ElevenLabs Voice Isolator: Remove Background Noise From Audio

The ElevenLabs voice isolator removes background noise, ambient sound and music from a recording and returns clean speech. It accepts audio and video files, costs about 1,000 credits per minute on subscription plans or $0.12 per minute through the API, and suits podcast interviews, field recordings and phone audio. Below: the steps, real per-plan capacity and the limits to plan around.

Last updated: October 2026. Plans and credit costs checked against ElevenLabs’ official pricing page.

What is the ElevenLabs voice isolator?

The ElevenLabs voice isolator is an AI audio cleanup tool that separates speech from everything else in a recording. ElevenLabs describes it as a way to isolate voices from background noise, and the output is a cleaned speech track you can drop straight into an editor, transcription tool or dubbing project.

“Isolate voices from background noise”

Source: ElevenLabs documentation

It is one of several audio tools in the suite, alongside text to speech, speech to text and the voice changer. The full review of ElevenLabs shows where it fits among them. In one line: upload noisy speech, download clean speech.

What kinds of noise can the voice isolator remove?

The voice isolator removes non-speech sound: according to ElevenLabs’ docs, it targets background noise, ambient sounds, music and other interference in audio recorded in noisy environments. It keeps the spoken voice and discards the rest, which is different from traditional noise gates that only reduce steady hiss.

“Noise reduction is the process of removing noise from a signal.”

Source: Wikipedia, Noise reduction

Five common problems it is built to handle:

• Room and street ambience: Traffic, air conditioning, crowd murmur and other constant environmental sound behind a speaker are stripped away, leaving the voice in front.
• Background music: Music playing under dialogue is removed along with other non-speech sound, useful when a venue or cafe recording has an unwanted soundtrack.
• Intermittent noises: Door slams, keyboard clicks and passing sirens count as interference, so short bursts that ruin an otherwise good take can be cleaned out.
• Phone and video call audio: Remote interview recordings with background clutter benefit from isolation before editing, making guests recorded at home sound closer to studio quality.
• Field recordings: Interviews captured outdoors or on location with wind and ambient noise become usable dialogue tracks for documentaries, vlogs and news-style content.

If the unwanted sound is not speech, the isolator is designed to remove it.

Which files can you upload?

The voice isolator accepts both audio and video uploads, so creators do not need to extract audio before cleaning it. ElevenLabs’ docs list common audio formats such as MP3, WAV, FLAC, M4A and OGG, and video formats such as MP4, MOV, MKV and WebM.

Each upload has a maximum file size and duration, so long recordings such as multi-hour livestreams should be split into parts before processing. Video uploads are useful for vlogs and interviews shot on a camera or phone, because the cleaned audio can be relinked in an editor without a separate export step.

Most standard creator files work as they are; only very long recordings need splitting.

How much does the ElevenLabs voice isolator cost?

The ElevenLabs voice isolator costs about 1,000 credits per minute of audio on subscription plans, drawn from the same pool as every other ElevenLabs product. On the Free plan’s 10,000 credits that is roughly 10 minutes a month, rising to about 121 minutes on Creator and about 600 on Pro.

Source: ElevenLabs pricing page (elevenlabs.io/pricing), checked October 2026. Minutes calculated by GrowthStackKit from ~1,000 credits per minute.

Cleaning full podcast episodes on a subscription burns credits fast: one 60-minute episode uses about 60,000 credits, which is double Starter’s monthly allowance.

Clean only the segments that need it, or use the API, to keep credit use under control.

How do you remove background noise from audio step by step?

Removing background noise with the ElevenLabs voice isolator takes five steps in the web app: open the tool, upload the file, process it, compare the result and download the clean track. The tool has few settings, so the main decisions are which parts of the file to process and whether the result sounds natural.

1. Open Voice Isolator: Signing in and choosing Voice Isolator from the ElevenLabs dashboard opens a simple upload panel, with earlier processed files available for download again.
2. Upload the audio or video: Dragging in an MP3, WAV, M4A or MP4 file starts the job, ideally trimmed beforehand to only the sections that actually contain noise problems.
3. Run the isolation: Processing separates speech from background sound and deducts about 1,000 credits per minute of the uploaded file from the monthly credit balance.
4. Compare before and after: Listening to the original and cleaned versions side by side reveals artifacts, clipped word endings or removed sounds that should have stayed.
5. Download and relink: Downloading the clean speech track and placing it in the editing timeline completes the fix, with room tone or music added back separately if needed.

Trimming before uploading is the simplest way to cut cost without changing quality.

Voice isolator API pricing compared

Voice isolator API pricing is $0.12 per minute of audio, billed directly rather than from subscription credits. That puts noise removal at the same per-minute API price as the voice changer and sound effects, making it easy to estimate the cost of cleaning a back catalog programmatically.

Source: ElevenLabs API pricing page, checked October 2026.

On the API, cleaning a 60-minute episode costs about $7.20, while transcribing that same hour with Scribe v2 costs $0.22. For podcasters with dozens of archived episodes, the API is usually cheaper than buying a larger subscription just for isolation; the ElevenLabs API quickstart covers setup.

Per minute, isolation is one of the pricier steps in an audio pipeline, so process only what needs it.

Where the voice isolator fits in a podcast workflow

In a podcast workflow, the voice isolator belongs right after recording and before editing, so every later step works with clean speech. Cleaning first improves transcription accuracy, makes cuts less noticeable and prevents background sound from jumping in and out at every edit point.

A typical five-stage chain looks like this:

• Recording: Each host and guest records on a separate track where possible, which keeps problems contained to one file instead of spreading across the mix.
• Isolation: Only tracks with audible noise go through Voice Isolator, saving credits on clean studio tracks that would gain nothing from additional processing.
• Transcription: Cleaned tracks run through a speech to text tool such as Scribe, producing more accurate transcripts for show notes, captions and searchable archives.
• Editing and mixing: Clean dialogue edits smoothly in any audio editor, with music beds, intros and room tone layered back in deliberately rather than by accident.
• Publishing and repurposing: The finished episode, transcript and short clips go out together, with clean audio making social clips sound professional on small speakers.

For the full production stack, see the guide to making a podcast with AI tools, and for transcription details the ElevenLabs Scribe transcription guide. Isolate early, and every downstream tool performs better.

What are the limits of the voice isolator?

The voice isolator’s limits come from what it was built for: it extracts speech, so it is not a general audio separation tool. ElevenLabs’ docs say it is not specifically optimized for isolating vocals from music, and every upload is capped in size and length, so plan around these before relying on it.

• Vocals in songs: Separating sung vocals from a music track is outside its design focus, so dedicated stem separation tools are a better fit for music production work.
• Intentional background music: Any background music in the file is treated as interference and removed, so intentional music beds must be added back after processing.
• Upload caps: Each file has a maximum size and duration, which means long livestreams or full-day recordings need splitting into smaller segments before upload.
• Credit intensity: At about 1,000 credits per minute, isolation consumes subscription credits three times faster than transcription at about 330 credits per minute.
• Free plan licensing: Audio processed on the Free plan is limited to non-commercial use, so cleaned audio for monetized podcasts or client work requires a paid plan.

Use it for speech cleanup, not music production, and budget credits per minute.

Tips for cleaner results

Cleaner results start before the voice isolator ever runs, because no tool can fully rebuild speech that was buried or distorted during recording. A few recording and preparation habits make the isolator’s job easier and leave fewer artifacts in the final speech track.

Five habits make the biggest difference:

• Close microphone placement: Speaking near the microphone raises the voice well above the background, giving the isolator a strong speech signal to keep and less noise to remove.
• Moderate recording levels: Keeping input levels below clipping avoids distortion, which the isolator cannot repair because distorted speech is baked into the waveform itself.
• Separate speaker tracks: Recording each person on a separate track lets only the noisy tracks be processed, saving credits and keeping clean voices untouched.
• Pre-trimmed uploads: Cutting silence, retakes and off-topic sections before upload sends only usable speech for processing, lowering the per-minute credit cost of every job.
• Room tone added back: Layering a little room tone or a quiet music bed after processing prevents cleaned dialogue from sounding unnaturally dry against the rest of the mix.

Good input plus selective processing produces the most natural output.

Should you isolate audio before transcription or dubbing?

Isolating audio before transcription or dubbing is worth it whenever the recording has audible background noise. Cleaner input helps speech recognition avoid missed or invented words, and dubbing benefits from a clean dialogue track that is easier to separate from music and effects.

For clean studio recordings, skip the step: you would spend about 1,000 credits per minute for little gain. For dubbing projects, the ElevenLabs dubbing guide explains how dubbing credits work per minute.

Isolate noisy sources, skip clean ones.

Is the free plan enough for noise removal?

The free plan is enough to test the voice isolator but not to run a show on it. Its 10,000 monthly credits cover roughly 10 minutes of isolation, and the Terms limit free use to non-commercial purposes, so monetized podcasts and client audio need at least the Starter plan.

Starter’s 30,000 credits cover about 30 minutes, which suits fixing problem segments rather than whole episodes. Creator’s 121,000 credits cover about two hours. The ElevenLabs free plan breakdown details every free limit.

Test free, then upgrade based on how many minutes you clean each month.

Voice isolator vs all-in-one podcast editors

The ElevenLabs voice isolator is a single-purpose cleanup tool, while all-in-one podcast editors bundle recording, editing and noise cleanup in one app. If you already pay for ElevenLabs for voiceovers, the isolator adds cleanup without another subscription; if you mainly edit podcasts, an editor may be the better home base.

Podcast-first platforms such as Podcastle and Descript combine recording, transcript-based editing and audio enhancement. The Podcastle review and the Podcastle vs Descript comparison cover those editors in depth.

Choose ElevenLabs for cleanup inside a voice workflow, and an editor for end-to-end podcast production.

Bottom line on the ElevenLabs voice isolator

The ElevenLabs voice isolator reliably turns noisy speech into clean dialogue, and accepting both audio and video makes it convenient for creators. Its weak spot is cost on subscriptions, at about 1,000 credits per minute, so process selectively or use the $0.12 per minute API for large back catalogs.

Voice Isolator earns its place as a fix-it tool, not a whole-episode filter.

Source: GrowthStackKit, ElevenLabs Review

Try it free on a problem clip, then decide between a plan upgrade and the API.

Frequently asked questions

The 12 most-asked questions about the ElevenLabs voice isolator.

What does the ElevenLabs voice isolator do?

The voice isolator removes background noise, ambient sounds, music and other interference from a recording and returns clean speech. It is designed for audio recorded in noisy environments, such as field interviews, remote calls and vlogs, and accepts both audio files and video files as input.

How many credits does the voice isolator use?

Voice Isolator uses about 1,000 credits per minute of audio on subscription plans. That is roughly 10 minutes a month on the Free plan’s 10,000 credits, about 30 minutes on Starter, about 121 minutes on Creator and about 600 minutes on Pro.

How much does the voice isolator API cost?

Through the API, Voice Isolator costs $0.12 per minute of audio. Cleaning a one-hour recording therefore costs about $7.20. That is the same per-minute API price as the voice changer and sound effects, and it is billed separately from subscription credits.

Can the voice isolator remove music from speech?

Yes, background music under dialogue is treated as interference and removed along with other non-speech sound. However, ElevenLabs says the tool is not specifically optimized for isolating sung vocals from music, so separating a singer from a song is better handled by dedicated stem separation tools.

Can I upload video to the voice isolator?

Yes. ElevenLabs’ docs list video formats such as MP4, MOV, MKV and WebM alongside audio formats like MP3, WAV and M4A. The cleaned speech track can then be relinked in a video editor, which saves a separate audio export step for vlogs and interviews.

Is there a file size limit for the voice isolator?

Yes. Each upload has a maximum file size and duration set by ElevenLabs, so very long recordings such as livestreams need to be split into smaller parts. Trimming files to only the noisy sections before uploading also saves credits, since processing is charged per minute.

Is the voice isolator free?

You can use Voice Isolator on the Free plan, whose 10,000 monthly credits cover about 10 minutes of audio. Free use is limited to non-commercial purposes under the Terms, so cleaned audio for monetized podcasts, YouTube videos or client work requires a paid plan from Starter upward.

Should I clean audio before transcribing it?

Yes, when the recording has audible background noise. Cleaner input helps speech to text avoid missed or invented words. For clean studio recordings, skipping isolation saves about 1,000 credits per minute with little accuracy gain, since Scribe already handles clear speech well.

Does the voice isolator change how my voice sounds?

It is designed to keep the original voice and remove only non-speech sound, unlike the voice changer, which replaces the voice. Heavily distorted or buried speech can still show artifacts after processing, so compare the original and cleaned versions before replacing audio in a final edit.

Voice Isolator or Voice Changer: which do I need?

Use Voice Isolator to keep your voice and remove background noise. Use Voice Changer to convert your performance into a different target voice. Both cost about 1,000 credits per minute on subscription plans and $0.12 per minute through the API, so cost does not decide between them.

Is the voice isolator good for podcasts?

Yes, especially for remote guests and on-location interviews. Place it right after recording and before editing, process only tracks with audible noise, then transcribe and mix. Because one hour costs about 60,000 credits on a subscription, regular shows often find the API cheaper for full episodes.

Can I use cleaned audio commercially?

Yes on a paid plan. The pricing page lists a commercial license from the Starter plan at $6 a month, and the Terms say you retain all rights in your output. Audio processed on the Free plan may only be used for non-commercial purposes.