TAXI.com

I find it interesting that libraries also seem to want to stay away from fully cleared AI-generated vocals as e.g. provided by ACE. I understand that one can even download individual licences to send along with any submissions. Is that not essentially WFH? I mean, the artists behind the voice models have agreed that it can be used for commercial proposes with no strings attached. Why is that not „safe“ enough?

CTWF wrote: ↑

Sun Dec 01, 2024 6:22 am

I find it interesting that libraries also seem to want to stay away from fully cleared AI-generated vocals as e.g. provided by ACE. I understand that one can even download individual licences to send along with any submissions. Is that not essentially WFH? I mean, the artists behind the voice models have agreed that it can be used for commercial proposes with no strings attached. Why is that not „safe“ enough?

My guess is it's because the legal waters are still too murky for them to take any risks. And why should they? It's not like there is a shortage of music out there performed by live artists. Libraries are very saturated with great music.

Another thing to think about, unrelated to your question, is that when you hire a "real" WFH" vocalist, you are supporting your fellow musicians, keeping their income stream alive. For now, while I can, I'm happy to do that.

Best,
Casey

Casey H wrote: ↑

Sun Dec 01, 2024 8:33 am

…

Hey Casey, what you wrote were my guesses for the reasons. Maybe even libraries think of the artists, some being run by artists. Nonetheless, the cat is out of the bag and at least with properly cleared AI clones there should be no legal issues.

Tom/CTWF

CTWF wrote: ↑

Sun Dec 01, 2024 10:24 am

Casey H wrote: ↑

Sun Dec 01, 2024 8:33 am

…

at least with properly cleared AI clones there should be no legal issues.

We are not qualified to say that. What a non-lawyer considers "properly cleared" may not be what their lawyers say is not. Also, keep in mind that the last thing a library wants is hundreds and hundreds or even thousands of tracks with the same "properly cleared" voice.

For instance, ACE Studio writes that with each use the artist behind their AI models earn money. So, it should also be appreciated that this kind of software can open up new streams of revenue for vocal artists.

Another problem for many producers is that the cost of vocalists may be prohibitive and it especially can be a problem if vocals are needed at an early and uncertain stage of the production of a song.

CTWF wrote: ↑

Sun Dec 01, 2024 1:55 pm

For instance, ACE Studio writes that with each use the artist behind their AI models earn money. So, it should also be appreciated that this kind of software can open up new streams of revenue for vocal artists.

Another problem for many producers is that the cost of vocalists may be prohibitive and it especially can be a problem if vocals are needed at an early and uncertain stage of the production of a song.

Interesting thread. A local singer/songwriter friend of mine has been busting out songs for an upcoming album that heavily uses an A.I. vocalist - "she" sounds really good;
you know: "contemporary and compelling". I like the name my friend gives "her" in the credits: Joy Aiken. See what he did there?
He claims it's totally legit for him to sell all that stuff on all the platforms as his own.

Our opinions really don't matter. Ask a few libraries and distributors. If they won't accept these songs, it's all moot. You can say "should" till the cows come home. moooo.
Casey

funsongs wrote: ↑

Sun Dec 01, 2024 2:04 pm

He claims it's totally legit for him to sell all that stuff on all the platforms as his own.

Yes, this is my understanding, too. Given this, it possibly won’t be all that long before the first libraries accept AI vocals. If the stuff is played on the radio already anyway …

CTWF wrote: ↑

Wed Dec 04, 2024 9:21 am

..... it possibly won’t be all that long before the first libraries accept AI vocals.

Potential problems include:
- The competition from other producers who can't sing or play (or produce) would increase massively, without paying singers and players.
- Libraries would be able to generate their own content without paying producers.
- TV companies would be able to generate their own content without paying libraries or PROs
- AI companies would be able to saturate the market with their content without paying anyone (other than lawyers).
- Also, a zillion kids from all over the world are growing up with AI and will soon be able to generate whatever audio and video content they want anyway, with no need for TV companies or libraries, or producers or PROs.

Against that backdrop, I don't think that having a few AI generated songs/vocals will be much currency. Neither will anything that I produce, but at least I'll have had fun making it and will be proud of it being my own creation using 100% HI .

I've had a lot of luck with using AI vocals to replace my crappy vocals, then I can use this as a demo on AirGigs to have an actual vocalist do a vocal. Prices range from 100 to 300, but there are some excellent singers on there. My favorite has been Sean Killeen. I'd post an example but that feels like shameless self-promotion.