← Voltar

Como funciona Removedor de voz Music remover Perguntas frequentes

Voz
Isolation

Isolate the voice in any recording and hear it on its own, in the original key and quality

MP3, WAV ou FLAC · até 20 MB e 5 minutos · 25 créditos por faixa

How vocal isolation works

  1. Upload the track. MP3, WAV or FLAC up to 20 MB and five minutes. Nothing is transposed or time-stretched at any point, so what you isolate stays in the original key and at the original tempo.
  2. The model isolates the voice. A network trained on real isolated stems reconstructs the vocal on its own. Because it is reconstructing rather than filtering, the isolated voice keeps its top end and its body — it does not come back thin and phasey the way an EQ-based attempt does.
  3. Listen to it alone, or under the mix. The page opens with the isolated vocal by itself. Lifting the music fader back to around 15–20% is often the more useful setting for study: the voice sits clearly on top while the arrangement stays there for context.
  4. Keep it. Export as WAV for transcription software or a DAW, MP3 for a phone. Whatever balance is on screen is what gets saved.

What people isolate vocals for

  • Transcribing lyrics — the line you could never quite make out, with the mix out of the way.
  • Learning a vocal part — hearing the phrasing, the timing and where the breaths actually fall.
  • Vocal coaching — working through a student’s reference track one phrase at a time.
  • Studying production — how many doubles are stacked, where the harmonies enter, what the ad-libs are doing.
  • Dubbing and captions — a clean voice track to sync subtitles or a translation against.
  • Reference before you track — hearing a finished vocal naked, right before you record your own.

Perguntas frequentes

Is isolation different from removal?
Same separation, opposite half. Removal throws the voice away and keeps the music — that is the karaoke case. Isolation keeps the voice and throws the music away. The split runs once and produces both, which is why this page and the remover cost the same and give you the same two faders; they just open on different settings.
Does isolating change the pitch or the key?
No. Nothing is transposed, stretched or resampled — the isolated vocal sits at the original pitch and tempo, sample-aligned with the track it came from. If you do want it moved, that is a separate step in the key and tempo shifter, which changes one without dragging the other along.
How clean is the isolated voice?
On most commercial mixes, clean enough to transcribe from confidently and to perform against. Two things leave traces. Reverb and delay tails belong to the voice musically but sit in the music statistically, so parts of them stay behind. And instruments that share the vocal’s register — a lead synth, a legato string line — are the hardest thing to pull apart, because in that band they genuinely look alike.
Can I isolate one singer out of two?
Not separately — the model isolates voice as a category, so a duet comes back as both singers on one track. Panning helps if the two are hard-panned in the original, but there is no per-singer control here and no separator on the market does it reliably yet.
Can I isolate speech instead of singing?
Yes — speech and singing occupy the same range and have the same harmonic structure, so the model handles both. If the source is an interview, a vlog or a voiceover rather than a song, the background music remover is the same tool written for that job.

Uma só separação, seis portas de entrada

These are all the same AI separation with the faders in a different place. Whichever one you start from, the split runs once and you get both tracks.