Tuesday, April 16, 2024


I fiddled around with Udio last week because of this Ars Technica article: New AI music generator Udio synthesizes realistic music on demand.

I'm particularly interested in AI-generated music because so much of music is artificial already. Algorithms, sampled beats, editing software, etc. There are entre genres of music now that have DJs instead of musicians. No one plays an instrument. It seems logical that AI is the next step, even for professional musicians.

I wrote lyrics for a short parody song for C and used Udio to see what it would sound like. In less than five minutes, I had four different versions, one of which I liked very much, and it was done. I'm not including the recording (for Cs privacy), but it sounded far better than I expected, and generating samples in different genres (I tried used the Beatles, Paul Simon, and Neil Young as keywords for styles) is fascinating. In the end, the recording I chose wasn't any of the above styles--it was Udio's own choice--but it sounded far better than the others.

There are ways to enhance what the AI generates, but I didn't need any of them. It's a powerful tool, and if you're curious, here's a link: Udio.

Site Meter