Audio tools: frequently asked questions

Choose your file, select Fast AI and preview 10 seconds containing a vocal phrase. Listen to Instrumental, compare another method if needed, then process the full track and download the instrumental WAV.

Yes. The public audio tools do not require an account, subscription or payment to process and download results. Device memory, supported formats and the upload limits still apply.

Voice remover usually means keeping the instrumental while reducing the singer. Music remover usually means keeping the voice while reducing accompaniment. Choose Instrumental for the first goal and Vocals for the second.

Yes. After full separation, the results include vocal and instrumental audio that you can preview and download separately as WAV files. Save them before closing the page.

Reverberation, stacked harmonies and sounds shared by voices and instruments can leave residue. Preview a difficult phrase with Fast AI and Quality AI and compare the backing music as well as the voice. Repeated processing may add more artifacts.

It can change instrument tone or remove detail. Separation estimates sources from a mixed recording; it cannot recover the original studio tracks exactly. Basic stereo cancellation is especially likely to reduce instruments placed in the center.

Start with Fast AI and compare the same 10-second phrase with Quality AI when the result is not useful. Choose the result with the best balance of vocal reduction and intact instruments for your purpose. A slower method is not guaranteed to win on every recording.

MP3 and WAV are practical starting points. Other formats, including M4A, FLAC and OGG, depend on decoding support in your browser. If a file will not open, use a supported audio copy; renaming its extension does not convert it.

The audio tools currently accept files up to 250 MB and 30 minutes. Large decoded recordings can exceed your device memory even within these limits. For long sessions, try a shorter excerpt first.

The browser downloads the selected model before processing: roughly 40 MB for Fast AI or 166 MB for Quality AI. It may cache these files for later visits. Processing itself depends on the device, recording length and selected method.

No. The model download is software data, not an allowance of song uploads. Browser storage may be cleared and require downloading the model again; the separate file and duration limits still apply.

The audio tools process recordings on your device. The browser downloads model files; it does not upload your song for separation. The website also collects limited tool-usage analytics without audio or filenames.

AI can estimate vocals from a mono recording, although results vary. Basic stereo cancellation relies on left-right differences and is a poor choice for mono material. Preview AI before processing the complete recording.

SurStudio does not import YouTube or SoundCloud links and does not export edited video. Use an audio file you own or have permission to process. For a video project, extract its audio with a suitable editor, process the audio here and replace the soundtrack in that editor.

No dedicated lead/backing split is available. Harmonies and background singing may appear in the vocal stem, the instrumental or both. The current modes separate vocals from accompaniment rather than isolating each singer.

You can compare previews from different genres. Dense guitars, distortion, crowd noise and room reverb can make separation harder. The recording and mix matter more than the genre label alone.

Use the cleanest source available, avoid clipped recordings and compare a quiet verse with a busy chorus. Check the drums and bass at similar listening levels. Export once you have a useful result rather than repeatedly separating an already processed file.

A vocal remover separates an estimated singing track from the music in a mixed song. It produces an instrumental backing and a vocal output. These are reconstructed estimates, so neither output is guaranteed to match the original studio recording.

An AI vocal remover uses a trained model to estimate vocal patterns in mixed audio. SurStudio downloads the selected model and runs separation in your browser. Its basic stereo method instead compares left and right channels; it does not identify a singer with AI.

Compare the same short passage for remaining vocals, intact drums and bass, processing time, privacy and usable downloads. A browser tool suits quick previews without installing an app. Desktop software can offer other models and hardware options. No method is best for every song; SurStudio offers Fast AI, Quality AI and Basic stereo for comparison.

Open SurStudio in your mobile browser, choose a local audio file and start a Fast AI preview. Listen to Instrumental for karaoke or Vocals for isolation. Keep the page open while processing and save the WAV to your device. SurStudio currently provides a website, not a native app-store application. Codec support and memory vary; shorter recordings may work more reliably.

You do not need to download a SurStudio installer. The browser loads the website and downloads AI model data when needed. After full processing, use the separate vocal and instrumental download buttons to save WAV files. Model data is not a music file or a monthly usage credit.

Use the correct website address, keep your browser updated and review any requested permissions. SurStudio processes audio on your device and does not require a password or payment for its public tools. Microphone access is needed only for recording or live microphone tools. Limited usage analytics and hosting requests are described in the Privacy Policy.

Check the software publisher, supported operating system, hardware needs and model download sizes. Open an audio file, choose a separation model and output folder, test a short passage, then export vocals and instrumental separately. Settings vary by program and version; desktop instructions are not controls available inside SurStudio.

Read the publisher's license and download instructions, use its official distribution and verify the file before installing. A free application may still need model downloads and substantial device resources. SurStudio cannot certify another application or guarantee its current pricing. Its own public browser tools require no subscription.

Some separators divide recordings into processing chunks. Segment size is the duration or amount of audio handled in one chunk. Smaller chunks can reduce peak memory use but may increase overhead or change boundary artifacts. SurStudio manages chunking internally and does not expose a desktop-style segment-size setting.

It depends on the editor, version and installed effects. Basic stereo reduction relies on centered audio and may remove centered instruments. AI separation requires a supported model or plugin. You can also separate your recording in SurStudio, download the WAV outputs and import them into your editor for trimming or mixing.

Older tutorials may show an effect that is no longer bundled, has moved, or requires a separately installed plugin. Check your editor's current documentation and compatible effects. Do not assume an old menu path will work in a new version. SurStudio runs in the browser and needs no editor plugin.

A stereo-cancellation plugin can reduce centered sound during playback, but it may leave vocal reverb and remove centered instruments too. It may not export a separate audio file. For a downloadable backing track, compare the Basic stereo and AI previews here and save the full Instrumental WAV.

SurStudio does not have a paid Pro tier for these public tools. Fast AI, Quality AI and Basic stereo are different processing methods, not subscription plans. File limits, browser memory and device processing speed still apply.

Here, lyric remover means reducing the singing voice in audio. It does not edit lyrics text, remove subtitles or rewrite words. Vocal removal can change parts of the accompaniment because voices and instruments overlap in the mix.

Vocals, Drums, Bass and Other. You can save individual WAVs, a ZIP containing all four stems or a mix using the stem controls.

Yes. After splitting, mute only Drums and leave Vocals, Bass and Other enabled. Export the mix and check for residual drum sounds.

Solo Bass in the stem mixer, then download the bass WAV for listening or practice. Some kick drum or low instrument energy may remain in that estimate.

Not into individual stems in this version. Both are grouped with other accompaniment in Other. Muting that stem removes the group, not just one instrument.

No. A metronome is a generated click for timing. SurStudio separates four audio stems; its practice and Mini Studio metronomes are separate controls, not recovered parts of the recording.

Sounds share timing, pitch and timbre in the original mix. The model can assign part of a source to more than one output. Compare the solo part and full mix before exporting.

Download the stems and add them as tracks in Mini Studio. Keep their starts aligned to preserve timing, then adjust levels or record an additional performance.

Not necessarily. Four-stem mode uses roughly 79 MB of model downloads and computes several outputs. Preview a short section on your device before processing a long recording.

The desired output is opposite. Vocal removal keeps accompaniment; acapella extraction keeps the singing. SurStudio produces both from the same separation.

You can experiment with the vocal output, but music-trained separation does not guarantee clear speech recovery. Noise and music may remain, and some consonants may be affected.

Source overlap can cause instrument bleed. Compare separation modes on the same phrase and choose the most usable result; EQ alone cannot recover a perfectly isolated voice.

No. It estimates a voice already present in your recording. It does not synthesize a new performance or clone a person.

Yes. Download only the vocal result after processing. The instrumental remains available as a separate download if you want it later.

Noise underneath louder audio remains. This tool attenuates quiet passages rather than separating noise from the main sound.

Yes. Keep the pitch shift at zero and reduce speed. The quality depends on the recording and the size of the change.

No. Modulation, percussion and ambiguous harmonies can confuse an estimate. Confirm with an instrument or your ear.

No. The tool creates a downloadable edit. Your original file remains on your device.

Up to ten tracks, totaling at most 30 minutes. File size and available browser memory also apply.

The recording format depends on the browser, commonly WebM or M4A. Other editing tools can export a decoded recording as WAV.

Enable live microphone monitoring and use headphones to prevent feedback. If you hear a delayed double voice, avoid monitoring through both your audio interface and browser at once.

Yes. Seek in the track to rehearse from that position. Check a short recording before a complete take, especially after changing your audio device.

No automatic lyrics, synchronized lyric display or video export is included. This tool handles audio playback and microphone recording.

Allow microphone access in the browser and choose the intended input device. Open SurStudio in a normal HTTPS tab and check that another application has not muted the input.

Use the separate Pitch & Speed tool, keep pitch at zero and reduce speed.

Yes, with Pitch & Speed. Set pitch to zero and increase its independent speed control.

This tool controls playback. Use Audio Cutter to create a downloadable version of the selected phrase.

It adjusts level. It does not remove noise, restore clipping or separate instruments.

EQ affects every sound in the same frequency range. Use Vocal Remover or Stem Splitter for AI source separation.

No. Compression changes dynamics within the recording. Peak normalization applies a uniform level change.

No. This tool normalizes peaks to approximately −1 dB; it does not target a platform-specific integrated LUFS value.

No. It provides a live reading; it does not correct your voice or record audio.