This article is currently just a placeholder for some useful information that Stephen posted on the M3 forums. This will later be fleshed out into a full tutorial, but for now this should be helpful for some users who are trying to create custom GEs with the KARMA Software for their keyboard model
You have the KARMA Software and have worked through all the tutorials included with the software, but you still don't have a solid understanding of how to create a new GE from scratch to capture the "feel" of a groove you've heard in a piece of music, or that you are hearing in your head.
At present there are two ways to create a GE on the KARMA Software:
1) Import a simple Phrase and turn it into a GE using the "Import To GE" Editor - you can do drum grooves, and melodic phrases this way, resulting in a Drum GE or a Melodic Drum GE, or
2) Create one from scratch (or edit an existing one), and this type of GE can be any of the RTC model types.
To create one of the second kind, let's say I want to duplicate some sort of riff or pattern, I would break the riff down into the various "attributes" in my mind, and try to make KARMA perform them.
For example, with a "CL1- Comp/Lead" RTC Model GE (one of the most commonly used RTC Models), the Note Series is contructed from the notes you supply as input (and by "input" I mean "playing a chord" or triggering a chord pad). The default Note Series parameters in an initialized GE of that RTC Model basically create a 3 octave or so arpeggio of the notes you supply. You play 4 notes, you get those 4 notes replicated across 3 octaves. The Rhythm Pattern then drives the Index Pattern to play through this series up and down, based on the Phase Pattern parameters.
So I might first temporarily set the Index Pattern so that it was a single value of 0, in both Phases. This will just play one note (the root note) over and over. (I might also temporarily set the Phase Pattern so that it is only using one Phase.) Then I can concentrate on the Rhythm Pattern, and make it do the Rhythm of the riff I am hearing in my head. Once I've got the rhythm, then I might play with the Index Pattern to try and get the sequence of notes. If there are chords mixed with single notes in the phrase, that's where the Cluster Pattern comes in. Then I would listen to the levels of the notes, and use the Velocity Pattern to get those as well, then the Durations, etc. In this type of GE, you build the effect using the individual attributes of the musical phrase.
This idea of the "separate musical attributes" is described conceptually in the following WIKI articles: