As its name suggest, this is an evolution of the Live Controlled Chaos Orchestra (lCCo v1) . All the tracks included in this set are played LIVE, and only recorded in one bounce. No post editing (except few bar cut here and there). Only the drums are writen, all the notes are randomly generated live! In the first version, i tried to develop a technique to control, with a natural feeling, a multitude of instruments simultaneously using midi randomisation, side chaining and re-routing between samplers and soft synths in ableton, all controlled by a keyboard.
In this version, i tried a different approche to get a more divers but predictable behavior and a better sound (using a pre arranged mastering using template to follow from stage one) with precise live control. In v1, impulses came from keyboard input randomised and scaled, then synths output where processed throu sidechain gate allowing/cancelling each other. In v2, midi impulses are generated by midi loops mostly extracted from drum audio samples. The great thing about that it fits the groove, and the overhall feeling of the beat! As well, with ableton groove management, it give a lot to play with! The rest of the control is done with knobs, faders and one X-Y joystic . It is ridiculous how powerfull just a 2 axes joystic can be when linked to series of filters, effects, synth parameters and ableton’s rack selector! Warp and shape sounds on the fly.[break]
Mastering wise, it is a bit of a chalenge. I faced the same problem than before with power processing. Even thou i now work on a twin quad core xeon (pc server running Mac OSX (hackintosh)) theorically running up to 16core with 24G of ram, i have been tempted to add bunch of vst/i rather than rearranging previous project for simultaneous track recording. Ableton seems to handle multicore processing alright when you have a bunch of tracks (synths – sampler – audio) playing simultaneously with short chain of effects but recording everything at the same time while keeping midi playable live is tedious due to the midi architecture of the projects. Midi pass by a number of different processing (one after an other) before reaching the all the synth/samplers (procedural processing). Any change to any of them will have effect on all the other down the line. Therefor it can’t be calculated simultaneously and have to be processed before reaching the instruments. Load of core is good, but in this case, higher clock speed would be more advantageous.[break]
An other difficulty i encountered trying to record is to define exactly what to record and “where” (midi/audio). Because of the multiple layers of ramdomisation, midi quickly gets complicated. If you record the input only, it will play differently each time. If you record the randomised notes, it will most likely play the same but any slight modifications to the racks will change everything. Not speaking about the mess it is to change from recording to reading. It is ok when you have few tracks but with 30+ instruments, it is quite a task. Same goes for the audio. I use a lot of side chain gates and re-routing witch complicate things greatly. Two solutions i’d like to explore is to either use rewiring with Logic or if the opportunity comes, to use a proper studio with multi-input recorder. Out of question at this stage.
This workflow works well when carefully tuned but get really messy when filters (vst or synth parametres affecting the frequency) are used. For example, a filter applied on mid sounds can give interesting suttle low or even sub movements but are gonna be crushed by the mid channel EQ if place before the channel and will miss frequencies if place after. To get around this, i try a lot of different ordering, re-routing and paralell channels to find what i am looking for. It can sometimes give unexpected effects opening doors to new discoveries.
I can feel the quality – power – presence increase after each projects. No secret, EQ are the second thing to pay attention to, every things need to have their place in the spectrum. The first would obviously be the generated sound itself. Good quality samples and well tuned synth are unavoidable. You can use tones of plugins, bad sound base will never give sound great! Get rid of low/mediocre quality samples (or keep them for reference). As well, slimer and well organised sound database is easier to navigate. Graphic EQ helps you visualise it and with some (swear only by the fabfilters) you can listen the part of the frequencies you gonna cut. Extremely handy! Subtle multiple parallel compression may be the third? When i look at the graphic wave of the recorded tracks, i can see that there is a lot of improvement to do, that there is space for little more compression, limiting and general mastering need more twiking.
Hopefully, Ableton has an amazing feature, the rack instrument. You can do so much with it. Pile tons of synth and sampler, plugins midi/audio, and play with the selector. The best part is that you can save all and re-call it anytime in any projects! Same goes for the midi control tracks. Finding the good combinations is long and difficult but once found, save it and re-call it : )
The only problem i found is when is use an indispensable plugin max dsp based the Chord arranger . It works in two module (identifier/receiver) and have to be inserted in this specific order. Each time it is used in a reloaded rack instrument, it need to be replaced.
The midi, that is the tricky part…
As i said previously, in V2, impulses are generated from midi loop generated from audio loop using abelton slice to midi internal tool (Ableton9 tool look to have had great improvement). It will generate following notes from each transient (metode i use) going crescendo. Those notes are then send to a chord vst so they all transpose to the same note. It only gives a midi impulse on each transients so each note is on the groove (can be offseted too). Interesting thing; depending on the lenght of the loop, the transposed notes will jump octaves witch in some case, gives cool variation in the melody but as well, if the midi spreading is define by octave range (ei, C1/C2 to group1, C3/C4 to group2 etc..) will unexpected sound variations. I have used side chaining too (gate and/or compression) but not as much as in v1 (not midi, but when tuned for, it can play a big role in the play.
One track loaded with the Chord arranger identifier module with the keyboard as midi input and the receiver module just after the chord vst (ableton internal tool) in the rack instrument (problem thou when used with racks it have to be reloaded manualy in specific order). Now, all the notes will correspond to the keyboard key input.
An other important part of scaling and randomisation is the Schwarzonator from Henrik Schwarz. It is a must have!
The Schwarzonator is one the amazing free plugin made with Max (found some bugs when several instances are used in the same project that that gaves really interesting behaviors! (Trying things like using tool in different way they are supposed to be used, connecting things that are not meant, trial and error is the most rewarding approch when in search of creativity :).
Max is the way to go when it comes to ableton and electronic music. If you have the skills, the ideas, the time and the will to learn max “programing” (if you can call it that way…) you can create about anything! Maxforlive website is a proof. Unfortunatly for me, i don’t the skills! but i am sure that the “midi machines” i build in each project could be synthesised with Max in some kind of crazy multiple output midi generator. I love to learn new stuff but going through the process of acquiring the necessary knowledge to build such a thing put me off a bit for now and rather spend more time on developing the concept. If someone with such knowledge and skills would be interested to collaborate, contact me !
Stay tune for lCCo v3!