The advent of AI music

Sunday, October 20, 2024 - 12:36

I think it's time to broach the topic of artificial intelligence being used in music production. New technologies are emerging that allow people to generate entire musical tracks using nothing more than a simple prompt and the click of a button. Although one might doubt the power of this seemingly simple process, modern AI is actually capable of creating very compelling music. I believe there are many issues to discuss surrounding this new technology.

First, many people have invested an enormous amount of time in honing their craft to produce music themselves, and this technology undeniably invades their space, threatening their market share. There’s also the issue that AI has been trained on a vast array of music and styles. At what point does this become copyright infringement? Where do we draw the line between tools like drum machines and arpeggiators, and AI processes that potentially create derivative works?

Another pressing question is that of artistic originality. Should we allow someone to claim ownership and receive praise for music that was created with the help of an algorithm? We've witnessed the rise of prompt writers who are capable of coaxing impressive art from machines, but is this truly an art form that should be celebrated? I see a parallel here with the advent of photography and how it affected painted art. In the end, we found a way to celebrate both forms of expression.

The application of AI in music production also signals broader advancements in the capabilities of AI. We’re inching closer to an AI-run world. Is this something we should embrace or fear? How do we engage with this technology in a responsible manner? I believe there’s no way to put the genie back in the bottle, so the question remains: what should our attitude be toward AI as it continues to shape the future?

MedicineStorm

joined 12 years 3 months ago

Sunday, October 20, 2024 - 15:24

Not sure this qualifies as broaching the topic, since there are already several discussions about it already present on the site (one example: https://opengameart.org/content/artificial-intelligence-assisted-artwork ) but I do agree it should be discussed.

--Medicine Storm

glitchart

joined 2 years 9 months ago

Monday, October 21, 2024 - 06:32

One of the things I´m interested on is the topic of generative vocals, the reason for that is because in the music field of generative composition I've never achieved anything remarkable (all I rendered sounded too random and useless to me) ... not that is not possible I probably did not invest enough time, but do you know you can actually add vocals to music tracks with such a realistic quality that it can almost be called "production ready"? Not only that, you can add lyrics and use any language as long as the trained cloned voice match the language (otherwise pronunciation sounds bad).

To give more details just in case anyone has more idea than I, the software with which I had more progress is called Synthesizer V from a company called dreamtonics: https://dreamtonics.com/synthesizerv/, https://svdocs.dreamtonics.com/en/home, they also have a forum here:https://forum.dreamtonics.com/c/synthesizer-v-studio/5

I hope it's ok to post that there, I don't aim to make publicity or anything, I just want to know if anyone has idea of vocals created with such a software can be licensed as creative commons provided that the used vocals are open.

* Update:

I found this f.a.q. which clarifies a few aspects of the license but no single word about the non commercial use of the software, in other words: there are only restrictions for the commercial use of it: https://vi-control.net/community/threads/faq-what-is-synthesizerv.136072/

The EULA says "Except as otherwise specified, you may use the synthesized sound generated by using this product, whether commercial or non-commercial, free of charge." https://www.ah-soft.com/synth-v/eula_e.html

This makes me wonder, is it possible to license a song as OGA or creative commons using the free voices?

* Update 2: I finally have an answer that seems correct to me:

what about license CC BY-NC which states that Only noncommercial uses of the work are permitted? can a work rendered with SynthesizerV be licenses as CC BY-NC or is there anything in the EULA that prevents it? if not, which creative commons license fit with the EULA if any? Someone answered me the following: "IANAL, but it looks like either BY-NC or BY-NC-SA will work, assuming you are complying with the EULA yourself by correctly attributing the Lite voice."

According to what chatGPT answered you can the combine that license like this:

If the music is under CC BY-NC-SA, the same conditions apply. The resultant work remains CC BY-NC-SA.
If the music is under CC BY-NC-ND, you can still mix it, but the resulting work must be licensed under CC BY-NC-ND, which restricts derivative works.
If the music is under CC BY (Attribution), you cannot legally mix it with CC BY-NC-SA vocals for a resultant work without breaching the non-commercial clause, unless you adhere to the non-commercial aspect of the CC BY-NC-SA license.
If the music is under a proprietary or commercial license, you would need explicit permission from the copyright holder to combine it with the CC BY-NC-SA vocals.

CC BY-NC is not listed and not wanted here, so no need to discuss wether the use of such software for submitted music is wanted here: it is not and I understand perfectly why, I'm not complaining.

Even if it is completely legal, is not my intention to submit anything that makes use of A.I. if it is not wanted for any reason including the legal uncertain it brings about as it is a new frontier and laws are still to be made about it (I'm not a lawyer and I see it only from a creative standpoint and I don't intend to replace my workflow with it.

I'll be glad to read your opinion in this regards and I'm also opened to know what you think of using such software to incorporate it in a creative workflow, even if it is negative or you have anything to say about this I'll be glad to know.