Extract a vocal track from a mixed song and take it into your next music project. Upload a supported audio file, preview the vocal and backing track, then download the result for a remix, sample idea, cover arrangement, or mashup.
Drop an audio file here
MP3, WAV, FLAC, M4A, OGG or WebM up to 50MB
No stems yet
Your vocal and instrumental tracks will appear here when the AI task completes.
An acapella is a starting point for a musical idea. Choose a use, then check timing, pitch, and background sound before it becomes part of a finished arrangement.
Place the vocal over a new beat, add transitions around its phrases, and try different arrangements. Keep a copy of the untouched extraction so you can return to the original timing while experimenting.
Audition short phrases as possible hooks or textures. Trim clean phrase boundaries, leave room for consonants, and check that edits do not cut off breaths or create clicks at the loop point.
Use the isolated vocal as a reference when rehearsing harmonies, studying phrasing, or sketching a new accompaniment. It can help you hear entrances and sustained notes that are harder to follow in a full mix.
Test a vocal against another song's groove or instrumental before committing to an arrangement. Check the key and tempo of both sources, and make sure the combination works across complete phrases rather than only a short preview.
Create a vocal-and-instrumental pair, then choose the vocal file for your project.
Choose an audio file up to 50MB in MP3, WAV, FLAC, M4A, OGG, or WebM. Use a direct public audio-file link only when it points to the file itself and does not require a sign-in.
Start the separation and let the tool analyze the mixed recording. It creates an isolated vocal stem alongside an instrumental stem, so you can compare the two outputs from the same source.
Preview the vocal from beginning to end, including quiet intros, layered choruses, and long reverberant notes. If it is a useful source for your idea, download it and continue editing in your DAW.
Extraction separates parts of a completed mix; it does not automatically adapt the vocal to a new song. Treat tempo and key matching as a production step after download.
Use your DAW's tempo analysis or tap along to estimate the original BPM. Listen for the start of the beat and compare several bars, because a small tempo error becomes obvious when a phrase repeats over a new rhythm.
Place the first clear downbeat or phrase entrance on the grid, then check later entrances. A song may have a pickup before bar one, a loose live performance, or tempo changes, so one automatic warp setting may not fit every section.
A tuner or key-detection tool can suggest the original key, but confirm it against sustained notes and the harmony. If a new instrumental conflicts with the vocal, try a modest pitch shift and listen for changes to tone, formants, and consonants.
Stretch a phrase only as much as the arrangement needs. Extreme changes can make transients, breaths, and vibrato sound unnatural. Compare the processed vocal with the untouched download and keep the original available for a different edit.
A finished mix can place instruments and vocals in the same frequency range. Some beat, guitar, reverb, or backing sound may remain with the voice. A cleaner source usually helps, but an extracted track is not guaranteed to be studio-isolated.
Before adding effects, check whether each phrase begins and ends where the new music supports it. Use short fades to avoid clicks, leave natural breaths where they help the performance, and avoid noise reduction that removes consonants or room tone unevenly.
The downloaded track is material for your session. Use focused edits to reduce distractions while preserving the character of the performance.
Use headphones and listen between words as well as during the main lyric. Quiet tails can reveal cymbals, bass, or chords that are masked while the singer is active. Check both the vocal and instrumental outputs to understand what moved between stems.
Import the vocal and your new instrumental into the same DAW session. A sound that seems noticeable in solo may disappear in context, while phase or timing differences can become more apparent when the source instrumental is also playing.
A two-stem separation groups vocal content together. It does not reliably deliver the lead singer, backing singers, doubles, and ad-libs as individually editable tracks. If layers overlap, edit phrases and levels in the combined vocal stem where possible.
Try a small EQ cut for a persistent tonal remnant, a gate only on quiet gaps, or a short clip fade between phrases. These tools can also damage the vocal, so bypass each change often and compare at the same playback level.
Save the original download on a separate track before stretching, tuning, denoising, or chopping. Name working versions clearly and keep an unprocessed copy in the project so a later arrangement change does not force you to repeat the extraction.
Check the downloaded file's sample rate, bit depth, and channel layout in your session. A compatible file can still need conversion to match an existing project.
WAV is a convenient choice for arranging and further processing because it is uncompressed. It can take more storage than a compressed file, so keep the project folder organized when collecting several stems.
MP3 files are compact for sharing a rough idea or listening away from the studio. For additional processing, avoid repeatedly converting between compressed formats because each lossy encode may add artifacts.
FLAC keeps audio lossless while reducing file size, when your editing software supports it. If the DAW does not import it directly, convert once to the session's preferred format before editing.
Place the vocal at the intended start point and verify that the DAW has not shifted it during import. Keep the source arrangement's timing in mind when comparing a phrase with a different beat or a newly recorded accompaniment.
Answers for producers preparing an extracted vocal for a new musical context.
You can use the vocal file as material for a remix draft or other project, but extraction does not grant permission to publish or monetize the underlying performance or composition. Check the rights and licenses that apply to your source and intended use.
The vocal retains the timing and pitch of the source performance. It is not automatically synchronized to a new beat or transposed into another key. Use your DAW to analyze the source, align phrases, and make careful tempo or pitch adjustments if your arrangement needs them.
Start with the cleanest source recording available and listen for where the residue occurs. Try a narrow EQ adjustment, a fade between phrases, or light noise cleanup in your DAW, then compare with the untreated track. Strong processing can also thin the singer or distort consonants.
This workflow creates a vocal stem and an instrumental stem, not separate lead and backing-vocal tracks. Harmonies, doubles, and ad-libs may remain together with the lead. You can edit the combined result in a DAW, but individual layers may require a different source or tool.
The tool does not change copyright ownership or grant a license. Whether you may sample, distribute, perform, or monetize an extracted vocal depends on the rights to the recording and composition, your permission, and applicable law. Obtain authorization when required.
Use WAV for an uncompressed editing copy when available and supported by your DAW. MP3 can be convenient for a compact preview, while FLAC offers lossless compression if your software accepts it. Check sample rate and bit depth after import and convert only when needed.
Upload a song you are allowed to use, preview the extracted vocal, and decide whether it fits your arrangement.