I've noticed the recent rise in AI-generated music stemming from a text prompt. The rise in the quality of generated music has been startling to say the least.
Another genre/category of music is the pop mashup [1]. It's a song that's created from snippets of existing songs blended into a single, cohesive track. Sort of a like a super-condensed 3-minute DJ mix that comprised of up to dozens of song snippets.
Initial thoughts on feasibility include AI not being able to reproduce pop songs verbatim due to copyright. To create a mashup in the style that we know today, it has to be the original song, and not say "a song in the likeness of Madonna".
What other hurdles would need to be overcome to make the generative-AI mashup a reality?
[1] Madeon - Pop Culture: https://redirect.invidious.io/watch?v=lTx3G6h2xyA
0 comments
—
Ask HN: What would it take for AI to generate pop song mashups? · HackerTrans
Another genre/category of music is the pop mashup [1]. It's a song that's created from snippets of existing songs blended into a single, cohesive track. Sort of a like a super-condensed 3-minute DJ mix that comprised of up to dozens of song snippets.
Initial thoughts on feasibility include AI not being able to reproduce pop songs verbatim due to copyright. To create a mashup in the style that we know today, it has to be the original song, and not say "a song in the likeness of Madonna".
What other hurdles would need to be overcome to make the generative-AI mashup a reality?
[1] Madeon - Pop Culture: https://redirect.invidious.io/watch?v=lTx3G6h2xyA