These songs don't have any hooks. Melodies don't get repeated. They all just meander along.
I think as they get better at making these great middle of the road songs, edgy music will reemerge. Whatever AI will be good at, will immediately devalue, in the realm of arts. Just how photography gave way to non-realistic art and how drum machines made sloppy drums (or ridiculous apex twins) hip. Then, the artifacts we see now as flaws will create its own sub genre.
So I see three ways this flows: mediocrity will be even more available, which will make artists who make mediocare music even less succesfull, pushing all human music further in human direction, except those using the unwanted artifacts of this new tech to create new sub genres.
Music, art, fashion is in the end all about changes. What we make now mostly means something in relation to what was already there. It's a big conversation, spanning millenia, and this isn't the last word.
> These songs don't have any hooks. Melodies don't get repeated. They all just meander along.
I've been playing around with it for a few days now. While I agree that it seems impossible to create songs with a more "sophisticated structure" (for lack of a better word off the top of my head), you still can get better results by fine-tuning, as is always the case.
If you just request "rock music" or "jazz", you get very dull, generic variants of the requested style. But on a second thought, isn't that exactly what should happen? You throw all the rock music on this planet in a blender, turn it on, and what you get is the most average rock music there is.
If you spend some time spicing up your prompt with flowery language or just a bunch of adjectives, you can get a sound that seems less bland. When you supply lyrics, using square brackets to denote verses, chorus and bridges can also result in a somewhat more structured song, but I found that the AI is pretty lackluster in that regard and you often need several attempts until it follows these inline orders.
So yes, in its current form it has mostly a novelty factor, this is Stable Diffusion for music, but I can easily see this being useful for a small indie gamedev who needs some BGM, or an alternative to the YouTube music library. Instrumental sounds fine, it's mostly vocals that have this clear digital distortion if you pay attention. It's surprisingly good, but still bad.
The melody of the instrumental pre-chorus is repeated consistently, and it's even used as the hook in the chorus; the pre-chorus is actually building up to the chorus. I'm impressed.
No need for AI or face recognition to do that: BT equipped system in elevator detects phones through their Bluetooth Device Address (sort of MAC address for BT), then calls a central repository of names to pair that phone with an identity, DNS style, then calls an advertising seller (for example, Google) that will happily give back the appropriate ads according to target, time, location and context.
Or, in a even more dystopian and much simpler scenario: some system app in the phone detects a beacon in the elevator through the above Device Address, then calls home (again, Google and others) and receives the ads to be shown in the elevator (or any other place with a audio/video system plus BT).
I know it isn't the point you are trying to make, but I can't think of a better way to re-enforce OP's point of "mediocrity will be even more available" than a mid-00s style, pop country song about eating hamburgers. "Toby Keith but with less to say" might be the gold standard of mediocrity.
Your comment seems to presume that AI will not get any better than it is now. Imagine an AI that understands how to create deep, impactful music better than humans do, because it understands how music works at a biochemical level. Imagine it can even predict the dynamics you're describing, about employing "unwanted artifacts" in the music as a way to evolve new genres. It would no longer need to create such obviously derivative works at that point, and it could generate music that sounds completely unique to us. It may take a long time to reach that point, but when it does, the kinds of music that it generates won't be able to be dismissed so easily.
Your comment seems to presume that the line between what a human is and what an AI is will stay clear. I'm predicting that this line will be increasingly blurry. Some people see smartphones as cybernetic extensions. When I call someone across the globe, did I do that or my phone? Is that a capability I posses or my phone? Does it make sense to separate the two?
Even if AI gets way better, the one thing that I don't foresee changing is what makes things valuable and or desirable. Sparsity. If everyone has it or can create it, it's not special. I think the GP was referring to this sparsity.
You're talking about agency there I think. No your phone didn't do that, you did.
However, if later down the line we create autonomous agents that can initiate the creation of said music themselves then I'd call that enough agency to say that the machine is "making music". Could probably almost do it now; tailored LLMs, image diffusion, music diffusion and you could have an ML agent that acts as a musical artist; posts to instagram etc with images of their persona "working on something new", releasing new tracks, bantering, etc. There are already AI OF stars apparently.
We could say "yeah but a human still set it up and told it to make music" but I would discount that; pretty sure no human has total agency, we are all impacted by our environment, peers, culture and all sorts of other external influences.
And I don't think sparsity changes things (maybe in the material world) but culture certainly does. Things are popular because they appeal to us in bulk, rarity/sparsity always result in higher effort for the payoff and so decreases popularity.
Alright, what if someone has a neural implant and that thing is running and or connected to an AI? Are you still sure the line is clear and sharp in that scenario?
This will be used for jingles and scoring by low cost studios and marketers. Making money at music was already hard and will only get harder. But maybe this makes sense in a way, music creation just became much easier and accessible to more people.
Alongside this prompt based music creation, we have AI powered autotune and voice masking, which allows even the worst singers to sing perfectly. Popular music was already being retuned during recording and this just makes it easier. In hip hop, old songs beats and verses are getting reused wholesale with little to no modification. See Jack Harlow’s First Class and Tyga’s Bops Goin Brazy. A lot of musical success is now business connections, studio promotion and timing, not being extraordinarily talented. Most rappers now don’t write lyrics down or freestyle, they record a line, pause, record another line, pause , erase the last one, rerecord, and then they’re done. This type of thing an LLM is made for. I think this means less works of art on the ‘radio’ and more garbage music forcefed by studios.
You can make them have a hook if you structure it right. The default prompt on Suno winds up as "verse, verse 2, chorus" for the structure, which indeed does sort of sound like a song, but not a song at the same time. When I changed things around and put a chorus in twice I wound up getting an actual song with a hook.
I assume the 2 minute limit has a lot to do with this. Often I find Suno generates anywhere between 40 seconds to 2 minutes, with the shorter generations having less of a recognizable structure as a song. If instead it was 3 to 4 minutes, I think it'd be radically different.
Yeah I have noticed that Suno's lyric generation is very canned sometimes, but also for timing - it oftens generates 3-4 (longish) lines for a verse where for the music there should be 2-3.
That's the ChatGPT style. It is extremely stereotypical when it comes to writing lyrics or poetry: 4 lines, rhyming, with metronome regularity. Both Suno & Udio seem to use ChatGPT for lyrics, and it's a terrible choice, which should be ripped out in favor of Claude or anything else. (Basically, any LLM which can pass the "write a non-rhyming poem" test prompt would be a better choice.)
> These songs don't have any hooks. Melodies don't get repeated. They all just meander along.
Perhaps this has more to do with the input rather than the capability of the AI. The license is somewhat meandering.
Also, there is a neat build-up to the ALL CAPS portion, when the song takes on a much more full and powerful quality. Plus the whisper on (the "Software"). It's obviously responding well to the features of the input itself.
Other than The White Stripes, what popular bands have had clearly unskilled drummers? I actually enjoy listening to more amateur music, so this might be a request for recommendations of a sort.
Meg White is widely regarded as a great drummer by her peers. "Unskilled" is definitely not true: she just plays in a style that is devoid of virtuosity, and it fits the music very well.
Fair assessment. A very skilled artist could use an unskilled style - see Nirvana's cover of Where Did You Sleep Last Night or even John Travolta's dancing in Pulp Fiction.
Maybe sloppy isn't the best term here. More like not being beat-perfect. E.g. using slightly off-beat hits or slight tempo variation as a way to emphasize other instruments or lyrics.
Pre drum computers these changes would more likely have been considered to be sloppy drumming, whereas these are more likely to signal authenticity these days.
Simpler Rhythms could be made creative. I am not a drummer or expert in researching popular bands but e.g The Ting Tings have a very interesting rhythm section.
Have you tried annotating the lyrics with stuff like (hook), (chorus), e.t.c.
I've used chatgpt to generate the lyrics for a random word, while giving it some artist names to draw inspiration from and suno manages to put in most of the appropriate musical characteristic of the annotated segments. Most notably (chorus) and outro (outro) in this example: https://suno.com/song/23e1d6de-569c-4c19-9551-cfda8e1bcd5a
I don't think it's fair to conflate "having structure" with "being formulaic".
Even the most abstract art is not about splashing paint in a canvas and calling it a day. It's about doing something close to this, but creatively and within a specific framework.
It takes much more skill to produce songs with hooks and refrains than to produce random music.
But it takes even more skill to produce something that is creative, not formulaic, but still has some structure that's pleasant/fun/interesting/etc to the listener.
The second best art of this time will always remain "who's afraid of red, yellow and blue?" after it inspired such fear in someone that they sliced it up with a knife.
I think as they get better at making these great middle of the road songs, edgy music will reemerge. Whatever AI will be good at, will immediately devalue, in the realm of arts. Just how photography gave way to non-realistic art and how drum machines made sloppy drums (or ridiculous apex twins) hip. Then, the artifacts we see now as flaws will create its own sub genre.
So I see three ways this flows: mediocrity will be even more available, which will make artists who make mediocare music even less succesfull, pushing all human music further in human direction, except those using the unwanted artifacts of this new tech to create new sub genres.
Music, art, fashion is in the end all about changes. What we make now mostly means something in relation to what was already there. It's a big conversation, spanning millenia, and this isn't the last word.