Skip to main content
Stem Splitter Docs

Getting Started

Sign in, choose a recording, and split it into parts you can listen to separately.

Course Access

Your instructor can request course access to Stem Splitter through the CUNY AI Lab. Once the Lab approves and activates the course, accept your instructor's invitation while it is open. You can then open Stem Splitter under My classes in the Lab portal.

Open Stem Splitter from My classes or sign in at stem-splitter.ailab-452.workers.dev. The app recognizes your course role. Lab members without a course can also sign in and work privately in their accounts.

Sign In

Choose CUNY Login at the top of the page and sign in with your usual CUNY account. After signing in, you can use Splitter, Classes, and Account in the navigation, with Log out at the right. Account shows your access and remaining allowances. If sign-in fails, use the app's Trouble signing in? help page to try again. You can use headphones to listen without disturbing others.

Getting Started

If you open Stem Splitter directly, Personal is selected under Share with, and you can practice privately. If you choose New class split on the Classes page, that class is selected instead. Before creating work for the course, read and check the acknowledgment described in Classes.

Stem Splitter before signing in, with CUNY Login, Insert Track, a field to search songs or paste a YouTube link, choices for two, four, or six splits, and an empty Session Rack
Stem Splitter before signing in. Upload and song search appear above the split choices.

To begin, you can open the Splitter tab and drop a file onto Insert Track, or click the drop zone to choose a file. The app accepts MP3, WAV, FLAC, M4A, OGG, and AIFF files up to 100 MB. Use music you are entitled to use for class.

Below Insert Track, you can type a song title or artist in the field labeled Search songs or paste YouTube link. Results appear as they arrive from Internet Archive, Wikimedia Commons, Freesound, and Library of Congress. Choose a recording from the list, then choose Fetch. With the list open, you can also use the arrow keys and Enter to select a result. If the field suggests a song, press Tab to accept it, then choose Fetch.

You can choose two, four, or six stems below the search field before importing audio. The app starts with four unless you have saved another available choice. Four separates vocals, drums, bass, and other; six adds guitar and piano. Two separates vocals from instrumental accompaniment. A stem may contain more than its label suggests, so listen to what is there.

You can follow progress in the Session Rack. When the new entry reads READY, open it and press play to listen to the stems together. Splitting a typical song may take a minute or two. You can leave the page and return to check its progress.

Connection Problems

If the Session Rack asks you to reconnect, choose Try again. On Account, reload the page when prompted. If your session has ended, choose CUNY Login again. Wait for your saved work to load before uploading another copy.

Listening and Notes

Once the recording is ready, choose which stems to play. You can mute the stems you want to leave out, or use Solo to bring one part forward. You can pause, slow down, or return to a passage, and add a note at a point you want to revisit. See Splitter Basics for the Session Rack and mixer controls.

You can also ask Listening Guy a question. Listening Guy does not receive the recording or its waveforms, so listen to check its descriptions. Course work is subject to the review policy you acknowledge, and the app deletes uploaded audio and stems 90 days after a split. See Privacy & Limits for details.