Personally, when I watch videos on YouTube, I’m amazed by programs like Suno or Mureka, which can produce not just a single musical composition but an entire album.
Personally, given the personal investment required to pursue my passion—an investment in DAWs and products like those offered by Native Instruments, Cubase Pro, Arturia —I tell myself that I’m lucky to be just a modest, non-professional musician who is accountable only to myself.
I suspect that Cubase, FL Studio, and many others will at least start using AI to generate modules that simplify mixing or the synchronization of instruments
But have you ever found yourself staring at a blank score and wondered if you might have just lost a little bit of your soul as a musician and composer?
Although I’ve had FL Studio for a few years now—before investing in Cubase—FL offers a chord progression generator similar to the one in Cubase, and also in my Instacomposer3 software.
I don’t think they’re comparable to AI, but I still have a hard time using them because, for me, putting together chords and letting software compose MIDI tracks for instruments—without any knowledge of harmony—leaves me a bit perplexed.
I played a lot by ear because, since I wasn’t really diligent in my advanced guitar lessons, I thought strumming four to six chords would do the trick.
That was before I started composing in Sonar 8, then FL Studio, and now in Cubase Pro.
There comes a time when you start to doubt the path you’ve taken—all that investment in DAWs, VSTs, and compositions that lead nowhere—and then today, just anyone can whip up a track without having spent time mastering the instrument, the practice, and the theory that goes along with it.
Well, after 45 years of music, I’m not giving up just to be lured by the siren song—and even if my musical compositions don’t amount to much, I’ll still be the one who truly stands behind them.
Well, I use Steinberg Spectralayers, so I guess I use AI.
All you 100% anti-AI guys - you’ve never used Spectralayers, or other stem-splitting tools? No judgement, just asking to understand your viewpoint better.
I also bought Izotope Neutron for some mixing, which has some AI in there I think I’ve read, but I haven’t used it yet.
Just to be clear, generative AI is what I’m discussing - not just anything that falls under the umbrella term AI which refers a lot of different things.
I got SpectraLayers 10 Pro with a Samplitude Suite…registered it with Steinberg…and have never opened it! I have Melodyne Essentials as well, I used it once a few years back to fix a single note in a guitar solo that bugged me! For the most part I record and if it’s not good enough I delete and record again…no comping or trying to combine 4,000 takes into a single track
For me, less than 0.5%. When you don’t have the computational resources to achieve fast results with them, it takes longer than it would take if you did the thing yourself. And since I can deliver subpar results faster than AI, I just do it myself.
In a very roundabout way, poverty is the solution to AI. Can’t afford hardware to run it, can’t afford to pay subscription to have it run on someone else’s hardware.
I have occasionally used the bandmate groove generator in ezdrummer 3 but I don’t think they claim it’s ai, it also struggles with fills in the right places etc.
Other than that I tend to play everything on guitar and then transcribe it to piano roll myself.
I am more interested in the wave editing possibilities of ai, for alignment and pitch change. Ai should be able to analyse singer traits and help - intelligent pitch correction, plosive control and debreaths.
