When I make audio or video content, an element with which I always fight is music. I do not mean a complete musical song or anything, but only a basic introduction and outro songs to relieve people at the beginning or at the end of simply start talking abruptly.
There are many good music libraries, but with the growth of many musical tools of AI, I thought it would be fun to see if they were better than me to compose an appropriate musical motive.
To try it correctly, I decided to quickly make a ‘podcast’ using another AI tool that had been playing, Google’s notebook. The general audio description function can convert documents, video transcripts and other sources of information in podcast style recordings generated by the between two AI hosts. By whim, I chose the glass blowing as a topic just because I am interested.
I gave him a handful of links and videos about the history of glass support and how it works. Soon, two voices of AI discussed him for more than 20 minutes. But for this, I only needed about ten seconds.
Suno serves
I tried some different musical tools, including Soundverse, Blessed and out loud. They all had their moments, but most did not break it. I tried short, detailed longer indications and even only keywords. Above all, they were fine, but often they were discordant or simply uncomfortable to listen.
After too long trying different indications and editions, I landed to ask Suno to make an instrumental track that works as an introduction to a podcast on glass support, as simple as that.
Suno produced two clues with the evocative title “Forming Fire and Sand”. One was fine, but the other had the perfect type of environmental and light tone that could easily imagine being heard before a nerd discussion about the annealing and the purity of the sand.
When I sewed everything, it worked, or at least it was not distracted from incoming speech. I executed it for some friends, and none of them noticed that the music was generated by AI, although they detected the real podcast as ai voices.
You can listen to how it was under.
I will not pretend it was perfect. I had to put fading manually, and if music were human musicians, you might think they were more excited than talented. Even so, Suno did a great job for a free tool without required complex production.
I don’t think I would choose it automatically about something from a human composer, even in a free song library. Personalization available with AI cannot cancel human creativity in most cases; The podcast ai demonstrates it. But, as an experiment, Suno made a harmonious addition.
