• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

AI and Future of Music Production

Quite possible, but as I already said, these were actually artists, and geniuses at that.
it's always been like this: in music and art in general, the percentage of garbage produced is very high, in my opinion, I would say 99.5%. It is the remaining 0.5% that gives us hope, that makes us forget the insult of being part of this miserable humanity
 
In the above given examples, which lyrics? , sorry i don't quite follow here.
for this kind of ehm.. lets for the sake or argument call it music, MP3 at the lowest possible quality is more than adequate.

I was thinking of the lyrics in the post you quoted. I sometimes forget that doesn't carry over when I quote (eg) your post. I should have included both in my reply so my post would make better sense.
 
I was thinking of the lyrics in the post you quoted. I sometimes forget that doesn't carry over when I quote (eg) your post. I should have included both in my reply so my post would make better sense.
No problem, i knew that you meant Lyrics in general:), and not above examples
 
If the artists you mentioned could have known AI, they would not have seen it as a threat, on the contrary, they would have found all the creative possibilities of the technology. If you like jazz, you will remember Herbie Hancock when he dedicated an entire album to hiphop-influenced electronic music (Future Shock): it was a scandal for jazz purists, yet that album went down in history by winning the 1984 Grammy Award.
AI will become a source of inspiration for artists, as technological innovations always have been
Great LP - Future Shock!
 
Seed stage Seattle startup Soundry AI creates text-to-sample generator

Article includes video demo and lists some competitors. Y Combinator is a very high profile startup accelerator followed by investors


 
Last edited:
Stability AI’s audio generator can now crank out 3 minute ‘songs’ based on a text prompt or sample

"The songs can actually sound nifty, at first, until the seams start showing. Then things get a bit creepy.

For instance, the system loves adding vocals, but not in any known human language. I guess it’s in whatever language that makes up the text in AI-generated images. The vocals sort of sound like actual people, and other times they sound Gregorian chanters filtered through outer space. It’s right smack dab in the middle of that uncanny valley. The Verge called the vocals “soulless and weird," comparing them to whale sounds. That tracks.

Stable Audio 2.0 makes the same weird little mistakes that all of these systems make, no matter the output type. Parts can vanish into thin air, replaced with something else. Sometimes melodic elements will double out of nowhere, like an audio version of those extra fingers in AI-generated images.
Sections"

 
Stability AI’s audio generator can now crank out 3 minute ‘songs’ based on a text prompt or sample

"The songs can actually sound nifty, at first, until the seams start showing. Then things get a bit creepy.

For instance, the system loves adding vocals, but not in any known human language. I guess it’s in whatever language that makes up the text in AI-generated images. The vocals sort of sound like actual people, and other times they sound Gregorian chanters filtered through outer space. It’s right smack dab in the middle of that uncanny valley. The Verge called the vocals “soulless and weird," comparing them to whale sounds. That tracks.

Stable Audio 2.0 makes the same weird little mistakes that all of these systems make, no matter the output type. Parts can vanish into thin air, replaced with something else. Sometimes melodic elements will double out of nowhere, like an audio version of those extra fingers in AI-generated images.
Sections"

Stable Audio is junk: it won’t follow the given text, generating a random soup of cacophonic sounds, also it wont generate any lirycs. Best AI remains Suno.ai for me, the only downside is the low quality encoding
 
Hi Forum,

here in Germany there is a word called "Kaufhaus-Musik" literally translated "department store music" that is the acme for "bad music".

This is the reason behind


to save money there is a market for inexpensive clones of well known titles from famous musicians to save what you have to pay to the GEMA organization

for this AI created music may be a solution, cheap and not so bad as the "cheap clones" for "department store music"

So far - so good, Stefano
 
I agree.

As a musician (pianist, singer, and woodwind player), I find it sad / disheartening, but I believe we'll be hearing full symphonies performed 100% in the digital realm (i.e., NO players, just sampled sounds indistinguishable from "the real thing") a LOT quicker than "2074". Same for solo instruments / soloists - a "perfect" violinist playing a "perfect" Guarneri or Stradivarius with a "perfect orchestra" that puts the Royal Concertgebouw to shame, etc., is not far away.
All that will be left to do will be to replace the listeners with non-biological ones, and the transformation will be complete. No humans in the entire loop from creator to listener. Complete the square, factor it, extract the roots. Done!
 
All that will be left to do will be to replace the listeners with non-biological ones, and the transformation will be complete. No humans in the entire loop from creator to listener. Complete the square, factor it, extract the roots. Done!
2024 might be a good time for folks to consider re-reading Ray Bradbury's story/chapter of The Martian Chronicles called There Will Come Soft Rains.
When humanity's all gone, no one will miss us. :(

 
It quite possible that there is no one else to miss us chickens. Bedawk!
 
So many things to unpack about what is happening and what will happen and what its impact will be...

I've played with Suno lately, on a month's subscription. I don't have it generate lyrics, I supply mine—from songs I've already written and recorded—and it's fascinating to hear how it "imagines" them. Both in different styles and in different takes.

"Like" this post if you want me to post a few highlights. I'm hesitant, as my words are involved, but if enough people want to know I could be swayed that it's helpful.

The one thing I feel is that when I go to a store and hear music playing overhead, I think, "This could be replaced, right now, with AI generated songs churned out ad infinitum, an no one would know the difference." Of course, that's partly because stores play pretty generic and non-challenging stuff that known to be likable. And AI is good at that.

But some of what it generates is pretty compelling. Sometime the whole thing (it kicks out two-minute pieces—you can have it extend them later) is interesting, but sometimes it's just interesting how it treated a certain phrase, or the chorus.

Personally, I've like it to sing my song over my music, and see how it phrases the lyrics that I might not have thought of.

I've also worked this week with ACE Studio, taking my singing and breaking it down into lyrics, notes, and inflections. I'm not interested in replacing my singing, which is what it's focussed on, but it's interesting nonetheless. Maybe though, I might do something like add a female harmony signing along with me on one of my songs.

But even if you don't like the idea of AI generation, where would or could you draw the line? Personally, I'd love having an AI producer or vocal coach that could listen to my song and say, "Here are a couple of other ways you could treat the chorus", and have it generate the alternate versions for me to evaluate. Or, "Have you considered not bringing in the drums until the pre-chorus build?"

But then, the next step is generating tracks, individual instruments as needed. I'm not a drummer, but enough that I can either play my V-drums kit and edit manually, or mouse in the beat. Or, I can send it to my friend to play the drums however he wants, which has the benefit of getting a different point of view involved. I could do it similarly on Fiverr. But at that point, why not give the song to an AI drummer, and either request a style, have it infer style, or just have it give 2-3 variations to choose from? I think drums are low hanging fruit for AI, with no pitch and harmonies involved. (I should check the web, I'd be surprised if this isn't already a thing.) Plugins and even DAWs like Logic have had "intelligent" virtual dummers for years, but with AI it could listen to the music it's drumming to and make better choices.

BTW, think of the serendipity that could come from being able to have AI audition your song in different styles? "Roxanne" was conceived by Sting in a bossanova style, Copeland suggested tango, Summers played it in straight four and Copeland switched it up to a reggae fell and Sting changed where the bass notes landed, accordingly. The Police were a punk band (that detested much of the punk habits) till then—would they have had as much success if they hadn't hit upon something else?

There are so many level this could be taken. Frankly, popular music is already dominated by writers who know the formula for mass approval and sell their services to the elite. If you say yes, that why it all sucks, I agree. But that's why I don't think AI is the doom of the music industry. Yes, it will impact earnings capabilities by musicians. But realistically, it's just an extension of what we have now, where musics complain they aren't paid fairly for their music. On streaming services. And a large part of that is there is so much competition, because this isn't the '70s, where you had to be signed by a big-money record label. Today, anyone can create and publish with a laptop. I think it's an unrealistic expectation, today, to expect a good living from making music.

It will also be an enabler. I write songs I want to give away—no pretensions on making money at my age. But for the music I write, I'd want to make them available as music videos. I don't have the budget to do that for free. But with the assistance of AI for the video, maybe I can.

But maybe...gee, a few years down the road when fewer humans can play anymore...maybe the local bar bands with be living the highlife, when people value real live humans playing in front of them more than what they can stream. :p
 
Stuff like Musicmaker and similar tools have made it possible to create pleasing music background loops for a long time. And such tools have been used in music production far longer than people called AI. Is Autotune voice correction an AI tool? Of course it is, but one I could live without and also one that shows art can be fraudulent... but also shows consumers can be stupid and undiscerning.
 
Stuff like Musicmaker and similar tools have made it possible to create pleasing music background loops for a long time. And such tools have been used in music production far longer than people called AI. Is Autotune voice correction an AI tool? Of course it is, but one I could live without and also one that shows art can be fraudulent... but also shows consumers can be stupid and undiscerning.
Well, I'd say those things are a little different from AI. Autotune is procedural, someone coded what it should do. An AI voice is one that is shown a singer's voice and singing, then, "now you do it—sing these notes using these words." And an AI composer is shown examples of country music until it understands what makes the genre unique, then "now, make a country song."

But yes, all tools—and where (and should) we drawn the line at what a tool is allowed to do? Paradoxically, our tools are what define us as humans.

Interesting times, because some of these things are good now, but we know that in two years today's versions will seem so primitive. Most music software up until now largely recreated what we already had. The job of a DAW is to recreate a recording studio—multitrack recorder, a multichannel mixer, and effects like EQ, reverb, and compression. Right now I see things with different specializations in AI (singers, video generators, etc.), because there is no single target, unlike the replacement of a recording studio. There will be different developers taking different paths.
 
I'll believe it when i'll hear it.
So far i've been more impressed with Big data than by what's called AI, chat gpt for instance is a powerfull tool but is really useless when it's about inovating.
Maybe I'm wrong but the source of great creativity has been proven to be a wide array of different interests and curiosity, you won't make a good musical AI if it doesn't catch the connection between music and... well everything else, all we know and sense, the human condition.
Basically I think it won't match humans as long as it hasn't become a life form.
Then I'll be impressed, until then it will just mimic, it will just be an '' à la '' AI, and will be of very limited interest to me.
I'll do a parallel with the many examples of artists that have been copying the style of the greatest geniuses of art, without ever managing to get that je ne sais quoi.
But hopefully I'm wrong, would be great to see that in myy lifetime, but so far AI hasn't been the revolution that was sold to me.
 
I want AI to do my job so i can have time to express myself creatively and do the things i actually want to do.

outsourcing journalism, music, movies, books and even now videogames to AI is just horrible and disgusting.

you can't even find a single search result on Google now that is not made by AI with zero backing or testing or human effort of any kind. Written internet is already ruined.
 
Back
Top Bottom