A lot of musicians are upset at the moment because they are worried their music has been ripped without any money changing hands to them. Interesting times
@@musictechinfo They certainly should be. The cat is out of the bag, the music industry is about to change a lot very rapidly. Unfortunately the laws have not kept up with technology
@musictechinfo but it would be great if Udio allows us to generate a new part from a certain point like Suno. I lost so many cool moments by some weird shit in the end xD And I like that Suno generates new parts separately and only then you can combine them. Because in Udio, REMIX can be a powerful tool... but only with the first part of the song. But if to look at both tools from a professional standpoint, and if compare to something like Stable Diffusion prompting and CompfyUI - they poop badly. Because, I think, that music is more complex than image and Midjourney style UI is not for it at all.
I agree we need more control with all these tools. I want to be able to do duets with different ai vocalists, and reference tracks, and export stems, and export midi my list goes on 😂
Ai music is dangerous but also cool because for those of us who let’s say wanted to use a copyrighted song on a video but the label would strike us or steal all the rev, now you can basically make weird covers in strange genres of that song that sound slightly like the song and evade copyright issues, it’s a huge threat to the copyright system in general
Yeah of course it scraped. That is how mashine learning is about. 😅 That is the same principle that a search engine is based on. And AI systems like Image to text (for accessability function reasons) that are out there for more than decates now. 🤷🏻♂️ Would mean bscly. Google would be illegal 😃 But you must also take in consideration what scraping means. There is no permanent download (on what possibility of hard disc space for an AI model would that function btw 😃). Thede databases are more like hyperlink-archives of stuff in the web. And the analysis is on a level your webbrowser deals with data. No way to win a copyright case on this, despite we illegalize the fundamental way the Internet works today.
LOL quote : "- it is his voice". Sounds like most auto tuned "singers", IMHO.
Of course it had to use copyrighted music, these programs train on all music data available and I'm fine with that
A lot of musicians are upset at the moment because they are worried their music has been ripped without any money changing hands to them. Interesting times
@@musictechinfo They certainly should be. The cat is out of the bag, the music industry is about to change a lot very rapidly. Unfortunately the laws have not kept up with technology
You can extend the song beyond 30secs. You can also make a longer song if you write the lyrics first.
Yes very useful I love the fact you can generate backwards too something that Suno needs to do fast 💨
@musictechinfo but it would be great if Udio allows us to generate a new part from a certain point like Suno. I lost so many cool moments by some weird shit in the end xD
And I like that Suno generates new parts separately and only then you can combine them. Because in Udio, REMIX can be a powerful tool... but only with the first part of the song.
But if to look at both tools from a professional standpoint, and if compare to something like Stable Diffusion prompting and CompfyUI - they poop badly. Because, I think, that music is more complex than image and Midjourney style UI is not for it at all.
I agree we need more control with all these tools. I want to be able to do duets with different ai vocalists, and reference tracks, and export stems, and export midi my list goes on 😂
Ai music is dangerous but also cool because for those of us who let’s say wanted to use a copyrighted song on a video but the label would strike us or steal all the rev, now you can basically make weird covers in strange genres of that song that sound slightly like the song and evade copyright issues, it’s a huge threat to the copyright system in general
The possibilities are endless
idk man i get better results from suno as of now
It’s a bit hit and miss for both from my side
Yeah of course it scraped. That is how mashine learning is about. 😅 That is the same principle that a search engine is based on. And AI systems like Image to text (for accessability function reasons) that are out there for more than decates now. 🤷🏻♂️
Would mean bscly. Google would be illegal 😃
But you must also take in consideration what scraping means. There is no permanent download (on what possibility of hard disc space for an AI model would that function btw 😃). Thede databases are more like hyperlink-archives of stuff in the web. And the analysis is on a level your webbrowser deals with data.
No way to win a copyright case on this, despite we illegalize the fundamental way the Internet works today.
Thanks, interesting times