I'm trying to study how arranger styles work. Does anyone here have any insight to the logic behind the styles? Offcourse the aproach may be differen't from each manufacturerer, but if you know how it works, please tell me!
I've done some investigation, but all the data I've found on the subject is that in the "old days" a style was essentially a midi file with different sections. The sections represented a loop consisting of x number of bars. This loop was repeated until some kind of user input requested another loop, such as a fill or variation. The midi file then changed to the corresponding section and started looping this section until another user input. However fills doesn't repeat they only play once. But how do they do the chords? Do they have one section for each chord? or do they do some kind of transposition? And filter transpose certain notes if the chord requires it? The chord recognition process, is that based on a basic array of data? and then match the input from the user to that array to retrieve the chord? Or are there other processes involved?

I'm doing this research because I maybe starting a development project for software based arrangers, so any info you guys could provide would be most helpful!

Doc-Z