Why Music Lab? Because I'm experimenting with music.
Side note: For best audio results, listen as loud as you can handle it, with good speakers. But don't hurt your eardrums. This is only an experiment.
Experiment #7503 - "Concave On Three Sides" - Create music with three different drum beats. On top of this music, add guitar and other instruments - later - using AI.
Using Ableton Live, write some music by hand for a 1 minute duration. I entered the notes myself with a midi keyboard. Just two synthesizers, with a smidgin of Ableton added arpeggio. Nothing special. No drums yet.
Now, copy and paste that music to make it 3 minutes long. Same music, strung together three times. Still no drums yet.
For the FIRST minute, add these drums called:
Darko, with Glitch Hop
For the SECOND minute, add these drums called:
Destructo
For the THIRD minute, add these drums called:
RMD Tom Kick Snare
To my ears, I found it interesting that each drum loop/sound give each "same music segment" a different "feel", for lack of a better term.
This is what it sounds like before AI added anything.
For a couple of months I thought about adding at least some guitar to this.
I found an AI music "embellisher" that would add instruments on top of my music, adding separate tracks that I could later put into Ableton Live.
Issue #1 I had was that there was only one "AI guitarist" to choose from. This guitarist is a real person that (I'm guessing) recorded enough of his guitar for AI to embellish.
Issue #2 I had, when asking AI to generate this guitar I could only control a few specific parameters like, strumming or picking / play with harmony or melody / add reverb and/or distortion. Wanted to control "style", like reggae or metal or mellow, but couldn't.
But I worked with it. If it generated something that I didn't like, I just generated something new. (It keeps every take, if I want to go back)
Generated something new many times and then had to select which takes were going to be added. Came down to "that take sounds kinda ok" to "whatever...".
Added some chorus and orchestration in the mix too, because I could.
I didn't remix it in Ableton Live, like I will eventually. Instead, I took the AI embellished music as was and had it "Mastered" using a website. It will analyze the music, try to remix it somewhat and changes volume so it sounds a bit better than the original.
I've listened to it, duh, and it's not exactly what the doctor ordered (or what he occasionally heard in his head), BUT it will suffice. It is an experiment after all.
Experiment #7503 may get some tweaks, time permitting.
Here's "Concave On Three Sides"