Lately I've been trying to improve the way I do things. Here's a catalogue of how it's been done. The first year I did music, I mucked around with a synth with a friend and made feedback noises using a borrowed tape recorder of my grandfather's. Despite how primitive this sounds, it was actually fun, and got me sort of thinking about sound. Then I spent a year and a bit in a sort synth/samply ambient band. Then I spent eighteen months with a four track recorder assembling various bits of synth, bass, feedback burblings. The result was always interesting, if poorly produced. Then I bought a sampler and used a computer to sequence it. From sort of low-fi early 70s ambient stuff I went to blippy, poppy danceish, 80sish stuff. It would probably make more sense now. The arrangements were often a bit complex, but the production was often appalling, and the drum tracks really unconvincing.

By 2001 I was over all that, and embarked on a modernisation phase, switching to AudioMulch from a PC. AudioMulch is a sort of cheap and cheerful VST host with some really good DSP granulation tools. It's conceived from an electric acoustic point of view, but with a consideration for ease-of-use and performance, so it's a sort of halfway house between the austere preciseness of C-Sound and the easy, everyone's-using-it feel of Live.

Compared with using a hardware sampler, Mulch has a greater flexibility with chaining devices and manipulating sound, and i have been able to refocus on the production side and improve things there.

However, there are downsides. AudioMulch is not geared towards note sequencing. Is has arpeggiator and bass synth contraptions (the AM term for a basic processing object), but these have a sequenced in patterns limited to two bars and one bar respectively. With up to 20 presets for each pattern you can potentially automate quite a bit of tonal variety, but it's timeconsuming, and I don't have time to fuss with all that. In short, I've gone from fairly tuneful tracks to tracks featuring much more bare and often randomly generated tone clusters(!).

Another downside is that Mulch's only supports the 4/4 time signature, and tempo changes cannot be preprogrammed. Most of the time 4/4 is enough for me, but sometimes it would be nice to slott in a 5/4 bar to kind of mix things up, or speed up the tempo towards the end of a track.

Finally, Mulch's rhythm and note sequencing elements are modelled on the 80s techno production paradigm (think 808s and 303s). This makes Mulch great for making techno tracks, but rubbish for anything else, aside from granular manipulation of file data.

Because of these downsides, I often feel like the tracks I write are more mechanical exercises than full musical statements. Too often they start out as a couple of looped harmonic or rhythmic textures, against which I then assemble some beats or basslines, and then proceed to shuffle these elements against each other until it gets boring, and then fade out. Hey presto, another classic Aquaboogie track.

Anyway, with a view to varying this up I've gone off and bought a license for EnergyXT2, another low-cost VST host, which may be itself hosted as a VST. I've been running XT2 within Mulch to provide synth part sequencing, and it works quite well, although there's a significant CPU cost involved. So far I've used it to drop the sort of lush four part chords I've been unable to do in Mulch (or rather, I could do, but I'd need two arpeggiators, each with transposed second oscillators - and to produce another chord I'd need to build two extra patterns and automate them - a complete pain in the arse). Eventually I'll probably start using XT2 as the basis for building drum patterns and synth sequencing, and use Mulch itself more for sample loops and general 'mulching' up the sound.

Once I have a more flexible environment to work with, I'll then be turning to the 'vision' thing. Given the current state of music fashion, and the state of the music 'industry', it's baffling figuring out what to do or even what can be done. One constant in my musical practice is continual incremental improvement. Actual musical 'achievement' has been thin on the ground, but improvement is always forthcoming.