The program, dubbed “neural karaoke,” was created by researchers from the University of Toronto and can take any digital photo and transform it into a computer-generated singalong. The program’s neural network has been trained with nearly 100 hours of online music to generate tones and notes that are often paired together. The program can take a musical scale and melodic profile, and produce a simple 120-beats-per-minute melody. It then adds chords and drums.
How does the first AI-generated Christmas carol sound?
Answer: not great
Step aside Michael Bublé and Mariah Carey, there may be a new sultan of swing when it comes to Christmas music — and it isn’t a human. A new, albeit tone-deaf, Christmas carol has been created using an artificial intelligence program that analyzed a picture of a Christmas tree and presents, and rendered words based on the visual input. Take a listen, if you dare:
The program, dubbed “neural karaoke,” was created by researchers from the University of Toronto and can take any digital photo and transform it into a computer-generated singalong. The program’s neural network has been trained with nearly 100 hours of online music to generate tones and notes that are often paired together. The program can take a musical scale and melodic profile, and produce a simple 120-beats-per-minute melody. It then adds chords and drums.
The program, dubbed “neural karaoke,” was created by researchers from the University of Toronto and can take any digital photo and transform it into a computer-generated singalong. The program’s neural network has been trained with nearly 100 hours of online music to generate tones and notes that are often paired together. The program can take a musical scale and melodic profile, and produce a simple 120-beats-per-minute melody. It then adds chords and drums.
