This sounds suspiciously like a story about how humans don’t have a clue what distinguishes music ‘with soul’ from music ‘without soul’; but have enough free-floating affect to allow them to deceive themselves on that score.
The line “Eventually, I stopped, because people got so mad I feared for my safety.” was particularly chilling. What kind of unhinged psycho gets so angry about being unable to distinguish machine output from human output?
When someone says that something “has soul”, what they mean is “I like it, but I can’t be arsed to articulate why”.
Nothing like a long article about music that has no links to audio examples.
Given that, in one of the examples cited, the same person described the same composition as both having soul and not having soul in two different performances, it doesn’t even seem to be that stable. If people were at least capable of stable (or evolving in an introspectively visible way, people are allowed to change their position) evaluations, I’d give them credit for being able to ‘like’ things. If they can’t maintain their evaluation, or even detect that it has drifted, even mere ‘liking’ seems to be giving them more credit than they deserve.
One time, it imposes a negative affect state on the listener, another time it imposes a positive one. For all we know, the ‘soul’-ness of the music correlates to the pleasantness of the waiter who served the listener dinner before the show…
Dan Brown for plot? Did he really say that? I can give you Dan Brown for plot.
- Tweedy professor type stumbles upon ancient secret
- Shadowy evil clique wants him dead
- Hot grad student chick provides clues
- Genetically odd but unkillable assassin terrorizes everybody, then gets killed
- Tweedy professor type turns his excellent adventure into a book about a tweedy professor type who stumbles upon ancient secret…
I swear he’s written the same novel fifteen times. The only plus with computers is they don’t expect royalties.
The more important question is “is computer-generated music any good?” The proof is in the pudding, and since the article doesn’t actually give any examples, the jury’s out. One factor to consider- new music (a.k.a. “modern classical”) is already pretty random and abstract, even when 100% human-generated. It’s also mostly mediocre. Like modern art, the art form has gone past the point of being relevant to people and being mostly an academic pursuit that an elite few give a shit about. Interestingly, even in the new music world, those that do it for a living would rather actually be listening to Led Zepplin or whomever than their own contemporaries’ work. So getting a computer to generate stuff that is just as shitty as the human-generated stuff is not going to be that tough. Let’s hear a computer consistently come up with a really great pop tune, or jazz composition, and i’ll be impressed.
There are audio samples linked in the piece for both Oberholtzer and Cope. Unfortunately, the links are yellow, so they can be hard to spot – a weird vagary of Co.Design’s “a color change a day” design. But they are definitely there, for all complaining.
I’ve heard of this music before (and heard it before), but then and now I’m just kind of… meh? I guess it’s a big deal that a computer can recreate music I didn’t even like when humans composed it, but it’s hard to care and even harder to play this “does it have a soul” game - I don’t know man, it’s like playing the pepsi challenge when you hate cola.
My only real opinion is that someone needs to have this technology analyze a pop singer’s music and have Hatsune Miku sing the melody for it.
This topic was automatically closed after 5 days. New replies are no longer allowed.