Musiio creates an algorithm to find new hits
When new hits come out or when some artists get great success it is also thanks to the experts and their work made of researches that can last months and months. Discovering new artists and new singles is a rather hard task because the characteristics of the music must match the people’s tastes of that specific moment.
An example is the story of manager Ryan Walter who, spending seven hours a day listening to new and unknown artists, managed to discover the singer Lewis Capaldi who became famous with his single “Someone you loved”. What if there was a way to simplify and speed up the long searches?
Musiio is a start-up based in Singapore, but its team is led by British and Swedish experts, that have created a new technology based on artificial intelligence. As Hazel Savage, CEO and founder of the company, said, the innovative system should be able to guess the song with the highest chance of reaching the top of the charts.
The algorithm created by Musiio will not only be able to classify the most recent music, but will also be able to define and estimate its potential, the goal is in fact to engage those who may be more successful than other singers. Musiio transforms audio files in order that the computer can read them, so it is possible to extract the sound characteristics and instruct artificial intelligence to find similar patterns. This process allows to recreate everything that human hearing is able to distinguish; talent, viral content, and quick identification of familiar melodies are the three most important aspects of an effective algorithm.
Musiio’s team chose about ten hits on YouTube, the data of these songs were then entered into a database with the data of other 10,000 songs from the Free Music Archive. After that, the innovative system of artificial intelligence had to rank the best twenty music tracks. In this way, the team of technicians carried out the verification to test the result of this ambitious project, which had a final success of 90%.
Musiio has thus developed a machine learning tool to identify the songs with the highest probability of climbing the charts and reaching number one, but as the company itself states, the algorithm was not created to replace the experts and their job, but, on the contrary, to give concrete support to their work making it faster and easier.