That and having Ai recompose my work in the styles of others or vice versa is something I would like to try.

That is probably a good explanation. I assume they also think that there is plenty enough of music in the catalogs to used cheaply. But as Ai gets more developed, easier and cheaper to adopt by none Ai companies NI, Steinberg, Arturia or some newcomer will pick up the ball and current DAW interfaces will be a thing for nostalgics that think it has a value in it self to click in notes on a grid rather than just lazily think them in.

yes. that was a long time ago. people were rigging the payout system. but spotify caught on and fixed it. the same with releasing your own song and having it play on auto play 24/7 all year long. ;-)

When I wrote "you need help" I meant "you need help with composing", not that you need help in general as a human being.

Good luck convincing the AI that you're right and it's wrong. Let us know how it goes.

Depends upon the definition of a long time ago, but I heard an interview in music radio this summer with an artist that made "soothing baby sounds for Spotify". He said in the interview that it was very lucrative, but he did it anonymously cause he knew from others that it would stain his brand producing other kinds of music if it came out. I have not looked at it but I assume the genre "soothing sounds for babies" are still played.

Seems like you are stuck on the idea that an Ai is an exact copy of a single human beeing that also has a normal stubborn approach to only help doing what it thinks is right.

Sorry to disappoint, but current Ais don not come with the human drama included. The are just mathematical models of human learning. They interpret the dataset they are given and when you ask them they just deliver what you ask for if they have it in their dataset and if it is not restricted by some legal or moral parameters in their code.

But I am sure somebody could make an arguments Ai in the spirit of Monthy Python as a service for all that want to have somebody to argue with.

Authors have already done so https://www.vulture.com/2023/09/geor...it-openai.html

But you shouldn´t expect too much of such campaigns. It´s like fighting cannons with swords back then. AI puts arts as well as the law into a "mirror cabinet" and all we talk about are just reflections.

AI doens´t have such such "bad habits" as many musicians do, so ...

What I was getting at was something different. There are things we can't do as artists and you as a composer are looking for help with some of those things. A collaboration with a human is something you don't want because as you implied earlier you want your vision to be fulfilled and not compromise. So what I was getting at was that the AI won't know what's in your head or how to get it done so it too will spit out things you will reject. That was really the only point I was making there - that AI won't perfectly fulfill your vision, just as a human collaborator won't (according to you).

Well that would be the next natural step for you I'm sure.

Its now a race to the bottom with no guardrails...
https://www.theguardian.com/technolo...ef=futurepedia

The difference is that an Ai will make it's best effort to 'recompose this song A in the style of Giorgio Moroders sound in G Locrian' rather than being "human" and arguing that it will probably sound shit and not be worth to even try.

And after doing that I will know in minutes if it was a crap idea or not. And further it will add to it's dataset so when somebody else ask it 'Do you think it will sound good recompose this song A in the style of Giorgio Moroders sound in G Locrian' it may answer. I was requested to do that by another person. He responded that it was no good. I can give that piece to you and you can decide if you share his opinion. So it will move things forward more.

I have no need for one. If I ever need that I can just post here as long as you are here

well, as they say.. bottoms up!!

Seems relevant. I don't think AI is going to level the playing field in the way that some do, and those who've struggled through the old ways of learning will likely still have a huge advantage over those who haven't gone through the process of learning difficult skills over time.

Those people that do things the old way will have an advantage until they are outnumbered by all those using technology.

I’ve seen it happen to where people who were not used to using computers have been shut out of the workforce for most jobs.

This tech is growing and will get more advanced. It will become the norm and the younger generations are not going to go back to “the old school ways” unless they have no other choice.

I should have been more clear - in the short term, I mean people who learned the old school ways & have the ability for deep focus will have a great advantage when they also use AI tools for their own benefits compared to people who only know how to use the new AI tools but lack deeper understanding of other subjects, focus and other critical thinking skills that seem to be waning. I’m encouraging everyone to experiment with these tools asap & figure out ways to use them in your work when they benefit, which they might, no matter your age or area of expertise.

The sad part of this is that Ai is not a simple tool for humans to use. It is a student with superior capacity. Ai is learning from what we put into it and it will learn to do the same. So if it is used by the very best it will also be trained by the very best and equal that level and produce stuff to that level a lot faster than any human. Just looking at the stuff Suno and Udio put out now it is totally good enough for most commercial purposes and it never asks it customer why he want that same booring pop song with that pathetic sales blurb as a text. The genie is out of the bottle and will shake the foundations of the music industry. The sad part is that I feel there will be a whole lot less opportunities for a whole lot of artists in a few years.

So I am very torn here. I am curious and want to use Ai tools for my personal projects, but I know that by doing that I also hand over my own "signature" to Ai to train upon. But I am also totally aware that what I do will have zero impact cause others will for sure use it and Ai will get it that way. I guess the bottom line is if it feels better to stand by and watch the progression or be a part of it?

I'm late to the discussion, but have had fun reading the thread up to this point.

I just wanted to point out some ways in which AI has already creeped into the work that I do. I'm not a composer, but I am a sound editor/designer and mixer. I work mostly on commercials these days, but I have worked on a lot of documentaries and reality TV in the past.

The good things I've seen with AI are the noise reduction tools for dialogue. These have made a huge improvement over the last few years and now dialogue that was previously unsalvageable can be saved.

The not-so-great part of that is that I have clients that insist on using tools like Adobe Enhance, which can do amazing feats of noise reduction, but also suck the life out of a lot of recordings-- and now the amount of time the I am hired for is reduced because the "dialogue has already been cleaned".

In the "I'm not so sure if this is a good thing or not" category are the AI voice generators. I've had clients use them to do scratch tracks of voiceovers; then once the client has approved the script the Prod co hires a real VO talent and records the final script. That saves some back and forth for the client, but the VO talent has lost out on any revenue from pickup sessions. How long until the production company just skips hiring the VO talent altogether? Probably not long. Many Youtube videos already do this. The reads are still pretty sterile at this point, but I'm sure they will get better, just as the image generators have.

I've also had clients replace words within a spot with an AI generated voice of the same person when they've had to make a revision to the spot for legal or creative reasons. I guess I'm ok with that assuming that they have the permission of the person being imitated, but I can easily see a scenario where this technology can be used maliciously.

More on topic to this thread are the AI sound effect and music generators that are coming out like Eleven Labs and Loudly.com. They create music or sound effects based on prompts. Again, I think this generation of tools are still a bit weak, but I can see them getting better and certainly good enough for lower budget productions in the near future.

I've been lucky enough to make a living doing what I (usually) enjoy and have had a decently long run at it. I've had some fairly high profile clients over the years, but the bread and butter of my work, and probably that of 95% of the working audio engineers and composers, etc., has been for lower budget projects. These are exactly the kind of projects where AI will do the most damage-- the projects where the budget is only enough for library music, or they can only afford one VO session with the talent. These are the clients that will certainly opt for the AI generated music that has the same wallpaper feel as library music, or a sterile VO that gets the point across and costs nothing.

I can see the same thing happening in the video editing world. Can AI create a narrative film? Not yet, but it can probably create a sizzle reel of beauty shots, or create a string-out of happy people going about their day (think of a commercial for some prescription drug, for example). It might not be good enough quite yet today, but I bet it will be next year. Those are all shots that were, up until this point, scouted, casted, acted, lit, filmed, Color Corrected, etc., even if they ended up in a stock footage library somewhere. Those too are all jobs that are at risk.

So maybe things will shift and jobs that we haven't thought of yet will pop up somewhere unexpected. I don't want to be a complete pessimist, but I am glad that I am nearing the end of my working life though because I don't know if my revenue streams will exist in the not so distant future.

it might become like woodworkers who only or mostly use hand tools and old techniques one day the way we are currently composing music, I am fine with that, don't mind 'getting left behind' because nobody can actually compete with me, including even the best AI

I’ve seen similar and share your perspective. For adobe enhance, I’m sure you do but for others, I would compromise and ask them to at least keep the original in the sequence. They don’t mind keeping 8 channel .wavs that only have three channels of audio.

For scripted or VO, I think the talent should still be getting paid for pickups and agree to have their voice “trained”. I guess that’s obvious. For unscripted, I think it’s the best tool for fixing creative edits. No different than searching for syllables through the shoot, which AI tools have made faster. But really fast using something like eleven labs. Still only replacing the bare minimum to make them sound more ”human”. The irony. I wouldn’t be comfortable changing their intent in the interview by either means. Of course still need permission. Though I can’t imagine that won’t just be assumed or included in contracts at some point.

The sound effects models I’ve tried have been pretty weak. Might be what they are being trained on. I could see soundfx libraries building micro ai models, trained off their recordings, but in a more focused category. Kind of like the great micro libraries that have come out the last 10 years.
Music - yeah, the race to the bottom has been going on for a while with cheap royalty free sites with way too good music for the price. Though well crafted bespoke music will stay around for people who still care about it. Also all of the legal consequences still haven’t been worked out for any of this stuff.

Adobe Enhance is an interesting tool. I used it extensively at my last job to save a bunch of terrible recordings made by non-engineers during Covid when production was shut down. Sometimes it really was magic, but it seems to "homogenize" the voice a little, removing some of the character so it doesn't sound exactly like the same person. I think part of it isn't just noise reduction, but rather a transcription that is then synthesized into a new voice, at least that's how it sounded to me.

I work in video production as well and think AI is coming for the jobs you mentioned in the near future, with narrative editing probably not that far off either...