AI Sound Separation

Separate and enhance mixed sound sources using AI

Explore AI Sound Separation projects

A man recording a video of himself on a cellphone outdoors

CASE 1

Windproof microphone with Xperia phones

AI Sound Separation is at the heart of an intelligent wind-filtering function that comes with Sony's Xperia 1 II. The wind filter uses AI technology to detect and isolate wind from the background sound, meaning that anyone with an Xperia 1 II can record videos outdoors without wind noise ruining the sound.

Blu ray box set of Lawrence of Arabia movie collection

CASE 2

Reviving classic movies and soundtracks

Classic movies Ghandi (1982) and Lawrence of Arabia (1962) have been remastered and given a new lease of life by modernizing and significantly improving their sound. Their original mono master soundtracks have been put through AI Sound Separation and analysed into different sound sources, such as dialog, shouts, horses galloping, and other sounds. The extracted tracks were then used to create an immersive Ultra 4K HD experience. This would not have been possible without AI sound separation. Watch example>>

Cell phone showing the LINE messenger service, with a microphone highlighted on the screen

CASE 3

New karaoke tracks in real time with LINE MUSIC

In Japan, LINE Corporation teamed up with Sony to create an innovative karaoke experience with their LINE MUSIC, that can take a recorded song and turn it into a karaoke-ready track in real time. It does this by applying the AI Sound Separation model, isolating and removing the vocal line. Normally, karaoke services are only available for a limited range of songs, and new tracks must be prepared in advance. Watch example>>

Composite image of school children on separate video screens singing

CASE 4

Bringing school kids together with AI

During the pandemic, most Japanese schoolchildren could not physically participate in afterschool activity clubs. To help make up for this, Sony Group Corporation organized a project Entame no Bukatsu (entertainment club activities), for kids to create their own music video. The idea was to combine the voices of all the children into one song, no matter where they were recording or on what platform. The project was made easier using AI Sound Separation to extract voices and combine them into one music track to create a music experience. Watch result >>

Semi-abstract closeup of a DJ's hand as it manipulates the controls of their equipment

CASE 5

Supporting the rights of music creators

Sound sampling is a common practise in popular music. However, artists can run into legal difficulties if they don’t get permission from the owner of the original recording they are sampling. Sony Music Entertainment Inc. has launched a service called Soundmain that aims to protect creators’ rights using blockchain technology and AI Sound Separation.

Interested in working with AI Sound Separation? Get in touch!

Submit your enquiry and we will get back to you as soon as possible.

Contact AI Sound Separation team