Interesting comment on an arranger forum! But I know how you feel. I think arrangers of any kind have still got a long way to go before they fool the discerning ear that it’s live. The ear of the player, that is…

Sure, use a live audio style, be extra careful to NEVER ask it to do something that exposes any glitches or weaknesses, you can fool most non-players for a while. Until they hear the exact same fill twice, or the inability to dial back the intensity for a quiet passage, or the impossibility of throwing in a 6/4 bar if needed.

The thing is, these are things that MIDI arrangers can do with ease. My Roland can adjust the velocities of parts (not the volume, but how hard the guitar is strummed, or the drums hit etc) in response to how hard I’m playing…

It’s things like this, the arranger’s tendency to plow on regardless of how YOU are playing that is the main area I’d like to see improved. It’s like playing with a bunch of really good deaf players!

Trouble is, this is a type of thing that would bring massive complexity and expense to fully live audio arrangers. Now you got to get your studio guitarist to not only play all chords, but at ever increasing intensity, and find a way to seamlessly crossfade between the parts.

Look, for me, audio vs MIDI arrangers is very much like sampling vs. modeling. Sampling will give you a supremely accurate picture of a sound in ONE state. To be more accurate, you need many, many recordings of multiple velocities (and round robins for repeated notes) before you can play it expressively. But it still sounds wrong as soon as you start to phrase with it unless you have a very sophisticated articulation library, and even there it falls a bit short. But that simple sample has now ballooned into a multi-GB sound.

Modeling, otoh, maybe not QUITE as close as a recording of one note at one velocity. But play a phrase, play with dynamics, and the expressive nature completely changes how you feel about that one note. No timbre shifts, no unidiomatic note transitions, simply the behavior of the real thing. For the true player, that’s worth any slight lack of realism playing one note. Heck, 40 years ago, I was using a breath controller on my DX7, and the sheer expressive nature of the sound still stands up to many modern arranger sampled saxes.

The simple fact is, yes, you can get utterly convincing backing with audio loops. But without considerable skill and considerable expense, you can do a damn thing with it other than just roll over and take it the way they give it to you. Conventional arrangers, otoh, can’t QUITE give you the same realism (though they creep closer and closer to audio each generation) but at least they are yours to alter and create your own individuality on, to shape into something unique.

At least for me, that’s worth its weight in gold! 🎹😍