When I make audio or video content, an element I always struggle with is music. I don’t mean a complete theme song or anything, but just some basic introductory and outro melodies to ease people in the beginning or end instead of just suddenly starting talking.
There are lots of good music libraries out there, but with the growth of many AI music tools, I thought it would be fun to see if they were better than me to compose a suitable musical motif.
To test it correctly, I quickly decided to make a ‘podcast’ using another AI tool I had been playing with, Google’s notebooklm. The Audio Overview feature can transform documents, video transcripts and other sources of information for AI-generated podcasts between two AI hosts. On a whim, I chose glass that blew like a topic just because it interests me.
I gave notebooklm a handful of links and videos about the history of the glass blowing and how it works. Soon, two AI voices it for more than 20 minutes. But for this I only needed about ten seconds.
Suno serves
I tried a couple of different AI music tools, including Soundverse, Beatoven and loud. They all had their moments, but most of them just didn’t quite crack it. I tried short requests, longer detailed and even just keywords. Most often they were okay, but often they disagreed or simply unpleasant to listen to.
After far too long to try different requests and edits, I landed on asking Suno to make an instrumental track that would act as the introduction to a podcast about glass blowing- as simple as that.
Suno produced two numbers with the atmospheric title “mold fire and sand.” One was just okay, but the other had the perfect kind of surrounding and easy classic tone that I could easily imagine being heard in front of a nerdy discussion of smoothing and sand rehearsal.
When I sewed everything, it worked the kind – or at least it wasn’t distracting from the incoming speech. I drove it with a few friends and none of them noticed that the music was AI-generated, though they caught sight of the actual podcast as AI voices.
You can hear how it went below.
I don’t want to pretend it was perfect. I had to set the fading manually, and if the music was from human musicians, you might think they were more enthusiastic than talented. Still, Suno did a good job for a free tool without any complex production required.
I don’t think I would automatically choose it rather than something from a human composer, even in a free song library. The available adaptation with AI cannot override human creativity in most cases; AI podcast proves it. But as an experiment, Suno made a harmonious addition.
