Bill Bruford • David Cross • Pino Palladino • John Parracelli • Louis Borenius • Paul Downing • Andy Holdsworth • Simon Pearson • Chris "Snake" Davies • Frank Felix • Roland Kerridge • Richard Niles

Simon Zagorski-Thomas had a dream; to get together an impressive array of the UK's foremost musicians, get them playing alternative MIDI controllers and make a CD of the resulting totally MIDI session — without a keyboard in sight. Here he tells the first part of the epic story...

What's the worst job that you ever had? Mine started a while back in those carefree days when I only worked 12-hour days and my biggest technical challenges were time-stretching loops from different places in an old jazz track and trying to get them in strict time and in tune with each other; or trying to program a synth sound for someone who could "hear it in their head — you know it's sort of fuzzy but dull and goes weeeow". Such was the idyllic simplicity of my life. Then a tragic thing happened — I had an idea. What about making an album where everything on it was played on MIDI controllers by real musicians and nothing was quantised? A compilation album where the only unifying factor was MIDI. The idea for the article grew into both a description of the sessions and the technical side of using MIDI in this way; talking to the musicians about how they got on with different controllers and how the musical side of MIDI was developing in terms of expression and reliability.

The first thing was to try and persuade 50 or so musicians to give up some of their valuable time to record the project on the understanding that the royalties from this bizarre mail order CD probably wouldn't amount to more than the proverbial hill of beans. Funnily enough, people were not only interested, but excited... and everyone also seemed to know someone else who played a MIDI controller. Our interest in MIDI at the Premises Studio means we have a Yamaha WX7 wind controller, a Wal MB4 MIDI Bass and a Simmons SDX Sampling Drum setup, as well as sounds I have programmed to respond to non-keyboard controllers. I knew of lots of players and a few of the more esoteric controllers like MIDI violin, Chapman Stick and vibes. I also found out about MIDI trumpets, the pedal steel guitar system, and heard tell of a fabled MIDI flute and even a MIDI accordion. Everywhere I turned, I found someone with a little bit more information and a huge store of enthusiasm.

The next problem was studio time. Luckily I'm a partner in a 24-track recording facility in Hackney, East London, called the Premises Studio. Unluckily it's very popular and so blagging time for this project was a bit of a problem. There are also eight rehearsal rooms there, so there was the possibility of sequencing one of the less complicated sessions in one of those. I've also been lucky enough to get some equipment on loan from various companies, which has made the first few sessions much easier and which will be of even greater use once all the editing and mixing gets under way.

Now it was simply a matter of getting everybody's Psion Organiser in the same place at the same time, getting all this complicated and temperamental equipment to interact properly with these complicated and temperamental people, getting everyone to play brilliantly together using an instrumental line-up they were completely unfamiliar with and getting everyone's performance into a sequencer, whilst at the same time providing them with an inspiring monitor mix... is that all?

Monday 15th February 1993

BILL BRUFORD: Simmons SDX Drums
DAVID CROSS: Zeta MIDI violin & Korg M1R
PINO PALLADINO: Wal MB4 MIDI Bass & Roland D110
JOHN PARRACELU: Roland GR1 MIDI Guitar

The first session: what could possibly go wrong? Well... our SDX could break down the evening before and John Parracelli could have to leave half way through the session.

I had the Unitor for my Notator and a Hybrid Arts SMPTE Track to do the sequencing on, so I had four MIDI ins into the two Atari computers; three on the Unitor and one on SMPTE Track. I had the MB4 bass sending on channels 1-4, the SDX on channels 5-11 and the violin on 12-16 (it's a 5-string version of the Zeta Jazz Violin). John Parracelli was on channels 1-6 of SMPTE Track.

I'd had some 5-pin DIN to XLR cables made up so that I could use mic leads as extensions for the MIDI lines and ran them through the door from the live room to the control room. If there'd been space in the control room for a four-piece band I'd have had them set up in there and doing the session through the monitors, but as it was I had to run the audio lines and the MIDI lines through from the live room and send a headphones mix back to them. This works better for the drums in any case, as the sound of sticks on pads is never drowned out by speakers and so the drummer gets a slightly distorted idea of the sound he or she is making. I also had to put a talkback mic in the centre of the live room, as everyone was being DI'd, and so I couldn't hear their plaintive requests for more level in the cans otherwise.

So we started off with Bill playing a kit he uses live — mainly acoustic percussion samples with an electronic hi-hat. David played a flute sound off the violin, John played a throaty sawtooth-based sound off the GR1 and Pino played a synthy finger bass sound with a slight filter sweep on the MB4. The session consisted of a long chat and a series of jams, but unfortunately John Parracelli had to leave after the first tune, a long medium-tempo number in B flat minor. Then came an ambient tune, in which Pino played chords on a string-y pad sound while David played an arabic-flavoured solo on a brassy analogue synth sound. The third tune was an up-tempo number in D; Bill played a conga pattern as the main driving rhythm, Pino used his sound from the first tune and David used the brassy sound again. Bill then changed his kit to an 808 snare and hat and a set of crossfading toms and cymbals that I'd sampled from the 808 and DX7. This inspired a short quirky piece based on 'Three Blind Mice' and then a very spacious medium-tempo tune in E, for which David played an ethereal choral sound and Pino used a synthy slap bass.

This turned out to be one of the few sessions where everyone involved was familiar with their MIDI controller before they started. Bill uses the SDX extensively and indeed helped Dave Simmons with the development of the instrument. Normally he uses the SDX mixed with an acoustic kit, writing tunes around specific samples and using the real drums to improvise on. He found that playing in a totally MIDI ensemble actually felt more natural than mixing MIDI drums with a live band. Bill finds the sounds he uses determine the way he plays and when he swapped from his own sounds to the 808 kit that I had programmed he changed the feel of his playing completely.

Pino has a Wal MIDI Bass (though he used ours for the session), but he hasn't used it much for session work — adding Moog lines to a few tracks but not a lot else. The main trouble with this system is not the timing but the sensitivity. In order to get the MB4 to trigger reliably enough to use in a live situation you have to remove most of the dynamic sensitivity. It also requires a very precise and clean playing style, which seems to suit Pino quite well.

David usually uses his MIDI violin in conjunction with the electric violin sound and takes off the sensitivity from the instrument to get more reliable triggering. We spent some time together trying to get it to be more responsive, and by making MIDI volume the sole controller of dynamics we got a much better feel from it. We did this by turning up the sensitivity on the violin to maximum (or just under), setting its output to be only note numbers, MIDI volume and pitch bend (with fixed velocity), and editing the synth's envelope to have instant attack and almost instant release. Without any amplitude envelope shaping in the synthesizer, the only factor that determines the volume of the sound is the MIDI volume — determined by the bowing strength on the violin.

John finds the Roland GR1 system the best guitar system he's found so far, though he hasn't tried them all, especially the more expensive ones.

All in all, the general consensus seemed to be that there were a few brief moments when they forgot they were "playing plastic" but, as John put it, there's still a lot of daylight between him and the sound. Bill was very happy with the SDX and thinks that it's by far the best drum system, but we had an interesting discussion about the distance between the player and the sound that there seems to be with MIDI systems at present. David felt that somehow the sounds that came out of the M1R when he played didn't 'belong' to him in the same way that the violin sound did. It also seemed that it was very important to maintain some kind of 'natural' relationship between what you do on the instrument and what comes out of it; it may be possible to get an electronic drum to play quieter the harder you hit it, but it doesn't make for a good performance.

The other important point that came out of the session was that if anything was going to be released it could not be a jam session; there had to be a tune to work on. Everyone agreed that if we could all fit our schedules together again we'd try and organise another session so that we could get something on the CD.

Sunday 21st February 1993

LOUIS BORENIUS: Simmons Silicon Mallet & Akai S900
PAUL DOWNING: Wal MB4 MIDI Bass & Roland D110
ANDY HOLDSWORTH: Roland GR1 MIDI Guitar
SIMON PEARSON: Simmons SDX Drums

This session was sequenced using exactly the same gear as the first one, with the bass going into Unitor/Notator on channels 1-4, the SDX going on channels 5-7, 9-10, the GR1 on Channels 11-16 and Louis' mallet going into SMPTE Track on channel 8. I then saved the SMPTE Track song as a MIDI file and loaded it up into Notator to get everything on one computer at the end of the session. Louis and Andy were both used to their instruments but Paul and Simon were coming to them for the first time, and so had to spend a while getting used to how they had to alter their playing styles to their respective instruments. We were working on a tune of Louis' called 'The Last of the Aztecs'.

We started off by using a kit that I had put together for the SDX, which had multi-sampled saucepans full of water as tom toms, a metal lampshade as a snare drum, a pair of Indian finger cymbals as a hi-hat and a cymbal sampled in a bucket of water. After a couple of run-throughs it became clear that this was a bit too weird to play along with and so we changed the kit to a set of fairly normal acoustic samples. The plan was (and is) to sequence it using the normal sounds and to play back and mix it using the saucepans. Simon found it surprisingly easy to adjust to playing the SDX and felt that, given time, he could get on very well with it.

Louis uses his setup extensively for live work and has been working with it for over two years. He used a pan flute mixed with a percussive wood block sound which he triggered from his S900. His comments on its playability were that the dynamic range doesn't match up to an acoustic instrument but that the triggering is instantaneous and it gives him access to a complete new range of sounds.

Paul had reservations about the MIDI bass, but he found it much better when we mixed a bit of the real bass sound in with the synth. He felt there was a noticeable delay, but this may have been a sensitivity problem because the fret sensing should avoid delay problems. He also found that its dynamic response took a lot of getting used to. The sound we used was a D110 sound called 'Mesh Bass' which I've edited to play better off the MIDI bass.

Andy used to have a Roland GR50 system and recently got the GR1, which he finds a much better instrument. He has used it almost exclusively for sequencing so far and so found it interesting to try it out in a 'live' situation. He had it set up to send no pitch bend information over MIDI and switched through a couple of the internal sounds during the tune, using a Fender Rhodes sound for the chordal work and a buzzy pulse wave sound for the lead lines.

Overall everyone felt a little strange to start with but adjusted fairly quickly; the general feeling was that using acoustic drum sounds and mixing the electric bass with the synth sound brought the whole thing together much better. There seems to be a very definite psychological barrier between the aural domains of acoustic instruments (including amplified electric instruments) and electronic instruments, and it's very difficult to put one's finger on just what the split is. Dynamic range certainly plays a large part, but the variation of a tone's frequency content is also vital, and it seems that the range of possibilities available so far through MIDI is not sufficient to prevent us from marking these MIDI-controlled sounds out as being qualitatively different. Take the digeridoo as an example — this instrument can sometimes sound incredibly electronic with its sweeping frequency peaks, but the plain fact is that the sound is too interesting to be electronic; there's just too much variation for it to appear artificial for more than a brief period of time.

Monday 22nd February 1993

CHRIS "SNAKE" DAVIES: Akai EWI Wind Controller & Oberheim Xpander
FRANK FEUX: Wal MB4 Bass & Cheetah MS6 synth module
ROLAND KERRIDGE: Simmons SDX drums
RICHARD NILES: Casio PG380 MIDI guitar

Roland Kerridge, who helped develop a large part of the original SDX library for Simmons, had come in to help me with the programming on this session and had brought his 4Meg Atari and another Unitor, so we had a slightly more compatible computer system, everything being recorded with the same software. The SDX was on channels 11-16 and the MB4 on channels 1-4 on one Atari, and the PG380 on channel 1 and the EWI on channel 5 on the other Atari.

We were working on a tune called 'Reptilian Stomp', which Richard had written especially for Snake Davies.

Frank, on bass, was the only person to be coming to his instrument for the first time and his response to it after a couple of hours work was unambiguous. "I really hated it. I use a lot of dead notes and sliding and have a generally filthy style which this thing just doesn't like. It's not like playing a bass at all." We ended up just monitoring the electric bass sound and sequencing the MIDI without even listening to it — storing up much editing work for the future. We also discovered that the EWI puts out a lot more MIDI information than the WX7, which was the wind controller I had worked with previously. I therefore swapped the EWI with the MIDI bass so that it was sequenced on the 4Meg Atari with the SDX, and the MB4 went down on the 1040 with the PG380. Unfortunately I forgot to check the MIDI channels, and so one string of the MB4 and the MIDI guitar were both on channel 1, which means I'm really in for some editing!

Chris has developed his technique for the EWI, and his preciseness prevents much of the glitching that other players have mentioned to me. People have said that the keys are difficult to get used to because they are so sensitive but Chris certainly has this beast well under control. However, he hadn't come across MIDI mode 4, which I found strange. This is Mono Mode, which is really useful for legato playing on a great many different MIDI instruments. In Mono Mode, a synth will work like an old monosynth — if a note is still playing when a second note is played, the attack portion of the sound is omitted and the synth moves in a smooth legato to the new pitch. "Why is this useful?" I hear you cry with one voice. With a wind controller there are two ways of moving to a new note; you blow again (which will just start the same pitched note again) or you change your fingering (which changes the pitch). For the synth to respond like a wind instrument, the attack portion of the sound should only be triggered by a new breath and fingering changes within a single breath should change the pitch while remaining in the sustain portion of the sound. There are three ways of doing this:

• The first is Mono Mode, where the controller only sends a Note Off after the new Note On has been sent if a new breath hasn't been taken. This is true for slides and hammer-ons/offs on a guitar controller too; the attack portion of the sound is only triggered when the right hand plucks a string. If the left hand plays a new note by hammering on then the sustain portion of the sound moves to the new pitch. The advantage of the Mono Mode system is that it helps to eliminate glitching through co-ordination problems — so if you're playing a series of staccato notes and your co-ordination between triggering the note (blowing or plucking) and fingering the pitch isn't quite perfect, then instead of getting two notes sounding very close together — the major gripe about glitching that people have — you get a single attack from the note and then a continuation of the sustain portion.

• The second way around the problem is to use pitch bend when a note is changed without retriggering. This is used on guitar controllers more than wind controllers; its limitation is that most synths only have a pitch bend range of an octave — so if you try to hammer on from an open string to something above an octave, it will only leap to the octave. The same is true of slides up or down a string — they'll sound like a chromatic legato slide but they'll only go up or down the octave (except if you have something like a Roland D110, which has a pitch bend range of two octaves up or down). This is why we use the D110 so often with the MB4, although the MB4 gives you the option of using pitch bend or MIDI Mode 4.

• The third method of getting a legato sound is to use an envelope shaper, which is what the EWI does. This would explain why Chris hasn't found the need to explore the wonderful world of MIDI Mode 4! The EWI sends out MIDI notes to a synth or sampler and then you feed the audio signal back into the EWI module, which uses the breath control to trigger a VCA which shapes the volume of the sound to the breath control. Chris found that when he uses an Akai S1000 and the envelope shaper he gets a slower response time than when he uses the Oberheim Xpander. This seems odd when you think that Akai make both the EWI and the S1000, but perhaps the answer lies in the fact that he doesn't use the envelope shaper on the Xpander, where the sounds that he uses don't have very strong attack characteristics.

Richard uses his Casio PG380 in the simplest way that it allows — with all the strings on one MIDI channel and with the pitch bend off. This requires very precise playing because of the reasons mentioned above about double triggering. Putting all the strings on one MIDI channel makes it easier to play any synth or sampler quickly but it means that you can't use Mono Mode — because you need a MIDI channel for each string for that — and you can't use pitch bend — because a change on one string will send the pitch bend to all the other strings. Richard thought that "this is a concept that is not yet fully developed; 128 steps are not enough to encompass what we do when we play our instruments. These things are all prototypes in a way and so all the performances that you get in these sessions are going to be pale imitations of the actual musicianship of the person concerned. Still, they are prototypes and it'll be interesting to see what we can do in ten years time."

Roland Kerridge knows the SDX as well as anyone does, with the possible exception of Dave Simmons. He knows its possibilities and its limitations and uses it to its full effect. Roland used a fairly normal acoustic kit sound and then took the disk away to edit it himself — getting rid of a few of the hi-hat glitches and suchlike. Who am I to turn down a little help in this gargantuan task?

Next month I'll be dealing with sessions from Courtney Pine, Jigantic, an African session, and BJ Cole playing MIDI pedal steel guitar. I should also have details about how to get hold of the CD, as long as all the copyright details and financial agreements can be ironed out. There's also the possibility that I'll be able to offer MIDI standard song files on Atari format disks for all the sessions; perhaps anyone who's interested could drop me a line at the following address:

Simon Zagorski-Thomas, (Contact Details)