Split audio stems with LALAL.AI

Lalal.Ai Try it →Watch full video →Sound Codex · Oct 2025

Demo summary

The narrator demonstrates how to use the LALAL.AI web interface by selecting the stem splitter, adjusting neural network settings like 'Deep Extraction' and 'De-echo', and uploading a file to generate and download separated tracks.

Step-by-step

Select the Stem Splitter product on the official website
Click the settings icon to configure the neural network and processing method
Enable 'Deep Extraction' and 'De-echo' if desired
Drag and drop your audio file into the interface
Listen to the generated preview in the UI
Download the individual processed stems once the processing is complete

Options

Choose between 'Orion' and 'Phoenix' neural networks
Select 'Clear Cut' processing to minimize cross-bleeding
Select 'Deep Extraction' processing for more detail
Enable 'Leadback Separation' when separating vocals

Watch out for

Deep extraction might increase cross-bleeding between stems
Processing time varies depending on the length of the audio file

Tips

Use the De-echo function to improve results
Use the built-in UI to listen to the preview before downloading to ensure you are satisfied with the result

Highlights

“Super easy.”

AI Vocal Remover & Stem Splitter

3:110:39Extract lead vocals using LALAL.AI Stem SplitterThe demonstrator uses the stem splitter tool, selecting the 'lead and back' separation setting to isolate Sabrina Carpenter's vocals from an instrumental for a cover project.LALAL AI
10:060:37Separate voice from music with LALAL.AIThe video shows LALAL.AI being used to upload a complex audio file and successfully separate the speaker's voice from background piano music.PiXimperfect
9:150:32Navigating the LALAL.AI web interfaceThe creator walks through the LALAL.AI web application, demonstrating a playback bug where the playhead moves without audio and the page refreshes unexpectedly.LANDR
1:082:46Split lead and backing vocals with LALAL.AI AndromedaThe presenter demonstrates the LALAL.AI web interface, selecting the V6 Andromeda neural network and dragging an audio file to split it into lead vocals, backing vocals, and instrumental tracks. He then plays back each isolated stem to showcase the separation quality.LALAL AI
1:320:34Separate vocals and music with LALAL.AI Stem SplitterThe narrator demonstrates how to upload a song to LALAL.AI to separate it into a clean vocal track and an instrumental background track.AI BORDER
1:523:06Isolate vocals and drums with LALAL.AI web interfaceThe user selects the 'vocal and instrumental' stem option, uploads a track, and uses the UI sliders to preview the isolated instrumental and vocal tracks before switching to isolate the drums.Parlogram
5:062:29Extracting bass and drum stemsThe narrator demonstrates selecting the bass stem and then the drum stem for the song 'Penny Lane' to generate and play back isolated previews.Parlogram
9:052:35Separate lead and backing vocals in LALAL.AI desktop appThe user demonstrates the desktop application by dragging a file into the workspace and using the 'lead and backing vocals' splitter to isolate three distinct tracks.Parlogram
1:051:41Extracting strings from a track with LALAL.AIThe user uploads a track and selects the 'strings' stem separation type to isolate the string section from the rest of the audio.Spinlight Studio
2:461:18Isolating piano stems using LALAL.AIThe narrator demonstrates selecting the 'piano' stem option and creating new previews to separate the piano part from a vocal and string arrangement.Spinlight Studio
4:530:48Synthesizer stem separation in LALAL.AIThe user tests the synthesizer stem splitter on an indie pop track to isolate a specific synth line.Spinlight Studio
5:410:55Extracting bass lines with LALAL.AIThe narrator demonstrates the 'bass' separation feature to pull a baseline out of a song for use in a backing track or cover.Spinlight Studio