To isolate vocals from a song, choose the audio file above and keep Vocals selected. StemVocal estimates a vocal track and an instrumental mix so you can hear what the model left behind. Download the tracks as WAV or MP3. Choose an MP3, WAV, FLAC, M4A or OGG input; there is no link or streaming-service import.
The result player selects Vocals first. The track buttons below it keep your listening position when you switch to Instrumental or Original. Free Vocals separations can return just vocals and instrumental; Four stems, Drumless and paid separations include all six tracks. The original comparison is only available before a reload or reset.
What the vocal stem actually contains
The processor works on a mixed recording, not the singer’s original microphone track. It estimates vocal sound separately from musical backing. That can bring the lead singer forward, but it can also include backing singers, vocal effects and reverb. A choir or harmony part may not separate cleanly from the lead. If the source is a duet, the two voices are usually part of the same vocal stem rather than two individual files.
Some instrumental sound may come along with the voice. High percussion, distorted guitar and synth tones can share frequencies and textures with vocal sounds. The inverse can happen too: a trace of voice may remain in the instrumental. Listen to the result with headphones, especially at the start and end of vocal phrases where reverberation can be more obvious. A separated stem is useful raw material, but it is not a promise of a perfectly dry studio recording.
If you only need the backing, the vocal remover page starts with Instrumental selected. If you want all four parts to edit separately, use the stem splitter. The tools share the same file limits and separation engine; each page simply starts at the result most useful for its task.
Ways to listen and work with an isolated vocal
For transcribing a melody, the vocal stem can make notes and syllable entrances easier to hear against a busy accompaniment. Open Practice a section below the player to slow the preview to 75% or 50%. Enter start and end times, or use the playhead buttons, then turn on Loop to repeat the phrase. Switching tracks keeps those settings, so you can compare the same passage with the instrumental. The individual track downloads keep their original speed and length. There is no pitch display, score transcription or lyric detection; you still need to listen and check the notes yourself.
For singing practice, compare the original recording, the vocal stem and the instrumental. The isolated voice can help you focus on breathing, timing and phrase endings. Then sing with the instrumental to see whether you can keep those details without the guide vocal. A headphone mix is useful so the separated track does not bleed into your microphone while you practice.
For remix planning, test a busy chorus and a sparse verse before you commit to an arrangement. A voice can sound clean in one section and carry cymbals or synths in another. The acapella-making guide explains how to bring the vocal track into an audio project, where you can edit breaths, pauses or artifacts. If you plan to publish the remix, obtain the rights required for the composition and recording. This tool does not provide a license to use someone else’s song.
Input quality and honest limits
Start with the best source file you are allowed to use. A heavily compressed file can already contain swishy high-frequency artifacts, which may be emphasized when the separator extracts a stem. Saving that file as WAV does not recover lost information. A clean source can improve your chances, but even lossless audio cannot guarantee that overlapped instruments and voices will be separated perfectly.
Free separation needs no sign-up. Keep the tab open until your tracks are ready.
Download your tracks before closing this tab. See privacy and terms for storage and use details.
To save the whole selected track with a different speed or key, open Practice a section, then Download with speed & key. Choose Export track, then Download WAV. It uses the native stem WAV for new results, runs on your device in supported browsers and uses no extra song credit. Loops are for playback; this export saves the whole track.