[GuB-42]

> Suno, the AI Music thing. I don't know if y'all have tried it, but it now produces really good stuff

I played with it a bit, and no, it doesn't! And I am talking as someone with limited music culture, musicians are likely to be even more critical.

For the first few tries, it sounds impressive and the tunes are catchy. It used to sound wrong in the background but they mostly (but not completely) fixed that. However, after a few dozen songs, it starts to always sound the same. It is all generic stuff, the songs tell no story, it is a bit like the kind of music that accompany corporate advertisement. You can try to be more precise in your prompt, but I never had any success, it will just ignore most of the details that could make your song interesting.

The most interesting result I had was actually when I managed to get it off rails, a bug more or less. I asked it to mix two very different genres together, and it made something unsettling in a way I don't remember hearing before. But as always, further working on it proved extremely difficult, as it always tried to go back to making generic stuff, ignoring the details you give it.

Suno can do remixes though. And it is a bit like with code. LLMs are very good at porting, when you already have something that works, it can make it work in another language. But if you just have an idea, it will screw up at anything original. If you want a LLM to implement your idea properly, you have to give it so much guidance that it amounts to writing the code yourself, while struggling with the ambiguousness of natural languages.

[monegator]

re: SUNO

i actually was discussing that with a guy i met the other day, an old school producer, did succesful stuff 30 years ago. He used SUNO to reinterpret old and ideas of his, in his judgement it did an excellent job and lets him create many songs daily if he want.

Sounds familliar? the good old "let AI be steered by experienced X and boost productivity".

All in all, gun to the head, i think i am so critical because to use these tools is surrendering to big corpos. It is not a democratic tool. If it was i would probably be using it. I have finally given up and started messing with local models (well, i did already with images) but general local models are useless.

OR maybe it's me? i cannot for one moment let go and converse with the machine. I can give order to the machine.

The tech is fantastic, but the fact that it's in the hand of corpos with all interests in never letting us be able to do shit without them, makes me one hundred and one percent against it.

[hootz]

Have you tried the open weight models, but not locally? Like, using it from a provider. That way, you get access to better models while still not using private closed models, anyone with enough compute can host them, not just the big AI corps.

[newaccountman2]

I think the general local models are better than you are giving them credit for--I have heard/read good things, at least.

[CamperBob2]

A local model that's as good at music composition as Qwen 3.6 27B is at almost everything else would be absolutely transformational.

It'll happen, eventually. How could it not?

[atomicnumber3]

Suno is completely incapable of producing heavy metal. I can't speak for other genres bc I don't listen to them, but what it produces is completely hollow and devoid of what makes metal metal. I also think most metal fans will categorically reject AI-made metal on principle.

[Applejinx]

Suno's incapable of making psytrance, which is mind-boggling as that is an intensely repetitive, machine-like genre that should be water off a duck's back to produce.

The problem is that it's doing it by diffusion techniques, so all its high percussion is totally vague and indistinct. Hell, it can't even do a decent psy kick because that too is unspecific and you can't have a psy track that is vague and blunted.

Turns out you can have a production that is hollow, weak and devoid of what makes purely synth machine tracks. It can't get trancey in a serious way because it's not capable of being sharp enough.

Got an example of the genre done properly: https://www.youtube.com/watch?v=Va1KBtI81TY or alternately you could just look up some Infected Mushroom early tracks :)

[alchemism]

That is an illustrative failure. It might be absurdly simple 4/4 pattern progressions which a calculator should be able to automate. But inducing a trance requires consciousness; electrical activity in a meat brain. They will never understand such things.

[skor]

just verified, it cant make a decent techno track, nor a drone track nor anything experimental. Its creativity is subpar, it feels like listening to a producer that knows where things go but is tired of playing, zero interest in creating/ performing, it gives off that kind of vibe

[Chu4eeno]

Just tested google's (lyria, integrated into gemini), and it made an honestly not bad progressive death metal song with female vocals alternating growling and melodic (though I accidentally used gemini pro to forward the prompt to the actual music gen model, so I assume it augmented with something to make it not "generic heavy metal").

[Atiscant]

I mean, even if could produce generic metal would it produce Igorrr? Meshugga? Tim Henson? Baby Metal? All of these are driven by other things then just producing metal. I agree pure AI music would properly rejected unless there was some point to it. I could see it have some part, but then as a weird instrument. Take a model for music, randomly mutate internal weights and then let it produce a drum beat. Keep doing that unless you hit some limit and perhaps that is interesting.

[NewsaHackO]

>will categorically reject AI-made metal on principle.

I think this is a huge part of the reason people sometimes find AI criticism so dismissible; there is always some factor other than the actual product it seems that AI-made assets are judged on. With Suno, the biggest ones I've seen are 1) hating AI-created music by virtue of it being AI-created, and 2) the hate is from people who attempt to generate income from their music production, and Suno made music cuts into that pie.

[hibgymnb]

So heres a suno criticism that has nothing to do with hating AI music on principle:

The most appealing part of my favrorite music is the human element. When I hear someone singing and I know they mean it.

When they tell me a story and I can tell it’s genuine.

When I can relate to what they’re saying and who they are as a human.

Suno will never be able to recreate that.

[pickleRick243]

This is obviously "hating AI music on principle". Your last sentence means that there is literally nothing Suno can produce to change your mind.

Not hating it on principle would be something like "Suno-produced music I've listened to is derivative/soulless and has that annoying AI quality that makes me want to turn it off immediately. Maybe one day it could produce something genuinely moving and beautiful, but I'm skeptical."

[customguy]

They didn't say "genuinely moving and beautiful", they said their favorite music is about "the human element / someone / I know they mean it / I can tell it’s genuine / who they are as a human".

They didn't even say they "hate" other music, either, just that it's not their favorite.

If someone says "I only like green paintings", that excludes red paintings, even ones that have the word "green" written on them. Nothing to "fix" there, if anything, the question is why some people just won't accept that. They are acting oddly, not the people who know what they like and why.

[pickleRick243]

I didn't say there was anything to fix (I didn't use that word). The poster I replied to is free to dislike AI on principle. I'm just pointing out that his dislike of AI is in fact based on principle.

[ai_brain_rot]

Music is fundamentally about connecting with other humans.

AI can’t recreate that.

It can create sounds that are pleasing to the ears, but music is so much more than that.

[LtWorf]

Currently AI music is quite recognisable as such. I don't know if it will get better. But people who listen to music do notice.

[lc9er]

Metal, punk, hardcore - any type of heavy music, really, should reject AI-made slop. If you’re a fan and/or maker of them and are not just wearing the genres as an aesthetic, you fully know they are a rejection of corporate and governmental control.

[spacechild1]

This is really true for most music genres outside the pop mainstream. The idea of AI free jazz is just as absurd as AI punk.

More generally, we think that music (and art in general) is a form of human expression and communication. The very idea of AI music just seems absurd, as it completely misses the point of what constitutes music as an artform. Why should I listen to something that has been produced entirely without human intent? Why should I prefer a cheap simulacrum over the original?

[skor]

follow the money, they wont be selling vinyl but generating streaming revenue, just type in what you know are paying niches and off you go to fill the hard drives with slop to be paid by advertisers on streaming platforms.

[odeono]

I think this is a question of how much control the user is able to have over the end product. Music creation in particular is very difficult... I've produced music for 4-5 years, and the granularity with which one has to control the finest pieces is often mindblowingly frustrating. It takes years to develope a decent ear for mixing.

By giving up that control, you do get to a quality end result sooner, but that end result can only be an approximation to your original vision, since you're giving up the control required to shape the sound to that granular level.

[andyfilms1]

Additionally, without the knowledge of how you got from A to B, you don't know what else is possible (or impossible.) In the process of doing something manually, you may stumble across a particular setting or effect that creates something you never even considered. And now, that is knowledge you can use on the next project.

[causal]

Yeah I have played with Suno a lot and I find that no matter how I change the genre, lyrics, etc. there's some underlying quality I can't quite name that my brain recognizes and quickly gets tired of. It's fun in a novelty sense, for now.

[gedy]

> But as always, further working on it proved extremely difficult, as it always tried to go back to making generic stuff, ignoring the details you give it.

It's like any LLM, it's not a tool for if you know exactly what you want with all these knobs and fine grained controls.

> The most interesting result I had was actually when I managed to get it off rails, a bug more or less. I asked it to mix two very different genres together, and it made something unsettling in a way I don't remember hearing before.

I don't think that's a bug or unexpected, it's what AI is good for. I do these (very) old Blues covers of modern songs and it's terrific at that sort of conversion thing.

[GuB-42]

> I don't think that's a bug or unexpected

In this particular case, it really gave some "uncanny valley" feeling. That is, on the surface, it sounds like something familiar, but something is off. The wrongness was completely unintended, not prompted for, but interesting, in the same way that eldritch horror is interesting.

I wanted to mix heavy metal and hardstyle, the idea I had in mind was to imagine a battle, one side being represented by one style, and the other, by another style, each side responding to the other. It didn't give that much thought to it, but it sounded fun. And instead of getting the back and forth I expected, I got... weirdness.

I then tried to add some matching lyrics, that is, something similarly uncanny, also using AI. It also turned out to be somewhat difficult and not all LLMs managed to pull it off (ended up with GPT-5.5). That is, for many LLMs, especially the small, self-hosted ones, I couldn't get the effect I wanted, even after trying different prompt strategies. Scary words, spooky scenes, etc... no problem, but that's the opposite of what I wanted. I wanted something subtly wrong, not an B movie horror scene. Also, by adding the lyrics to the song, Suno lost a bit of an edge to the song.

If you are interested, here are the links, with and without lyrics. Don't expect anything good, it is just a little experiment: https://suno.com/s/Amu1FcrjkHsB2WTt , https://suno.com/s/gXMfNnv1g453PiaS

[gedy]

I think it sounds cool, the style mix is really interesting. Had some fun doing similar:

https://suno.com/song/a24e349b-de3b-4f98-a733-f7f70949571f

https://suno.com/s/Ff9n8A7R9k7iTpBD

I definitely come from the angle of appreciating the novelty and newness of sound from this generated music, especially when tasked with mashups. I'd not use if I had a strong vision around refining a particular song though, maybe that's the challenge with the "battle" inspiration you were after.

FYI though you can give some hints to the generation about what you want, if you look at the 'lyrics' for the two instrumental songs above you'll see them in the brackets.

[doctorpangloss]

In 2024 some people were saying, illustrators will be fine, the models can't even get the number of fingers right! They were wrong.

[citrin_ru]

A big threat for illustrators is not LLMs per se but slop acceptance. We've seen the same with machine translation - it was (and to some extent still is) worse than human translation but people accepted it because it's free. I recently seen and ad on a street which caught my attention - it was a photorealistic image of a taking off plane. It looked off and when I looked at details it's become clear it's AI generated - the 1st giveaway - stationary turbines on a flying plane, an angle relative to the runway is wrong, small details, especially in the background are off. Someone prompted LLM, seen the result and decided it's OK to print this advert. IMHO they would be better off with buying a stock image.