The latest sequencers are trying to move away from the 'tape recorder' emulation idea by offering much more flexible ways of presenting the musical information. But the very way that sequencers deal with data forces you to use them in a particular way.

At the lowest level, all MIDI sequencers deal with MIDI messages containing information about note ons and offs, program changes, channels, velocity and so on. For really fine control there is no substitute for a list of messages and a text-style editor that lets you tweak the actual values. For slightly coarser control, a graphic display like a grid editor lets you make rapid changes to several notes at once, and can begin to look rather like the symbology of conventional music notation. For larger blocks of notes, sequencers use either graphical displays or lists to enable arranging of blocks of notes. Beyond this is the idea of a playlist — which play these arrangements of smaller blocks in order.

All of these different ways of displaying and manipulating the musical information are nothing more than alternative 'views' of the basic MIDI messages. You still record musical events into the computer and then play them back — it may be a sophisticated tape recorder, but it still forces a linear, one-way method of composing.

Is there any other way of doing it? The problem with having powerful and capable tools is that they blinker you to other approaches to doing the same thing. For example, suppose you wanted to produce a piece of music which consisted of 12 different sections, and you'd like to be able interactively to put the sections together in any order. A conventional sequencer would have no problem with this using the arrange facilities, but suppose that you wanted to also transpose the sections as well as arrange them into order? Although most sequencers will let you move the sections around in time, transposing a section is not quite so easy, and usually requires a permanent change to the information. Suppose that transposing was as easy as changing the order of playback — say by calling the sections 1 to 12 and then supplying two lists: one showing the order, and another showing the transpositions in semitones.

Section Order: 1 2 3 4 3 2 1
Transposition: 0 5 7 2 9 4 8

This is a very concise and 'easy to understand' way of representing what we want to occur — it would also begin to stretch most sequencers. For a start, you would need two versions of sections 1, 2 and 3 to cope with the different transpositions, and these technicalities quickly obscure the structure of the music itself. If you look at that list of transpose values, it could also be represented as notes: C F G D A E G#, so you could use another section as a transpose list, but this would be well beyond the capabilities of even the most powerful current sequencers. Instead of transposition, why not have another list which details the lengths of the notes in a section — you could then change the rhythm of a section just by changing the Note Length list. If all these changes are easy to do in text, but hard with a sequencer, then what is to stop you doing it (with the right tools) on a computer?

Nothing. The tools are beginning to appear now. They may look heavily programmer-oriented and mathematical from the outside, but this is just because of the flexibility and power of the facilities they provide. (Music notation looks just as strange and confusing to computer programmers at first.) Most importantly, the software tools let you name things (anything from a couple of notes to a list showing the order to transpose sections in) yourself and then use those names as parts of larger names, so you work with constructs that make sense to you. This use of symbols (Major seventh, Lydian, Chorus, Percussion, etc.) helps make the text more 'readable' and easier to understand.

What are these tools that let you throw away the constraints of conventional sequencers? Well, names like 'Object Oriented' and 'Expert System' are unlikely to mean much to most musicians, although you may have heard of Music Composition Languages (MCLs) in the context of computer music. A couple of specific examples are probably a good way of showing the sort of thing that is now available.

Hybrid Technology's AMPLE MCL was my first real introduction to the whole concept of producing music under the control of a text file on a computer. Seeing a page or so of concise text converted to a performance of Pachelbel's Canon had me hooked, and I must confess that the fiddly and unforgiving nature of both step and real-time recording with a sequencer has annoyed me ever since. AMPLE has its roots in the original BBC computer, and is strongly supported in schools via the Master series of Acorn computers. A beta test version of an AMPLE-to-MIDI file convertor for the Atari ST has been released, but I have heard no more of a final version. Contact Hybrid Technology (Contact Details) for more.

Symbolic Composer from Tonality Systems (Contact Details), described as a "6th Generation Common Music Language" by its developers, is a tool for exploring music production. It enables Atari ST (and Mac) owners to produce MIDI Files (type 0 & 1) of music from nothing more than a few lines of text, as well as rather more complex and detailed functions like mapping any data to produce MIDI data, DIY fractals, 'random' music generators, and many other uses.

Symbolic Composer runs best within Dr. T's MPE (Multi-Programming Environment) with the KCS or Omega sequencers, and it requires an allocation of 2MB of RAM — so you will probably need a 4MB ST, and perhaps a hard disk too, for all that music you will be making. It uses Lisp as the basis of its programming language, and so it uses lists in very much the same way as the example outlined above. Lisp programs are mostly made up of functions (which modify or calculate values), macros (an easy way to have a sequence of functions), variables (containers to hold values: patterns notes, velocities etc) and lists (which can contain symbols, numbers, functions and variables in all sorts of combinations). All sorts of scale, chord and other mapping functions, macros, variables and lists are provided, as well as lots of examples on disk and detailed tutorial material to get you started.

The sort of capabilities Symbolic Composer provides include:

• Mappers — converting symbols and note-length values into melodies and rhythms.

• Generators — using various algorithms (for example: recursive, fractal, digital synthesizer and brownian) to generate symbols and numbers.

• Convertors — converting from numbers to symbols and vice-versa. Also melodies using sine waves and rhythms using chord symbols.

• Neural Expert — grow your own 'virtual' music players. Use rules to process symbol and length patterns, or even make rhythms directly from symbols. You set what should happen, and Symbolic Composer does it for you.

• Library Manager — provides access to large numbers of drum, rhythm, chord, symbol, and velocity patterns by using template-controlled libraries. You just extract the information you want from the library by specifying the sort of things you want in a template. The libraries contain things like standard jazz progressions, rock and world music rhythms, fractals, and even drum patterns. (The Mac version comes with an excellent advertising 'tour' in the form of a HyperCard stack, and the list of capabilities detailed above have been paraphrased from just one of the informative 80 plus pages...)

Consider it a sort of super customisable PVG (Programmable Variations Generator) and more! It can free you from the rigours of playing and improvising, and leave you free to do more enjoyable things like just thinking and experimenting. With fully-featured sequencers costing £500 or more by the time you add on all the extra modules, the £250 price tag (ST version) for Symbolic Composer's power begins to look like a bargain.

This sort of aid to composition is frequently found in academic use, but is rarely used by the more commercial side of music, even though many of the techniques can be successfully used in potentially 'popular' music. You may be very surprised at what you can do away from the restraints of your favourite sequencer...