; one slider notch either way and the sound would collapse or expand out of all proportion. It was all experimental and I only had an hour to mix the whole track. What you hear on the CD is the abrupt end of the main theme, leaving the flute swirling around in its own effects. The 3D L&R channels are very slowly faded up, while the flute fades out, and the aux sends on the flute are pre‑fade so that the signal still reaches the effects even though the fader is down. The send to the auto‑panner is now slowly brought up and the image starts to swing. Finally, the speed of the pan is gradually increased to 'insanity' before cutting off, leaving a wake of reverb. There are many documented ways to alter perspective using phase, and our version just happened to work. Have a go yourself and see what you come up with by experimenting — or go and buy a Roland SDE330!


By the end of the 14th session we had approximately 74 minutes of mixed material on DAT. Let's make a CD!

Mastering was done in two stages. First, the mixes had to be arranged in the right order, with three crossfades to execute. Second, the resulting master DAT would get further EQ and hard disk computer treatment, before final transfer to PCM 1630 for PQ coding and glass mastering. The first stage utilised three DAT machines, necessary for creating the crossfades. A Sony DTC77ES and Aiwa HDX1 played the tracks to be faded into the desk, and the result was captured on a Sony DTC1000ES DAT at 44.1kHz via the analogue inputs. We could only afford one day in the digital mastering suite at SRT, so the crossfades were done in our attic studio, spending hours getting them right.

Stage two meant a journey to St Ives, Cambridgeshire [famous as the home town of a certain leading European hi‑tech magazine! — Ed] to use the mastering room at SRT. We spent a day with Nick Watson on a SADiE hard disk system, equalising each track with a very powerful Sony DSP1000 digital EQ. The individual nature of the timbres of each track made accurate EQ essential for an even‑sounding CD. I took a pair of Sennheiser reference headphones to check the final version before the final transfer. To me, excellent headphones are the equivalent of DI'ing into your brain. They are the microscopes that let you see the sound close up, before any room acoustics inflict themselves upon it.

One of the wonderful things about computer editing is the level of precision available. A tiny but audible click from a footpedal was removed in 'Thundermental'. A cartoon pencil under mouse control re‑draws the offending waveform at single‑sample resolution — try doing that on a Revox. Two stick clicks on 'Elberoto' were removed using fast, equal power crossfades. The analogue transfer grunge between the tracks was replaced with digital silence, and the starts and fades were 'tucked in' seamlessly. I edit 'books on tape', performing around 7500 quarter‑inch edits a month, and seeing this sort of control in action for the first time was mind blowing.

After mastering to 1630 U‑matic, the final job was to place the markers at the transitions of the crossfaded tracks for CD track numbering and PQ coding, which gives you the final track times and total length for the finished product. We took a CD‑R home to the attic studio, sat back, listened, and knew we couldn't have done it any better — it sounded superb. Now we had to decide what it should look like and what we should say in the CD booklet. My sister happens to be an MA(hons) graduate from the Royal Academy, so she was sent a tape and given these simple instructions. "Use canvas and oils, then turn on the tape and paint what the music tells you. When the tape stops, turn it over and continue or stop. Do not paint in silence."

Paul Medley photographed the resulting painting, Steve Martin put all our text and artwork together on the Macintosh, and we sent the whole bundle off to SRT to finish the job. Eventually we received 1000 CDs and 500 cassettes. The entire project had cost us £3400, most of it on the artwork and post‑production, and since its release we've sold over 2500 copies, without any distribution or mail order service. Every attempt to find a distributor has been greeted with the same old negative "sorry, can't find a pigeon‑hole for you" approach. It seems there isn't a space for us in the great universe of music, so we exist in a parallel one!

Taking The Wookey: Going Underground With Aquabats

In January 1994 we realised one of our dreams — to record inside Wookey Hole, Somerset, a system of huge underground caves five minutes from Cheddar Gorge. Earth music, performed inside the earth — irresistible! We had the whole place to ourselves for four hours, and a guide with a nuclear torch kept us from getting lost. Despite the fact that the temperature inside the caves remains constant (9C) throughout the year, so that it actually felt warmer inside than outside, it's not a friendly environment for video cameras, digital recording equipment, or PA systems. Limestone caves have superb powers of percolation, and since it was raining very heavily outside, this resulted in very heavy rain inside. It was also extremely muddy. Still, the Aquabats' spirit of adventure and the fact that it just sounds so damn good in there made it a very memorable session. And nothing blew up!

The session was recorded with the two AT4033 mics, Symetrix and Sony DTC77 DAT setup. The reverb inside the cave is haunting and dark, and lasts forever, and the effect on the flute, voice and drums is spectacular. We played two improvised mini symphonies and spent the rest of the time making single‑shot samples of various instruments to DAT. We also had a scary time submerging two AKG C1000 mics into one of the drip pools. All you need is a couple of condoms! They are very strong and can accommodate a C1000 quite comfortably. Use tank tape (plumber's gaffa) to seal the end around the cable, and in you go! We spent the last five minutes hitting a 13‑inch gong and plunging it into the pool, thereby choking the sound in a dramatic way and entertaining the crew at the same time. I should note here that this is a highly dangerous exercise and should not be attempted by anyone who might blame me in the event of a catastrophe.

Sonic Tricks

An Aquabats' speciality is the vocal drone, and these appear throughout the album, in various forms. They were created using a DIY Infinite Reverb which utilises the SPX1000 patch called 'Echo Room'. This patch invites you to build your own space, giving you control of height, width and depth up to 100m, which can then be multiplied by up to 10 times, the maximum being 1000m cubed. There's nothing to stop you having a room 1000m x 5m x 10cm — except, possibly, good taste! Set the reverb time to 480 seconds and multiply that by 10 — 80 minutes! This is obviously not infinite, but when sent in small amounts via an aux send to a digital delay with a long delay time and a hold button, it becomes as infinite as your electricity supply allows. The secret to smooth drones (holding a low voice note), is to start droning (say "AHH") into the mic and then slowly raise the level on the aux send to the reverb. This results in a sound with no percussive edges, which get caught up in the delay and create an obvious loop. Turn down the aux send as soon as you've finished droning, then repeat the process with a different note. Add as many notes as you like to build a huge chord. Then raise the reverb faders on the desk to hear the result. The drones that appear on the CD were recorded to one channel of the Tascam 38 and then treated with reverb in the mix to give it some stereo spread.

I mentioned earlier that the Tascam 38 was running at low speed (‑12% at 15ips). We also recorded many sounds with varispeed at +12%, to slow them down during normal‑speed playback. Although this practice is nothing new, some instruments can be totally transformed. On the track 'Chakatu', a conga drum turns into a giant earthenware pot, and the flute plus a little delay becomes a distant ship's foghorn. Since recording the album, I have acquired two Digitech RDS delay units (four and eight seconds delay). All the delay parameters are adjusted with knobs, which means that you can change more than one at a time — unlike most digital delays with parameter buttons and scrolling menus. Because we improvise our material on stage in the same way we do in the studio, the 'knob' delays have become an important performance tool, used to widen, chorus, flange, double, loop, phase and occasionally even delay any sound we choose to throw at them, on the fly! Frei and Sue use the footpedal equivalent, the Digitech PDS8000 — six of them!


Magiko was reviewed by SOS Demo Doctor John Harris in the March 1995 issue. John liked the album enough to award it Top Tape, complimenting the band on "interesting playing, mixing and use of effects" and a "fine job" by Tim Turan. You can get hold of a copy of Magiko by writing to Aquabats, 6 Peel Place, Oxford, OX1 4UT.

A Soft Approach To The Future

Our trip to Wookey Hole has inspired us to perform an Earth Symphony, consisting of four movements. (corresponding to the four elements — Fire, Water, Air and Earth.) We will improvise different atmospheres and digitally splice and crossfade the sections together. Since we started gigging, we've built up an impressive array of new toys, instruments and sounds. I can't say anything about the material because we won't know ourselves until it's recorded. All I can tell you is that there will be some unfathomable sounds and textures, some of them created using a version of a new 'morphing' idea I'm currently working on. The second half of the new album will again be a seamless journey, this time through the best moments from the live tape archive. Every show, all our busking, and special appearances have been recorded, and consequently there's over 60 hrs of material on tape and CD‑R. It's going to be a tough decision to select just 35 minutes from that lot.

This time the project is being put together on a newly‑acquired Apple Macintosh with Digidesign Session 8 (8‑track) and Sound Designer software. I'd already decided to buy a computer, and having used the Alesis and Tascam digital 8‑track machines on previous projects, considered them both. However, it worked out that the cost of the Digidesign system for the Mac was the same as buying either of these machines. The ability to execute long crossfades, draw waveforms, and undertake absolute pinpoint editing allows Aquabats to realise the more unfeasible ideas that usually get binned, due to the analogue 'technical threshold' being reached. It also sounds good, which in my opinion is the single most important factor, and the reason why my Fostex E2 quarter‑inch tape recorder is still used for production mastering.

If I have any niggles with the software, one is not having a scrub tool to 'rock 'n' roll' over an intended edit point in the Session 8 program. This is infuriating on a recording and editing system of this sophistication, and a basic editing requirement. In Sound Designer II the metering is inadequate, and the 5‑band graphic EQ is actually a 5‑band parametric — it just looks like a graphic. This results in a huge difference in control power (upwards fortunately). I'm sure, also, that more than one level of 'undo' could be written into the program, and it would be great to be able to change the gain of individual channels. The ability to import files would also help — maybe these facilities could be addressed in a future software upgrade? Still, aside from these few gripes, which were discovered during the first session, I think it's an excellent audio manipulation tool, and I couldn't bear to be without it.