One of the things I love to do is to try and find interesting insights using our AI. This week, I asked our AI this question: “Can it hear when a child singer develops their adult voice?”
When Justin Bieber broke the internet with his angelic voice and incredible smile, he was just a little kid. Being an artist at such a young age, his journey is very unique and gone through many evolutions - notably, his voice has too. Using songs at four different stages of his career: (i) 14-15, (ii) 17-18, (iii), 20-21 and (iv) 25-26, I wanted to see if technology could spot the change in his vocal capabilities. Here are the results…
Our AI tagged his tracks when he was 14-15 years old as “female”, understandably because his voice had yet to break. There was then a little bit of an in-between stage at the age of 17, where our AI tags his vocals as “mixed” - which simply means that his vocals showing both “female” and “male” characteristics. Actually, there are a number of artists who use this to their benefit by singing male and female parts of a song as if it was a duet - Both Prince and Michael Jackson would be great examples of this.
So why does this matter, or why is it useful? Well, if we think about it in terms of a music request; Perhaps the music request is something along the lines of "we're looking for an energetic, dancey, pop track with a female vocal lead". Music supervisors and catalogue owners would understandably either work from memory or naturally filter by gender - even though the track that's closest to what the client was looking for, from a sonic perspective, was Somebody to Love by Justin Bieber.
Using AI to listen to music without prejudice and without pre-conceived notions of gender, or genre for that matter means that music supervisors can uncover music better and find those hidden gems than ever before.
Interested in trying our AI search for free? Speak to us, or try our tagging demo here!