LALAL.AI Makes Stem Isolation Easier Than Ever

New Online Audio Service Uses AI to Get Unprecedented Stem Separation Quality

LALAL.AI introduces a new and improved way to extract vocals and instruments from audio and video sources. The service uses a machine-learning algorithm to precisely identify and carefully separate stems. It utilizes the power of AI to improve the lives of millions of people — DJs, producers, vocalists, dancers, karaoke lovers.

Artificial intelligence has previously been used as a solution for the complex task of vocal track isolation, but LALAL.AI is the first service to surpass even Spleeter by Deezer and outdo PhonicMind in accessibility and quality of results. We've compared all available audio splitting methods and proved that ours is the best.

The service first operated on Rocknet, a neural network created by the LALAL.AI team based on 20TB of training data to isolate instrumentals and voice tracks from songs. Now LALAL.AI uses Cassiopeia, a unique neural network superior to Rocknet as well as rival solutions, providing more precise and clean results with significantly fewer sound artifacts.

In addition to vocal and instrumental stems, LALAL.AI gradually gets more and more musical instruments in the list of the supported stem separation types. In 2021, the service has become the world’s first 8-stem splitter capable of extracting voice, accompaniment, drums, piano, bass, acoustic guitar, electric guitar, and synthesizer.

LALAL.AI helps developers and business owners to expand the range of their services by offering them server solutions. High-quality isolation of voice and instruments can be installed on a client’s own server or utilized via LALAL.AI’s infrastructure.

The LALAL.AI team is planning to continue creating new opportunities for sound professionals and creative people around the world while keeping the bar high for stem separation quality.

Don’t just trust our word — see for yourself! Compare the quality of isolated tracks extracted with LALAL.AI and other services.

Examples of split songs:

Let's Call it by Lawrence Trailer