Voice isolation keeps the speaking voice in a recording and removes everything else: traffic, room tone, hum or a music bed. It is applied to a finished file by a model, or live as a microphone mode. To pull a voice off a music bed in your browser, use the free vocal remover.
The live version is what Apple calls Voice Isolation on iPhone, a microphone mode that, in Apple's words, prioritizes your voice and blocks ambient noises (support.apple.com/en-us/101993). On a file, noise removal, voice isolation and stem separation overlap but are different jobs. Older noise reduction turns down bands of frequencies, which suits steady hiss or hum and can take some of the voice with it. A voice-isolation model rebuilds the speech as its own signal, so traffic or a music bed can drop out almost entirely while the voice stays full. Stem separation splits a finished song into vocals, drums, bass and other. For a voiceover with background noise, Studio Sound cleans it in the browser; to split a track, use the vocal remover.
The wins are largest on material you cannot re-record: an interview beside a road, a phone recording in a busy room, archive footage. It is also a preprocessing step: transcription, dubbing and lipsync all work better on an isolated vocal than on a full mix.
There are limits worth knowing. Reverb is part of the voice's own signal and is harder to remove than noise sitting behind it. And aggressive isolation on already-clean audio makes things worse, because the model starts removing parts of the speech it is unsure about.
In practice
- Run it before transcription or dubbing, not after.
- Reverb and clipping resist isolation; steady background noise does not.
- Compare against the original at full volume; over-processing is easiest to hear on breaths.
Voice Isolation: common questions
What does voice isolation mean?
Voice isolation means separating a speaking voice from everything else in a recording, such as traffic, room noise, fan hum or background music, so only the voice remains. It is different from stem separation, which splits a whole song into vocals, drums, bass and other parts.
What is voice isolation on iPhone?
It is one of the iPhone's microphone modes. Apple describes Voice Isolation as prioritizing your voice and blocking ambient noises, during calls and in compatible apps. It works while you talk; the voice isolation this page defines is applied to a recording you already have.
The mistake to avoid
Isolating audio that was already clean. There is nothing to remove, so the model removes detail from the voice instead.
Where you will run into it
- Isolate Vocals From a Track — Strip the instrumental. Keep the voice.
- AI Video Editor — Upload the clip. Say what's wrong with it.
- AI Caption Generator — Drop a video, get timed captions. Free, on your device, no upload.
- Vocal Remover — Free tool. Split a track into vocals and instrumental, or four stems, on your device. Solo, mute, download WAV.
- Studio Sound — Free tool. Use on-device RNNoise neural cleanup for stronger voiceover enhancement than light DSP denoise.
Related terms
Speech-to-text
Speech-to-text meaning: converting spoken audio into written text, the transcript captions, translation, and search all depend on.
Stem separation
Stem separation meaning: splitting a finished mix into parts (vocals, drums, bass, other) as separate audio files.
Speech-to-speech
Speech-to-speech takes a recording of one person talking and re-renders it in a different voice, keeping the original performance intact.
AI dubbing
AI dubbing meaning: replacing spoken audio with another language, usually keeping the speaker voice and optionally re-syncing lips.
Burned-in captions
Burned-in captions meaning: subtitles rendered into pixels, so viewers cannot turn them off or lose them on re-upload. Closed captions are a separate track.
The all-in-one AI studio for creators. 60+ models for video, image, voice, music and lipsync, in your browser or on your phone.
Free on iPhone. On a computer? The same account works on Versely Web, no install needed.