Hi, (warning: long post!)
When it comes to composition and ideas, I find it hard to see a pattern or regularity in my way of working but there definitely is one:
I can wander for many hours through my equipment doing "research" on how specific sounds are generated. I can try to build strange bassdrums on the md, create patches on the Virus ti and reverse engineer existing patches to see how they were made. I learn a lot of it and one thing that has been very helpful is simply sampling one bar of an existing piece of music and loop it inside the md with the aim of coming as close as you can get. It's funny to see how much variations it generates by just trying to replicate what others have made. Also this forum, the md tips from tarekith and the mnm tips of tib are great sources...although I have made some contributions to them myself, all I can say is: Thanks! It really helps understanding your equipment better and it boosts creativity and experimentation.
What others create is every now and then also a severe source of frustration as it all looks (and sometimes is) so simple. Especially techouse...just listen to some tune on beatport, proton radio or di.fm and many times you hear simple loops, made up of simple components. It frustrates me that with all the technical knowledge I have, it doesn't pop out of my machine on a day by day basis. Record labels like pokerflat for example are bad for my ego as they have some sense of melodic minimalism and simplicity which proves to be very hard to replicate.
The other extreme off course exists as well: For example the German label Perlon records comes up with very minimalistic and abstract productions which are so far above what I can imagine....doing so much with so little....
Creating the movement, the development in a tune, keeping the listener interested and triggering emotions...that's the hard part for me.
However, for me creating music is not about replication or being a look-a-like but simply having fun, learning, being in control and create something that I like, finnished or not, recognised or not. In the end I am not a full blown artist and my life (fortunately) doesn't depend on this. I' m simply an engineer...a computer/maths expert...not someone with a creative aura who has sparkles of creativity around his head but at least someone who understands what he is doing, if needed all the way up to the level of Fourier analysis.
But every now and then, creativity strikes and then, things suddenly move at lightning speed. Once a melody pops into my head, complete with rhythm, bass, chords, etc. I experience just 30 minutes in which every knob I touch turns into gold. A good bassline is programmed in just one minute on the mnm while I previously had been wandering through all kinds of arps and bass patches without any result. Rhythm programming on the md goes fast and sounds really good with every machine I introduce...It fits the bass, the cowbell fits, the hihats...eq-ing to fit it in...filtering to correct...rate reducing to destroy....it all comes together...
So the bottom line is that if I have something to work to...something that lives in my not too creative mind, things go fast and then the experience and the knowledge boost the speed. All the other acitivities are there for learning, fun and research.
There is however one thing which has improved my understanding of producing with the Elektrons. It is a purely technical issue, but I think it is worth to point out here as it helped me in overcoming some producing and arranging problems. The approach is technical and systematic:
In some way, music can be considered a rhythmic repetition of sounds. The repetition aspect can, however not be exaggerated. Just imagine creating a one bar loop and let it run for 4 minutes...that becomes boring after one minute already. Conclusion: music needs a development, a change....while at the same time some elements like rhythm and/or bass must stay the same to keep the song cohesive. So creating music is balancing on the tiny line between change and constants.
This brings up the question how many "change" the Elektrons allow for. The first answer is off course: plenty, just use the parameter locks! But let's look a little closer to that and perform some simple calculations:
The mk1 units have 8 banks of 16 patterns each so that is 128 patterns. There is also memory for 64 kits. So on average, one kit drives two patterns. Simply sticking to this rule means you never run out of kits, although theoretically you could run all 128 patterns on just one kit.
A kit stores the chosen machines and their "default" settings. The pattern, however, stores the triggering of notes (including swing and accent) and off course the parameter locks and glide settings.
So the pattern (triggering of notes and the lock system) is what allows you to create change and variety while the kit forms the basis.
Now let's assume that we want to create a song in which no bar sounds exactly the same, that is: every bar of the song is unique in terms of triggers and locks. Then, how long can a song be?
Let's see:
A pattern has at most two bars on the mk1 units, so one pattern is 8 beats.
As we have 128 patterns inside the Elektrons this means we have:
128 x 8 = 1024 beats.
Assumed that you create music at roughly 128 bpm, this means that you can create at most:
1024/128 = 8 minutes!
So if you use the mk1 units to their full potential, you can create at maximum only 8 minutes of continuously changing (every bar unique) music. On the mkII units, Elektron has doubled the pattern length which brings it to 16 minutes, but it is still not much.
Now, my reason for pointing this out is that if you are creating a full song, everything that poses a unique part of your song costs relatively a lot of pattern memory. So if you want to create a full song on the Elektrons, you will probably end up with at least half or more of the pattern memory used.
It is therefore not strange if the intro or roughly the first two minutes of a song already eat two full pattern banks alone.
Off course I know that not everything in a song is unique. Things are repeated every now and then and existing patterns are reusable.
But if you run into troubles because you are reusing a pattern in songmode (with a different mute configuration for example) and your song is sounding dull, boring and too linear then just ask yourself the question whether that pattern should have a unique element which makes it different from the others. If it does, then you are not struggling with your creativity but you bumped against the constraints of information theory. You are trying to compress (information wise) what cannot be compressed as it is unique. Therefore: copy the pattern and create another one with the needed changes to break out of the boringness...be greedy in using patterns, given the calculations made above.
There are also a few other "tactics" that can be used to break out of the 8 minutes unique issue:
-live tweaking: many people do it and (audio)record straight into a daw or sampler.
-use random lfo's as every time the pattern plays, the affected parameter will sound a little different. (It is not a coincidence that the more randomness or entropy you find in a signal, the harder it gets to compress. True random sequences therefore cannot be compressed.) The lfo's would gain a lot more power if they could affect the trigger timing as well. For example: putting a square wave lfo with depth 4 on a trigger would trigger the note 4 steps before or after the step it was programmed on. (this is probably impossible to implement as it would allow triggering outside the current pattern.)
-copy a track living on track x to track x+1 and program the needed variation by editing the triggers and/or locks. In song mode you selectively (un)mute the appropriate tracks thus creating a variation while it does not cost you an extra pattern. Off course this cannot be used if all 16 or 6 tracks are used.
-(ab?)use songmode by playing only a part of a pattern and mangling it internally. Very powerful and creates unexpected results.
-Put 2 variants of 1 bar drumrolls in one pattern. (you can stash 4 in one pattern in the mkII units) and use songmode to play the appropriate part.
If you made it this far: thanks for reading. This is just an attempt at putting a couple of things in a perspective. Perhaps it helps you in understanding and possibly overcoming loopalitus. Anyway, I still have a long way to go... but it will be fun for sure
M.