A.I. For Music Industry

 

 

BETTER SOUND 2019 | Category: 6. Services and Sound Start-ups

 

Entrant

Musiio

 

Abstract

Musiio develops A.I. solutions to assist production music (sync) companies, streaming services and record labels with their curation needs. Through the use of machine learning, our tech is able to listen to large volumes of music in a short period of time, thus increasing efficiency and accuracy in providing a personalised offering.
 

Description

Initial Outset
With more than 14 years of experience in music technology, UK-born Hazel Savage and Swedish data scientist Aron Pettersson met in a start-up incubator programme in Singapore. Hazel has worked in music companies such as Shazam in London, Pandora and Universal Music in Australia and Aron has been writing code for more than 17 years. Combining Hazel’s expertise in music and Aron’s expertise in machine learning and neural networks, they started brainstorming for inefficiencies in the music industry that could be solved with the use of Artificial Intelligence.

Idea Concept
With the invention of the internet and the rise of resources intended to make it easier for music to be produced and recorded, there has been a massive boom in the amount of content being released into the world. Before 2007, there were approximately 10 to 30 songs being released monthly due to the high costs (such as rental costs of studios, equipment) involved in releasing music. In 2018, it is estimated that approximately 30,000 new tracks are uploaded onto Spotify daily. However, the role of curators and Artists & Repertoire (A&R) executives has not changed to this day and they are still manually listening to tracks to discover talent, or to find relevant tracks/audio for ‘sync’ companies.

Solution Process
Using Machine Learning, our product is able to listen to huge volumes of music in a short amount of time to find sonic matches, thus reducing the amount of time needed to create a fully personalised offering. By increasing efficiency, Musiio will in turn increase accuracy by effectively “listening” through tracks that may be buried deep in large databases.

Results
The AI will be able to recommend a list of tracks that it has deemed to be sonic matches to the curator’s search enquiry or fully automated using our API. By listening to the “seed” track fed to the AI by the curator (via a YouTube link, an MP3 file upload or using a recommended Musiio demo track), the AI is able to provide a list of recommendations that are sonic matches to the “seed” track. This increases efficiency and accuracy as the AI is able to listen to the database quickly to provide recommendations that are the best matches.