Reading about the book that inspired this, written by an artist who closely studied the styles of others— that’s impressive.
Reading an AI-generated site about some AI model that shits out some fake Bill Evans on command—that’s not. Any moron with enough tokens can do it. Congratulations.
I think the last sentence of your comment is the most exciting thing about this whole article!
There are at least three enormous gulfs here:
First, just understanding music and even jazz as a prerequisite to being able to do this.
Second, being able to do this using machines instead of a piano itself, which is unaffordable for a lot of people.
Third, the capturing of the je ne sais quoi of musical style, which is almost impossible to articulate.
So for me bringing this all together, it allows you to capture an improved understanding of different styles without needing to invest a huge amount of time because you can feed the same input into multiple style models and see how it diverges. That's a control which can't usually be obtained!
As a technical feat, what AI research allows across domains is a remarkable thing.
> First, just understanding music and even jazz as a prerequisite to being able to do this.
Do you have to understand music to train an ai to do this? You can be a fan without understanding music. You can be motivated and have an ear for what Cedar Walton sounds like without knowing anything about what harmonic and rhythmic vocabulary he had or didn’t have.
> Second, being able to do this using machines instead of a piano itself, which is unaffordable for a lot of people.
What? You can get a piano or a keyboard for a couple hundred dollars.
Unless you mean actually training and being good at the piano being unaffordable, no, that's just lack of skills and/or will to train to become good at it. Anyone can afford to buy an instrument and become good at it if they cared.
I'm sure no actual musician is actually impressed by this slop, and as the GP comment says the actual book written by a human to whose experience I can relate to and be inspired by is really impressive.
This is amazing. I like very much that the approach is using symbolic music information instead of audio. It might help to better understand music some day (i.e. "really understand", not the helper projections the traditional music science works with). It's also interesting to notice the slight time jitter, likely due to the more discrete and limited "information density", or possibly round-off errors due to limited resolution of the weights. Maybe the paper answers it.
The Oscar Peterson one does seem to mostly understand that he had less mobility in one of his hands after the stroke but it uh... doesn't really get much else right.
Yeah the four samples I heard all sounded kind of formulaic. Too obligated to continue the mold of the last thirty seconds. Impressive and quite pleasant but no big ideas and no thrill of flight.
That “transcription” using piano roll (but without even showing the actual keys on the side, and without bar lines) is impossible to follow and doesn’t make sense for jazz, where chromaticism, key changes and syncopation are so common. I see a note and then another note that looks a little higher, are they a major third apart? Or maybe it’s a minor third? Was the first note a pickup? Or is it on the one? No way to know
Reading an AI-generated site about some AI model that shits out some fake Bill Evans on command—that’s not. Any moron with enough tokens can do it. Congratulations.
There are at least three enormous gulfs here:
First, just understanding music and even jazz as a prerequisite to being able to do this.
Second, being able to do this using machines instead of a piano itself, which is unaffordable for a lot of people.
Third, the capturing of the je ne sais quoi of musical style, which is almost impossible to articulate.
So for me bringing this all together, it allows you to capture an improved understanding of different styles without needing to invest a huge amount of time because you can feed the same input into multiple style models and see how it diverges. That's a control which can't usually be obtained!
As a technical feat, what AI research allows across domains is a remarkable thing.
Do you have to understand music to train an ai to do this? You can be a fan without understanding music. You can be motivated and have an ear for what Cedar Walton sounds like without knowing anything about what harmonic and rhythmic vocabulary he had or didn’t have.
What? You can get a piano or a keyboard for a couple hundred dollars.
Unless you mean actually training and being good at the piano being unaffordable, no, that's just lack of skills and/or will to train to become good at it. Anyone can afford to buy an instrument and become good at it if they cared.
I'm sure no actual musician is actually impressed by this slop, and as the GP comment says the actual book written by a human to whose experience I can relate to and be inspired by is really impressive.
If you don't care enough to write it, I don't care enough to read it. Will be checking out the Hyman book though.