290 comments

[ 3.0 ms ] story [ 249 ms ] thread
Feel they didn't really test these genres

"Cajun synthpop chant" has no chanting or synths, it sounds more like country music with french woman vocals

Agreed, Suno still very often does not follow instructions
Yes, suno has definitely focused more on lyric and melody than prompt following. If the prompt is more than 3-4 words then it deteriorates pretty quickly. I think one reason is that there's not a lot of high quality descriptions of music around - you can guess the genre of the artist, and you can scrape reviews and the like, but that will be pretty noisy.
You can do a lot more detailed prompts with v4.5 than previously and instructions in [brackets] also go a long way now.
Could you elaborate on the instructions in brackets part?
Brackets such as [Verse] help provide waveform separation in the edit view so that you can easily edit that section without manually dragging the slider.

Others such as [Interrupt] will provide a DJ-like fade-out / announcement (that was <Artist name>, next up..." / fade-in - providing an opportunity to break the AI out of repetitive loops it obsesses about.

I've used [Bridge] successfully, and [Instrumental] [No vocals] work reliably as well (there are also instrumental options, but I still use brackets out of habit I guess).

Sure, you can do a lot of things here... stuff in [brackets] isn't sung.

For example I was trying to steer a melodic techno prompt recently in a better direction by putting stuff like this upfront:

    [intro - dramatic synths, pulsing techno bass]
    [organic percussive samples]
    [rolling galloping pulsing gritty bassline]
    [soaring experimental synths, modulation heavy, echos, sound design, 3d sound]
    [lush atmosphere, variation]
    [hypnotic groovy arppegiation arps]
    [sampled repetitive trippy vocal]
All of this is just stuff I kind of made up and wanted in the song, but it meaningfully improved the output over just tags. I think "steering/nudging the generation space" is a decent idea for how I feel like this affects the output.

I also often use them to structure things around song structure like [intro], [break], [chorus], and even get more descriptive with these describing things or moments I'd like to happen. Again adherence is not perfect, but seems to help steer things.

One of my favorite tags I've seen is [Suck the entire song through vacuum] and well... I choose to believe, check out 1:29 https://suno.com/s/xdIDhlKQUed0Dp1I

Worth playing around with a bunch, especially if you're not quite getting something interesting or in the direction you want.

When I try to mix genres that are too different, it chooses only one of them and ignore the others... let's see if it's better in the new version
Makes it weird they chose to demo it like this imho when it's not good at that. This design makes it look like its specifically that being shown off.
None of the Acid House samples have anything to do with Acid House either. The Jungle samples sound more like Liquid Drum & Bass.
> The Jungle samples sound more like Liquid Drum & Bass.

So another slop?

I tested around five genres although they were more mainstream than what you picked and they were quite good, for instance French Ska. If someone told me it was an actual French Ska band I wouldn't have doubted it. Or Klezmer music, but maybe the Jiddish wasn't totally correct.
All of the non-English lyrics are so cringe as to border on intentional parody.
Good call. Need to test some songs in my mother tongue.
I tried Urdu. The music quality seems good but the pronunciation is wrong (e.g. slight aa sound instead of ee sound) for many simple common words and helper words.

But I don't understand why its wrong. If its trained on lots of Urdu/Hindi music, no one pronounces those words like that. How does it get the a/e wrong while still singing almost correctly? It's weird.

The vowels sound subtly wrong for English on the alternative rock station, but that might just be Eddie Vedder being in the training set...
Eddies in the training set
That's a sick UI gimmick.

And it's even cleverly mobile friendly.

I can't get Suno to obey the style tags, it's as if they are not used at all in generating the output, or massively changed. Maybe they do the same trick Gemini used to generate the black queen of England. Or it's a trick to waste our credits faster, the only way to get something decent is to spin the dice many times.
It seems to lean heavily on the lyrics when choosing the style; if you pick a style that doesn't match well with the lyrics, it'll kinda ignore it.
I'm still experimenting with what works and doesn't work. Currently for style I am trying things like:

-----

STYLE: Earth Circuit Fusion

INSTRUMENTATION: - Deep analog synth bass with subtle distortion - Hybrid percussion combining djembe and electronic glitches - Polytonal synth arpeggios with unpredictable patterns - Processed field recordings for atmospheric texture - Circuit-bent toys creating unexpected melodic accents

VOCAL APPROACH: - Female vocalist with rich mid-range and clear upper register - Intimate yet confident delivery with controlled vibrato - Layered whisper-singing technique in verses - Full-voiced chorus delivery with slight emotional rasp - Spoken-word elements layered under bridge melodies - Stacked fifth harmonies creating ethereal chorus quality

PRODUCTION: - Grainy tape saturation on organic elements - Juxtaposition of lo-fi and hi-fi within same sections - Strategic arrangement dropouts for dramatic impact - Glitch transition effects between sections

---

One thing I have noticed with the new model is that it listens to direction in the lyrics more now, for example [whispered] or [bass drop], etc.

There are clear limits. I have been unsuccessful in spacial arrangement.

EDIT: I realized I didn't specify, this is when you do custom and you specify the lyrics and the style separately.

v4.5 is a lot better at adherence with detailed descriptions!

Still not totally adherent, but if you can steer it with genre, detailed descriptions of genre, and elements of the genre it's way better than v4. Some descriptions work better than others so there's some experimentation to figure out what works for what you're trying to achieve.

You can also provide descriptions in [brackets] in the lyrics that work reasonably well in my experience.

Disclaimer: I work there as a SWE.

Have been using Suno over the last few iterations and can vouch for this; steering the final product with style tags is a lot better now and I can use more natural language rather than trying to come up with what the genre specific "wordings" for certain styles of music would be. Good to know the tip about the brackets in lyrics too.

Some examples of style descriptions I've used that generated results close to what I had in mind are "romantic comedy intro music, fast and exciting, new york city" (aiming for something like the Sex and the City theme) and "mature adult romance reality tv show theme song, breakbeats, seductive, intimate, saxophones, lots of saxophones" which did indeed produce cheesy porn music.

Please make noise internally that the product desperately needs a tutorial and a prompting guide! It’s amazing if you hang out on Discord and spend a lot of time learning, but people just sitting down with it have no idea how to use it.
Duly noted, thanks for the feedback.
That is a very cool UI; super fun to just hit random and find new niche genres/styles. I'd never heard of klezmer, for example, but such a nice style! I don't know if it's the music, but it's been a while since a website has put this big a grin on my face!

I keep wanting to save some of the songs I hear. Damn, I don't think I would really be able to tell in a blind test that these were AI.

> That is a very cool UI;

I really don't like that UI. It's hard to read, and when I found something it slips. Too much form over function

> I keep wanting to save some of the songs I hear.

Just click the title of the song. If you have an account you can add to favorites, download, etc.

Ah the title isn't visible on mobile for me!
Well, I never used Suno on mobile, perhaps it's that! Here on desktop it's ok
On my 16gigs octa-core Ubuntu box, the UI stutters.
They should make it easier to download the songs. There's so much music that could be used commercially instead of expensive licensing. Someone could even set up a venture to record era-appropriate AI music to a cassette or vinyl and start selling them.

Oh, the joys of infinite public domain music!

If you like klezmer, chances are you'll like Balkan and/or Romani music too
"I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes."
If you enjoy doing art and writing then great, AI will never take that from you. If it's your job, then it can take some of the load from you so you can go do things you enjoy.
Aren't they still in active lawsuit with Sony etc about training on music without license? Yet still releasing new products?

I guess they are hoping for the Uber outcome where they earn enough money during the illegal phase so they can pay some tiny fine and keep going.

Illegal? If the record companies were so obviously going to win, they could've obtained a preliminary injunction to stop Suno's business. They didn't, so the service continues.
Pretty sure napster also continued to operate during ongoing legal actions but I am not a lawyer
AFAIK the lawsuit is still ongoing.

Suno admitted to train their models with copyrighted music and are now defending the position that music copyrights and royalties are bad for the future of music.

I think it's more straightforward to argue it's transformative fair use.
fair use requires it be non-profit which doesnt work for suno
I don’t think that’s true. Parodies are protected as transformative fair use and plenty of them are for-profit.
you're right it's a grey area, but in the field of music specifically, it's some of the most stringent requirements. the more creative it is, the more it replaces the market for the original, the less of a parody it is (not weird al), the more you will struggle to use fair use as a defense. suno generating music hits all the parts of "this is hard to defend" while also being in the the most litigious possible industry for fair use
Fair use has consideration for whether the derivative work disrupts the market for the original, it doesn't mean you can't use it for profit as we saw with the Google books case.
Not according to US federal statute.

Fair use encompasses a lot of possible scenarios involving copyrighted works which may or may not be commercially licensed: transformative derivative works, short excerpts for purposes of commentary, backups of copies obtained legally, playback for face-to-face instruction, etc.

Transformative is one of 4 conditions that the court uses to decide if something is fair use. This question is still in active litigation in the courts.
I hope no songwriters have listened to Sony music or they could be in trouble.
Suno is not a songwriter.
It's a simulated brain that writes songs ...
Do most companies just stop releasing product while in legal disputes? I don’t think I’ve ever heard of that.
They don't unless a judge orders them. It would be nonsensical to do so as lawsuits typically take years.
lawsuitS with Sony, Universal, and Warner Music, indeed. And the RIAA. and the GEMA.
I've been doing music composition and songwriting as a hobby for a decade. 4.0 is where Suno added enough features where workshopping things conceptually there first made it worth it, even for someone who can and often will, break apart stems into composite instruments and then manually adjust as needed. People always worry about what this means at the low effort "spray and pray" approach to music, but ultimately, it also allows for faster and cheaper iteration and development for all involved. Is the finished product ultimately "better" though? You be the judge.

For comparison, here's a song where I forced myself to do everything within Suno (took less than a week):

https://www.youtube.com/watch?v=R6mJcXxoppc

And here's one where I did the manual composition, worked with session artists, and it took a couple months and cost me several hundred dollars:

https://www.youtube.com/watch?v=I5JcEnU-x3s

The biggest impact Generative Music had on my life was for wedding skits. Families that wanted a funny song about the bride filled with anecdotes didn't need talent anymore.

The first time I heard it, it was incredible. The 2nd wedding that did it, it started to feel boring. The 3rd time, everyone hated it.

Similar to image-generation, we're getting tired really fast of cookie-cutter art. I don't know how to feel about it.

(comment deleted)
I enjoy playing with Suno as a toy to flesh out bits and pieces of creative ideas I have that I cannot complete at my current stage in life.

Weird, stupid things. Writing theme songs for TV shows that don't exist, finding ways to translate song types from culture A to culture B, BGM for a video game you want to make, a sales song for Shikoku 1889 to sell Iyo railway shares, etc...

Some of us have zero cultural influence and services like Suno mean we aren't listening to the original brainrot (popular music). Sure, you might create garbage but it's your garbage and you aren't stuck waiting for someone to throw you a bone.

I love Suno, it's a rare subscription that is fun.

I agree, you can make stupid ideas happen without having to make a huge investment in something you want to hear as a joke. There was a metal song I thought had lyrics that would also work as pop-country and I did quick cover of it on Suno to see if I was right.

I'm pretty sure that I actually could, if I really wanted to, create this cover legitimately and even put it on Spotify with royalties going to the original artists (it seems they have a blanket mechanical license for a lot of works). But it was a "gag" song that probably has a market of just me, so hiring a team of people would be a lot of time and money for 3 minutes of a giggle. I also would have to worry about things like if it's changed too much to be a cover and getting sued for putting in extra effort.

Distribution services like Distrokid, CDBaby, Tunecore etc will handle the mechanical license for covers. As long as you don't change the lyrics or melody, a cover will remain a cover, even if you change a genre from metal to country. The "derivative work" carveout is to protect people from changing the lyrics to e.g. something offensive and the original rights holder being unable to do anything about it.

That being said, your idea isn't original; there's already a flood of automated AI-generated cover songs being pushed onto Spotify, and they + distributors are (allegedly) starting to actively combat this.

My "idea" was to get human artists to record it, which is, yes, very unoriginal. I guess that was a bit ambiguous.
I'm sure if you lied and told everyone that you hired someone to create the tune they would like it again.
3 weddings in a couple months, lucky you!
thanks, early 30s are exhausting haha
It's probably just the lyrics, not the musical content. Popular music is mostly the same. It's clever lyrics and good meter that's more important imo. You can just dump something in from GPT or use Suno for it, but unless you spend some actual time on lyric composition, it will absolutely be campy as hell.
Suno songs also sound really poor from a technical perspective. The high end of the frequency spectrum is always very washy, reminiscent of the days of 128kbps mp3s. It sounds ok in isolation but it's very noticeable when it's thrown into a playlist of professionally mixed/mastered music.
I'm not a pro, but this seems different on the 4.5 model. It seems much crisper.
Yeah 4.5 definitely improves a lot, but I can still hear washiness in the high end. Probably not noticeable to most in isolation or poor audio devices like low-end phone speakers, but would be very noticeable on a good soundsystem in a playlist with professional songs.
It still doesn't capture meter or phrase structure well in classical songs, and there are a lot of weird artifacts. It's better than the old models, but there's still a long way to go.
> The high end of the frequency spectrum is always very washy, reminiscent of the days of 128kbps mp3s.

I have a feeling that’s by design. Firstly for computation purposes, secondly to avoid someone making a studio-quality deepfake song.

(comment deleted)
AI art is like dreams. I'm amused by my own but never want to hear about anyone else's.
I love this analogy.
(comment deleted)
Your real poetry on the other hand, pretty good!
AI art is like a photoshop drawing. If it's done by someone who sucks, which are most users if the tool is accessible enough, you will just think "That's a bad photoshop drawing". You will recognize the standard tools, the standard brushes, bad masking – all the stuff that is easy to do and that everyone will do.

That's not a tool issue. It just means that working on a raised floor is not the same a being able to reach a higher ceiling.

> AI art is like dreams. I'm amused by my own but never want to hear about anyone else's.

I don't know. Scrolling the Sora image generations feed is pretty fun.

It's got trendy memes, lots of mashups, and cool art. They've also managed to capture social media outrage bait on the platform: conservatives vs. liberals, Christians vs. atheists, and a whole other host of divisive issues that are interspersed throughout the feed. I think they have a social media play in the bag if they pursue it.

It feels like Sora could replace Instagram.

Meta's already starting to fluff up their content with AI slop. How is it better when someone else does the same?
Honestly, this is the kind of anti-AI argument that makes me care. It also acknowledges just why those of us who like it are so passionate.
Feel great: humans value creativity, and novelty.
I've cousin who's a song writer who for a brief few hours was very excited with the silly poems she could get an AI to write, and she shared them with the family group. The first one or two were ok, but eventually several of us started pointing out how vacuous the poems seemed, and how sad that she -a great song writer in her own right- was excited about that generative AI crap. She stopped right quick.

I'll say that this "Suno" thing makes good-sounding music to these non-musician ears right here, but trying a few of these I'm starting to notice it seems fake. But that's not very interesting. What's interesting is that they're going to get good enough to get past the phoniness.

> I don't know how to feel about it.

I know how I feel about it: I don't like it one bit.

https://x.com/sama/status/1899535387435086115?lang=en-GB

It's personal taste, but this is significantly better than the last couple of fiction books I've read (which were both well reviewed).

I think it's good enough that it's hard to argue is emotionally vacuous, unless you define that to mean 'it was written by a machine'

I think increasing we'll find AIs are extremely good at emotional 'manipulation' (which I mean in the same sense has how a good tearjerker is in some sense emotionally manipulative).

You say "she stopped right quick" as if telling someone the things they are talking about are vacuous, sad, and crap is supposed to lead to anything else (comments rightly deserved or not). Even when such comments are tempered in delivery it's still no more than a comment saying you said you didn't like it so they stopped involving you in it.
Our daughter's kindergarten has their own generated hymn. The kids love it. In times the teachers rarely have musical skills it compensates for this.
I find it so weird that in this age of credentialism and unnecessary qualifications, most kindergarten teachers have no idea how to play the 3 chords on a ukelele which allow them to play basically every kids song and lead a singalong
If you actually know some of the languages and you realize they are just singing jibberish (much worse than actual real songs), it's impossible to listen to. The instrumental ones can be great.
I speak Japanese — I'm pretty sure it isn't gibberish when I put in custom lyrics. (It does sometime read Kanji wrong but not when you put in the pronunciation)

I can't say anything about autogenerated lyrics.

Nope, on the Spanish stations I listened to, it's great.

Which one are you referring to?

I love listening to African folk music. I don't understand the lyrics but really enjoy the music.

I've found some okay but listening to "meaningless" music doesn't sit right with me

The French ones I listened to where legit
I had not known that this was AI until reading the comments here. I was really enjoying the 'anti-folk big band' station. Now I'm sure that's just a nonsense genre but that nonsense was more enjoyable than the stuff I've found on Spotify. I'm not sure what that says about me or the state of music but I did not expect it to be this capable yet.
The music that Suno generates as anti-folk is pretty aesthetic, but when you read into what anti-folk is meant to be as a genre, I can't help but feel that an AI algorithm spitting out music and lyrics is pretty far from anti-commercial ethos espoused by the antifolk movement.
It’s too bad AI is associated with commercialism when it’s a product of academia.
In the context of my comment and the linked website, could you explain your point?
Honestly it baffles me at this point that Suno keeps trying to generate lyrics when everything else it generates sounds so good.

All of these example get ruined by the most simple and boring lyrics imaginable. Poetry is an art and clearly the model doesn't yet grasp all of its nuances like it does for the rest of the "composition".

At this point the only thing that gives this away as AI generated are the vocals.

>Take the plunge in the strength of the breakbeat,

>Among the stars, where the dreams and freedom meet.

>Finding the ecstasy of life’s uncharted quest,

>In every pulse of the music, feel the zest.

Like... what?

> At this point the only thing that gives this away as AI generated are the vocals.

Only because the bar for music is so low nowadays. Thankfully poetry hasn't been commodified yet liked music has.

Yeah, I find it funny that they have this super strong music generation model, but then they use what could only be GPT 3.5 for lyrics.
I think it's because, like a first year writing student trying poetry, it tries a little too hard to make the lyrics coherent and storytelling in a script read from a page. A lot of real song lyrics have a whole lot of repetition, non sequiturs, and break grammar rules to fit the meter better
I mean, perhaps I should have made this more clear, but later models of GPT are leaps and bounds better than whatever suno is currently using.
Use GPT for the lyrics then, you don't have to use Suno for that.
Have you tried out the Remi lyrics model in custom mode? It’s a lot more unhinged and creative in both very good and very… interesting ways. I’d check it out if you’re interested in lyrics generation.
Love the site, and impressed by what they generated there. With that said... I'm starting to feel like music might be the last thing to be affected by Generative AI.

I IV V with different accents over the music and different drum sounds is fine, but thats not really music. It's pretty bad when you can pick out the chords progression in 5 sec. Cue the infamous 4-chord song skit by Axis of Awesome.

> I IV V with different accents over the music and different drum sounds is fine, but thats not really music

Music is more about the human that made it and their relation to you than the sound properties themselves. Same as other art. The more indirect the music process and the further you are from the living experience of the human creator, the less it resembles art. I feel art is more of an spectrum rather than a binary switch and the metric is how much direct human involvement did the audio experience have in terms that you can relate to.

Remove the human completely and you just have sound. It is likely that something like bepop, gabber or industrial synthwave would have been considered "sounds" rather than art by medieval folks or Mesopotamian people if they heard the sounds without knowing whether the source was human or not. Same with us if we were to heard some music from the year 3200 or 4500, we would likely not consider it music.

thats exactly what I'm feeling. GenAi for images or text is useful, it feels like the resolved values can be added to things or accomplish a purpose. The GenAI music feels like sound (as you put it) - like great its there, but thats not music.
I actually am of the exact opposite opinion. With image / text / video, I am much more able to differentiate between AI vs human.

However in music - there is so much badly done human music as well, for me it's nearly impossible to understand the difference between a badly done human music and a high fidelity AI music (the chord progression, happens as often in human music). Moreover, I have put Suno AI on playlist mode before and it's actually been enjoyable, and I am a big AI sceptic! Sometimes even more so than Spotify's own (although they've been accused of putting AI music on playlists as well - but I am fairly sure the weak stuff that put me off was by humans - did I say I cannot differentiate?).

Especially some music genres - like Japanese Vocaloid ; Power Metal, some country, where certain genre specific things overwhelm the piece, AI does a very good job in mimicking those from the best of the best and put meagre efforts to shame.

Here is one AI song I generated in an earlier version of Suno - let me know if anything stands out as AI: https://www.youtube.com/watch?v=I5JcEnU-x3s

and another I recorded in my studio with an artist: https://www.youtube.com/watch?v=R6mJcXxoppc

Agreed, am a musician too, and especially for popular music, AI music is as complex and often indistinguishable from music created by humans.

Kind of sad, especially for composers (which I am trying to be). Ah well, can only keep moving forward.

I’m only an occasional hobbyist, but I am super excited for how AI can empower me to write ideas I want, but which are beyond my ability and / or not possible using normal tools. I really think we’ll see a revolution in music theory once it’s easy to incorporate microtonal, multi-tempo, and other crazy stuff.

Also as we can blur the line between instrument and audio; why can’t my piano morph into an organ over the course of a piece? (I’m familiar with the Korg Morpheus and similar; I mean in a much more real sense).

I do agree with you about AI music powering a revolution - it really creates some amazing music and it's still early in the technology - But for us musicians who studied literally decades to learn all these techniques (harmony, counterpoint, music form, polyrhythms, piano, partimento...), it's painful to see that others can create a piece of music in seconds which took us a lifetime to learn how to do (I write mainly classical music and it was a punch in the gut when I first heard Udio's classical music generation. Very impressive - sigh).

An no disrepect towards anyone using AI to create music, it is here and unstoppable, but I don't currently use generative AI in music myself. Yeah, think for performed works for a live audience (at least in classical music), most people want to hear music composed by humans (for the most part). Hopefully, will stay this way for awhile, otherwise, I've been going down a road that goes nowhere. Ah well, wouldn't be the first time:)

100% agreed. But I think it is like any other revolution that pushes human creativity to a higher level of abstraction.

Someone on HN, don’t remember who, made the observation that some artists mistake mastery of tools for the art, whereas artists who focus on the actual art can roll with changes to the tools.

Agree with this for the most part. In fact, I write my pieces by hand on paper, use notation software (Dorico) to create the sheet music, listen to the piece using playback software (Noteperformer) which is connected to multiple virtual instruments (BBC Symphony Orchestra, etc.). These are all wonderful tools, and make writing music easier. For me at leas, they're not absolutely necessary, but are definitely helpful.

But, AI isn't just a tool, it's actually generating musical ideas at a highly finished level. For the first time, we have something that takes over a substantial amount of the creativity used to write a piece, a process which has always been the domain of people, and it's doing it at levels that are close to what the best skilled humans can do. Yeah, this isn't just something that aids creation, but is doing the creating itself.

Maybe one day, will use AI to create substantial amounts of the music I write, but am not nearly at that point yet - don't think most classical concert audiences want to got to a concert hall to hear AI generated music, but that may change. Guess we'll have to wait and see.

Suno is way more than that. Listen to the jazz tags or something, you're being way over dismissive here.
But why would I ever want to listen something generated by someone else when I can just generate infinite amount of the same stuff myself?
Because it sounds good, you can do both.
But what's the point of spending time and thought for music prompted by others? What I can generate sounds exactly as good and has the exact same value...
Can’t you say the same thing about why listen to human jazz musicians?
I am not a jazz musician. I can't generate their music, all the nuances, the improvisation or the feeling... in other words the human factor of their craft 1 to 1 with a press of a button.
It’s not replacing the music I listen to, but its definitely capable of replacing the random music I hear on the radio
I recommend clicking around on genres for languages you don't speak. The songs sound great as long as you don't know what they are saying. Lyrics are still a very weak point for AI music, still, infinite enjoyment as long as you're ignorant of the flaws!
One of Suno's biggest weakness is their lyrics generation, and that you can't generate lyrics without also generating a song. I think it's better to use a different LLM to generate and iterate on lyrics, which you can then pass to Suno in order to generate a final song.

If anyone here has a subscription and they can spare the tokens, I think it would be fun if someone shared a song about Hacker News.

I'm hoping that in the future tools like Suno will allow you to produce / generate songs as projects which you can tweak in greater detail; basically a way of making music by "vibe coding". With 4.0 the annotation capabilities were still a bit limited, and the singer could end up mispronouncing things without any way to fix or specify the correct pronunciation. This blog post mentions that with 4.5 they enhanced prompt interpretations, but it doesn't actually go into any technical details nor does it provide clear examples to get a real sense of the changes.

Another thing I did with LLM which I found very useful, is to give the LLM an existing song lyrics and ask him to do a similar one with different subject I give him
you can do this... just toggle to "custom" mode on song generation
Their biggest weakness is that every voice has a persistent synthy quality, like it's a vocaloid it's being sung into one of those tinny microphone toys for kids. I find Udio has much more natural-sounding vocals.
Udio has these random "halfway" singer voices that I find really interesting to listen to - a little uncanny-valleyish sometimes but interesting nonetheless. Depending on genre it can be a Nat-Sinatra, a Whitney-Dion, etc., it's really cool. It would be great if it could be saved from a song then loaded back for a new one on demand.
This song was produced a long time ago from a verbatim hacker news comment, and got released on Spotify and Apple and became a favorite at home.

Your comment inspired me to upgrade it to 4.5 because it did have that AI tinny quality. https://suno.com/s/tbZlkBL7XeLVuuN0

It sounds better but has lost some magic.

Here is the original comment - https://news.ycombinator.com/item?id=39997706

In that spirit, from the same “artist” here is your comment - https://suno.com/s/AumsIqrIovVhT0c9

And

https://suno.com/s/YGlpHptX6yXJVpHq

Not sure which I like more.

Fascinating. I find that song to be unlistenable and something about the thought of anyone enjoying it seems really wrong.
Well, that’s about as low effort as Suno gets, so I’m not sure you should expect to like it more than a laugh. But when you put effort into the songwriting and shaping you can produce legitimately good music. This is my third highest ranking streaming song and it actually got some airplay on US radio stations - https://music.apple.com/au/album/too-much-winning/1788045206...
Check out custom mode -- we've added a lyrics writing flow/editor to help create and edit lyrics, as well as Remi, a more unhinged lyrics model.

We can do better on user instruction for sure, duly noted. In my experience a lot of different stuff works (emotions, some musical direction sometimes, describing parts/layers of the track you want to exist, music-production-ish terminology, genres, stuff like intro/outro/chorus), but I think of it more as steering the space of the generated output rather than working 100% of the time. This can go in the style tags or in [brackets] in the lyrics. Definitely makes a difference in the outputs to be more descriptive with 4.5.

I got suno to render a song that I used to sing to my daughter when she was a baby. After some amount of fiddling, it produced something quite nice and for this I am eternally grateful to AI.
I still hear hissing on many of the examples, has that improved?
I honestly would like to see an open source music model that can generate the stems
Love the UI! Beautiful and great for exploration.
Immediately got a banger: "Dimensions of being" https://suno.com/song/dd3dbde0-4df6-4aec-a9be-bd2f64c281c8
Hmm. Still has the biggest issue I’ve observed with AI songs in the past: the lyrics don’t have a consistent or pleasing rhythm. It seems like the lyrics are generated with little regard to rhythm, then the melody is generated and tries to fit the lyrics without being able to change them. The result is a song where verses constantly have too many syllables (resulting in them spilling into the next meter when adjacent verses don’t), or too few syllables (resulting in awkward rests), or syllables with the wrong stresses. Occasionally the singing tries to compensate by just skipping syllables. In the second verse, the word “whispering” is replaced by an indistinct two-syllable word that sounds like “forming”. In the chorus, the phrase “the dimensions” merges the first two syllables to sound like “the mentions”.

Someday, I’m sure, Suno will find a way to fix this issue. But today isn’t that day.

You don't feel like this is incredibly generic?
I have a hypothesis that AI music and other arts will not take off.

My reasoning is that the fact that it was made by another human is really important.

Not only because you might think a piece of music is lame because it was made by AI vs a human.

But also because all the things that bring you back to a piece of art is wrapped up in the person that made it.

People who are immense fans of the Beatles, Taylor Swift or Kanye West illustrate this point.

You keep coming back because you liked this person's music before, and so you can't wait to preorder their music in the future.

Same goes for books, paintings and really all other art I can think of.

An artist develops a following that snowballs into their music being broadly consumed.

There are "AI music artists" that have been around for a decade. Miquela is the one I know about. But in that timespan, hundreds of human artists have developed followings and cultural sway that far outweigh what Miquela has done.

It seems more and more that AI is simply another tool for humans to use. Rather than a replacement altogether of humans.

> My reasoning is that the fact that it was made by another human is really important.

This presupposes that people are still able to tell the difference between computer-generated and human-generated music, which, in my opinion (as a trained musician and former producer), is no longer the case for the majority of people.

Music and musicians undoubtedly also fulfill an "idol function." But the industry has long since provided an answer to this, for example with highly successful, artificially optimized boy or girl groups. With "Milli Vanilli," they took it to the extreme by having the "musicians" no longer sing themselves, but were chosen solely for their effect on the audience. This also works with computer-generated music, only much cheaper.

"Wow this song is great! Who's song is this?"

"It's AI!"

"Oh."

I've literally seen this happen, in person.

The song doesn't get passed to friends. The artist themselves is the transmission mechanism.

> With "Milli Vanilli," they took it to the extreme by having the "musicians" no longer sing themselves

This simply proves my point, not detracts from it.

You can try to make an AI into an actual human (see Miquela) but the fakeness of that is simply insurmountable. Miquela hasn't had breakout success.

There's something very condescending in the elite culture of the past 10 years (it's starting to go away) that thinks you can get successful by being fake. It assumes the masses are stupid.

They are not.

My wive just discovered the "ambient house p-funk" style on Suno (https://suno.com/style/ambient%20house%20p-funk) and listens to it all day, not caring a single second who played it and whether the author is human. And if you look at the developments on Spotify you can well see that most people fetch music by mood or playlist and rarely care about who composed, produced or played the music, even if the meta data are available; music has become an anonymous commodity. I also spend several hours just pressing the random selection button, curious and stunned about whatever track this system generates. I don't miss the human aspect at all: it just sounds human.
I think you keep on missing my point.

It's not that the music itself is bad. I've listened to AI music while coding for example.

But for a piece of art to "take off". It needs to go beyond you. Not only do you need to tell your friends about it, but they need to tell their friends about it, and so on.

And oftentimes during the transmission process you don't even remember the name of the song, but you remember the name of the artist. Skrillex for example.

Humans care about what other humans do. We couldn't care less about what robots do.

The music market is likely goint to converge to the following two extreme segments: A) audiences who are mostly interested in idols and lifestyles, where music is merely a means to an end, and B) audiences which are mostly interested in the music itself (either as main act or in the background). For the A) audience, music is actually secondary and more of a supporting function. For the B) audience, the musicians or composers are secondary (i.e. mostly relevant as a source reference for getting more music). Since serivces like Suno have reached an excellent quality level for most genres, it opens rationalization potential for both groups of audiences. The production companies who focus on the A) audience can massively reduce their production costs (i.e. instead of a "ghost composer" and behind-the-scenes musicians they just use generated music); the only obstacle for them to overcome are today's copyright laws which don't protect computer generated music (which I personally still prefer to state-imposed monopolies enabled by patent or copyright law). The audience B) instead has access to an inexhaustible supply of new music that interests them, without having to wait for new releases.

Whether people will still learn an instrument or become musicians is a question that is difficult to answer today. The decline of this profession actually began with the invention of recording technology and has steadily increased since then. It is now almost impossible to make a living from it, and that was already the case before Suno and co. Services such as Spotify have taken anonymization and commodification to the extreme. Nevertheless, people still learn instruments and make music. It may well be that the creative possibilities offered by services such as Suno will even inspire people to make more music again.

Yeah, I think you're incorrect on many fronts.

Won't split into 2 extreme segments. It's just 1 segment- people who listen to music, often recommended by their friends or peer group.

Also, Profession hasn't declined.

Personally, I have little interest in typing a text prompt and getting a complete song as an output. However, I will gladly pay serious money for a tool that interactively collaborates with me in a granular, iterative process of generating, adjusting and mixing individual instruments and sections toward a finished multi-track song project.

Allow users to creatively engage by providing suggested starting places in the form of BPM, key and chord progressions or as brief audio and/or MIDI sketches. For example, let me give the AI a simple sketch of a couple bars of melody as MIDI notes, then have it give me back several variations of matching rhythm section and harmonies. Then take my textual feedback on adjustments I'd like but let me be specific when necessary, down to per-section or individual instrument. Ideally, the interface should look like a simplified multi-track DAW, making it easy for users to lock or unlock individual tracks so the AI knows what to keep and what to change as we creatively iterate. Once finished, provide output as both full mix and separate audio stems with optional MIDI for tracks with traditional instruments.

Targeting this use case accomplishes two crucial things. First, it lowers the bar of quality the AI has to match to be useful and compelling. Let's face it, generating lyrics, melodies, instrumental performances and production techniques more compelling than a top notch team of humans is hard for an AI. Doing it every time and only in the form of a complete, fully produced song is currently nearly impossible. The second thing it does is increase the tangible value the AI can contribute right now. Today it can be the rhythm section I lack, tomorrow it can fill in for the session guitarist I need, next week it can help me come up with new chord progression ideas. It would be useful every time I want to create music, whether I need backing vocals, a tight bass riff, scary viola tremelos or just some musical inspiration. And nothing it did would have to be perfect to be insanely useful - because I can tweak individual audio stems and MIDI tracks far faster than trying to explain a certain swing shuffle feel in text prompts.

Seriously, for a tool anything like what I've described, I'd be all-in for at least $30/mo if it's only half-assed. Once it's 3/4-assed put me down for $50/mo - and I'm not even a pro or semi-pro musician or producer, just a musical hobbyist who screws around making stuff no one else ever hears. Sure, actual music creators are a smaller market than music listeners but we're loyal, not as price sensitive and our needs for perfection in collaborators are far lower too. Plus, all those granular interactions as we iterate with your AI step-by-step towards "great", becomes invaluable training data - yet doesn't require us creators to surrender rights to our final output. For training data, the journey is (literally) the reward.

Honestly I don't think Suno knows who's supposed to be using their product. There's valid use cases for typing a text prompt and getting a song with zero fuss: background music in YouTube videos or corporate training videos, intro music for events (Starcraft fans know the ASL has been abusing Suno songs for a few seasons now), an AI-generated playlist running in the background in a cafe, etc. But those use cases don't seem worth too much, when you consider that music was effectively already free--you can hop onto Soundcloud and filter by Creative Commons and there's an unlimited number of songs you can use for these cases.[0] It remains to be seen how big the market for content creators who would actually pay for AI-generated songs rather than just pull CC tracks from Soundcloud is.

So then there's the casual end-user who's making music for themselves to listen to. IMO this is largely a novelty that hasn't worked out. I haven't heard many people regularly listen to Suno because, again, music is already incredibly cheap. Spotify is ~$15/month and it gives you access to the Beatles and Rolling Stones. The novelty of AI-generated "Korean goa psytrance 2-step" is fun for a bit, but how much will people pay for it, how many, and for how long?

I do think there's a lot of potential targeting musicians who incorporate AI-generated elements in their songs. (Disclaimer: I am a musician who has been using vocal synths for many years, and have started incorporating AI-generated samples into my workflows.) However as you point out, the functionality needed for Suno to work here is very different from the "write prompt, get fully complete song" use case.

It'll be interesting to see where it goes from here. In general, AI-based tooling does appear to be pivoting more towards "tools for creators" rather than "magic button that produces content", so I'm hopeful.

[0] One notable one is the artist "009 Sound System", who had a bunch of CC-licensed tracks that became popular due to YouTube's music swapping feature; since the list was sorted alphabetically, their tracks ended up getting used in a ton of videos and gaining popularity. https://en.wikipedia.org/wiki/Alexander_Perls#YouTube

> the functionality needed for Suno to work here is very different from the "write prompt, get fully complete song" use case.

Yeah, AI music gen is super fun to play with for a half-hour or so - and it's great when I need a novelty song made for a friend's wedding or special birthday - like, once a year maybe. But neither of those seems like a use case that leads to sustainable, high-value ARR. I'm starting to wonder if maybe most AI music generation companies ended up here because AI researchers saw a huge pile of awesome produced content that was already partially tagged and it looked like too perfect of a nail not to swing their LLM hammer at. And, until recently, VCs were throwing money at anything "AI" without needing to show product/market fit.

I'm not sure they fully thought through the use case of typical music listeners or considered the entrenched competition offering >95% of all music humans have ever recorded - for around ~$10/mo. As you said, another potential customer is media producers who need background tracks for social media videos but between the stock music industry offering millions of great tracks for pennies each and the "Fivver"-type producers who'll make a surprisingly good custom track in a day that you can own for $25 - I'm not seeing much ARR for AI music generators there either.

Currently the launch hypothesis of AI music generation puts them in direct competition against mature, high-quality alternatives that are already entrenched and cheap. And those use cases are currently being served by literally the best-of-the-best content humanity has ever created. Targeting replacing that as their first target seems as dumb as a SpaceX setting "Landing on Mars" as the goal of their first launch. There's no way to incrementally iterate toward that value proposition. Sure, targeting more modest incremental goals may be less exciting, but it also doesn't require perfection. Fortunately, music producers have needs that are more tractable but still valuable to solve - and not currently well served by cheap, plentiful, high-quality alternatives. And music producers are generally easier to reach, satisfy and retain long-term than 'music listeners' or 'music licensers'.

I mean there is a Cursor integration for Ableton Live: https://www.youtube.com/watch?v=QXSImhfS15k

Plus Ableton Live itself has a lot of generative tools these days:

https://www.youtube.com/watch?v=_RNXVfo-oLc

But I honestly don't see the point. The journey is the whole point when making music or any art really. AI doesn't solve a problem here. There never has been one in the first place. There is more music out there than you could ever listen to. Automating something humans deeply enjoy is misguided.

Generative tools and AI are great for finding inspiration. Same as using presets (which many do still think is "cheating" despite everyone doing it). Ultimately all listeners care about is the end result, not how it was made.
There wouldn't be a point in seeing artists play live if people only cared about the end result.

Listening to the studio mix on my headphones at home will always be better sound than being in a crowded concert.

I mean you are right to a certain degree, if it works, it works and if generative tools inspire you to make better music that is great. I am not so sure about that though.

I am forced to vibe code at work and it has not make me more creative. Is has made me want to quit my job.

Artists playing live is the end result though. If anything, the fact that people go to see Katy Perry in concert and not Max Martin, or how Tiesto is still a massive festival draw despite everyone knowing he's used ghost producers for 2 decades, are great examples of how little people care about the process of music being made versus the end result.

I'm not saying you need to use generative tools, but if it helps you make music you should do it. Ultimately what you're sharing with the world is your taste, not your technical abilities. To slightly expand on a famous quote in the music world -

> I thought using AI was cheating, so I used loops. I then thought using loops was cheating, so I programmed my own using samples. I then thought using samples was cheating, so I recorded real drums. I then thought that programming it was cheating, so I learned to play drums for real. I then thought using bought drums was cheating, so I learned to make my own. I then thought using premade skins was cheating, so I killed a goat and skinned it. I then thought that that was cheating too, so I grew my own goat from a baby goat. I also think that is cheating, but I’m not sure where to go from here. I haven’t made any music lately, what with the goat farming and all.

There’s not one right way you art. . . Plenty of artists like Warhol and Dali used mass production and automation.

If you enjoy writing music your way, great. But I strongly disagree that it’s a mistake to enable people to approach it differently.

I actually used to have the exact same position when stable diffusion came out.

All my artists friends where criticizing it and I was thinking it was some form of Neo-Ludditism that they were following. Why not embrace progress? No one is stopping them to not use it but if it helps lower the barrier of entry isn't that great? Surely generative AI could be used to enhance of the workflow of an artist?

Oh, how I have been wrong. In reality it has only been used to replace artists. To devalue their work. It has not place in a artists pipeline.

https://aftermath.site/ai-video-game-development-art-vibe-co...

I think the use of generative AI or at least of generalist LLM's is something fundamentally different than artists embracing new media and new processes. Like digital drawing is still roughly the same process as drawing on paper. The process is largely the same and most skills carry over. You are still in control. Using a prompt to create images is something that is not drawing.

I also recommend:

https://www.youtube.com/watch?v=1L3DaREo1sQ

I see the argument but can’t agree. To me it feels like pining for all of the lost typist jobs when word processors appeared.

Artists aren’t entitled to a job any more than I am; to the extent they (or I) do work that is easily replaced, by AI or offshoring or anything, that’s a career risk.

Those same tools enable artists to be more productive, more creative, more capable. The people suffering are those who are attached to the tools they learned with. We in tech know that’s a bad long term idea, and it is surprising in the arts. But on the whole, human creativity and overall output of great art will benefit from this change.

Your initial opinion was the right one. Too bad you let yourself be bullied into changing it by the vocal minority of online personalities. Artists that are not influencers have already embraced AI.
Try aiva.ai
Thanks for the suggestion. It looks closer to what I want than anything I've seen so far. Probably still not granular enough but much closer and headed in the right direction.
i'm not really into music, at all, but would easily pay $50/month to be able to have this functionality, too. even just the tools being what they are i've spent over a dozen hours playing around with it because it's so much fun. add the ability to separate parts of the tracks and select portions specifically to change and i'd use this constantly, all the time
have you looked at the Music AI sandbox that Google announced? it seems more instrument and editing based.
The real potential of tools like Suno isn’t in cranking out radio-ready hits. It’s in creating music that doesn't have commercial incentives to exist. Case in point: Functional Music.

I started using it to generate songs that reinforce emotional regulation strategies -things like grounding, breathwork, staying present. Not instructional tracks, which would be unbearable, but actual songs with lyrics that reflect actual practice and skills.

It started as a way to help me decompress after therapy. I'd listen to a mini-album I made during the drive home. Eventually, I’d catch myself recalling a lyric in stressful moments elsewhere. That was the moment things clicked. The songs weren’t just a way for me to calm down on the way home, they were teaching me real emotional skills I could use in all parts of my life. I wasn’t consciously practicing mindfulness anymore; it was showing up on its own. Since then I’ve been iterating, writing lyrics that reflect emotional-cognitive skills, generating songs with them, and listening while I'm in the car. It's honestly changed my life in a subtle but deep way.

We already have work songs, lullabies, marching music, and religious chants - all music that serves a purpose besides existing to be listened to. Music that exists to teach us ways of interacting is a largely untapped idea.

This is the kind of functional application is what generative music is perfect for. Song can be so much more than listening to terminally romantic lyricists trying to speak to the lowest common denominator. They can teach us to be better versions of ourselves.

Yup. My favorite genre by FAR is baroque. High quality recordings are not as wide as you’d expect, and no one’s really pumping out new baroque. V4.5 is noticeably better, even if the model shows the real “plagiaristic” aspect.

Still, I’m excited about the product. The composer could probably use some chain of thought if it doesn’t already, and plan larger sequences and how they relate to each other. Suno is also probably the most ripe for a functional neurosymbolic model. CPE wrote an algorithm on counterpoint hundreds of years ago!

https://www.reddit.com/r/classicalmusic/comments/4qul1b/crea... (Note the original site has been taken over, but you can access the original via way back. Unfortunately I couldn’t find a save where the generation demo works…but I swear it did! I used it at the time!)

I've mentioned it before on HN, but Sid Meier worked on an application called (appropriately enough) CPU Bach for the 3DO that would algorithmically generate endless contrapuntal music all the way back in 1994.

https://en.wikipedia.org/wiki/C.P.U._Bach

Ohhh this looks very cool. Thank you for sharing! Will dive into this over the weekend.
It's almost week-end! But I'm pretty sure these were shared merely as historical anecdotes, and that Suno 4.5 is the bleeding edge here...
Would you mind making these available to others?
Sure! I "re-mastered" my mini-album with v4.5 since that's more in the spirit of the discussion.

https://suno.com/playlist/e6c3f3d1-a746-4106-bea1-e36073d227...

Side note: It feels a little vulnerable to be sharing these. They genuinely helped me through difficult times and I wasn't really expecting anyone else to ever listen to them.

That's incredibly generous of you. Thanks for sharing.
These are beautiful. Thank you so much
Thank you for sharing these, I will try listening on my drive to work.
Yup, give us a tool and we'll use it. I've mostly used it for meme songs.

But I really think they've made a mistake with direction, realistically it should've been trained on tracker files, and build somgs via the method (but generate the vocals,individual instrument sounds for midi, obvs).

I think the quality would be higher since the track can be "rendered" out essentially, but also only then would it be a useful tool for actual musicians, to be able to get a skeleton file (mod, etc) for a song built up that they can then tweak and add a human touch to.

I used Suno to generate rap versions of Feynman's first lecture. It turned out really well!

https://suno.com/playlist/d2886382-bcb9-4d6d-8d7a-78625adcbe...

> we will see something similar to what is shown in Fig. 1–1
This is a brilliant idea. It been a very long time since I wanted to listen to a lecture.
This is absolutely brilliant. Is there a way for you to put this on Spotify or make the files downloadable? Not familiar with the pricing, capabilities and licensing of Suno, so sorry if this is not possible.
Amazing idea. I can hardly stand audio books, and some musical updressing might completely change that. Having a beat to it just makes it so much more fun/engaging/digestible/memorable. Kids are already taught their ABCs using songs for that reason. Why stop there, if all sorts of material could easily be put in that form? Honestly it could revolutionize teaching.
Wow, love it, so good!
I thought i would hate it, but it's actually really engaging!
Oh. Music I know can play repetitively in my mind, which is annoying at sleep time, so I tend to listen to slow music I am not familiar with at bed time. Suno did generate good such music just now for me, even though it failed miserably to generate music for genres I like during daytime. Fascinating.
I have the same thing, really annoying. I can wake up at night at the same song too. I stopped listening music hours before bedtime, but that doesn't always help and I also tend to forget. I'll try out suno for this purpose. What do you use for prompting, if you want to share?
If it's not too personal, could you share one of those lyrics?
Have you heard of "Conscious Rap"? It's the practice of creating rap music around the topics one wants to be educated, and then listening to it ambiently. That subgenre has many tendrils, as people with unrealized lyric talent make raps about politics, finance, how to maintain, and developing critical thinking skills and then in the Conscious Rap social media sharing sites their personal music takes off.
That's new to me! Do you have any recommendations or starting points you could suggest? It's a fascinating use of music
I like going to BandCamp.com and they have a Conscious Rap section, in that section you find all kinds of interesting music. Each band/mc has their own band page, and on those pages they list other sites they can be found and those listings are gold. Tiny little online labels, or just lone artists, and some are not really music artists at all, more like collage audio philosophers.
This and other subsequent tales have to be some of the most soulless and depressing misuses of other people’s collective IP. Utterly disparaging to see it applauded.
IP is a legal fiction created by capitalists to territorialize creativity.
This is exactly on point. My budget gym plays constant low quality eastern european remakes of US hits that are completely unbearable. I count the days until they switch to AI music that is assumably even cheaper but much more acceptable.
Huge yes to this

A friend recently made a simple app/site where he can pick the bpm of the music (he likes running listening to 180bpm), then see a bunch of songs and create a quick playlist he can load on Spotify with 1 click and go running

That got me thinking it would be cool to use suno/ai to create activity-specific songs, like songs about/for running or biking, or studying, or working, or painting. Instead of trying to curate popular hit songs to fit the task

(comment deleted)
This comment is absolutely on point. It is simply fun to listen to songs that are specific to your micro-situation, enjoy them a few times, and then life has moved on. I get it to make little ditties for our customers about how great our products are; about specific moments or events in time, or as little theme songs for projects or releases. It is delightful.