Frequently Asked Questions

Vocal Split is in internal testing now. The next step is a certified code signature for the installer; once it is signed, a beta with invited testers follows, and the public release comes after the bugs they find are fixed. There is no fixed date yet — apply for the beta or ask to be told about the release on the same form.

A signed installer tells Windows who published it. Without one, Windows SmartScreen warns that the app is from an unknown publisher, and many people rightly stop there. We would rather not ask anyone — testers included — to click past that warning, so no build is handed out before it is signed.

Fill in the beta application. It asks what you do with audio or karaoke, for how long, which tools you know and what you would test first. Places are limited and go to people who already work with karaoke, stems or mixing, because their feedback is the one that finds the real problems. We write to the people chosen when the signed build is ready.

The signed installer and a Pro Edition beta key that unlocks every feature — 4K, MP3+G, batch, the mixer, custom looks and the four parts — for the test period. In return we ask for a few hours over about two weeks on your own songs, and a report of what you found.

Prices will be announced at release. There will be a Free Edition to try the whole app on the first 60 seconds of any song, a Home Edition and a Pro Edition; the editions table shows what each includes.

Yes. Separation with the default model, karaoke videos, MP3+G, the party and everything else work offline. Only the two optional AI models and the word aligner are downloaded — once, when you first choose them.

No. Everything happens on your own computer. There is no account, no cloud processing and no tracking of what you separate.

MP3, WAV, FLAC, M4A and OGG for separation. The Party page also plays karaoke videos (MP4, MKV, MOV, AVI, WEBM, M4V) and MP3+G files, and the Mixer also takes AAC, Opus, WMA and AIFF tracks.

Vocal Split uses Demucs, one of the best open separation models available, followed by its own cleanup pass that removes most of the bleed between voice and music. Results depend on the recording — heavily reverberant or very dense mixes are always harder — but for karaoke the instrumental is usually clean enough that nobody notices anything is missing.

Because nearly all of a bass line is below 150–200 Hz, where laptop, phone and TV speakers play almost nothing — the meters show it moving, but the speakers cannot reproduce it. Listen on headphones, or turn on Bass on small speakers on the Effects page (or pick the “Small speakers” preset): it adds the bass's harmonics, which small speakers can play, so you hear the bass line.

It depends on your processor. On a modest two-core laptop a 4-minute song takes around 14 minutes; faster processors take proportionally less. You can mark just the part you need, and a batch can run overnight.

You paste them — or import a .txt or .lrc file. Vocal Split deliberately does not download lyrics from the internet: you stay in control of the words and of the rights to use them.

Drag the dividers on the vocal's waveform to move where a line starts, right-click to change the singer, and watch the preview update instantly. Turn on “Stop after the timing” to check everything before the video is written. Your corrections are kept with the song.

Yes — any language and any alphabet. For Bulgarian, Serbian and Russian, Vocal Split can also add the lyrics in Latin letters under the Cyrillic line.

Write who sings on a line of its own before their lines — for example [Anna], [Mark] or [Both]. Each singer's words light up in their own colour, and lines sung together in a mix of both. You can also right-click any line on the timeline to change who sings it.

Videos are standard MP4 files with H.264 video and AAC sound — the combination practically every TV, media player and USB stick setup accepts. Choose 720p for very old sets, 1080p for most, and 4K for big screens and projectors.

The words fill the whole screen and the keyboard runs the party: Space starts and pauses, → skips, ← restarts, ↑ ↓ change the key, and Esc brings you back to the app. With a TV connected as a second screen, the words go to the TV while you manage the queue on the laptop.

Yes, in the Pro Edition's Mixer. Add a take you sang, a backing track or a guide, and it is lined up with the song automatically — its start, its speed and its key — then mixed with the parts you choose and saved as one file. If it does not sound enough like the song to be placed with confidence, it is left at the start for you to move by hand in 10 or 100 ms steps.

MP3+G is the classic karaoke format used by karaoke discs, players and hosts' software (which can often change the key live). If you only play karaoke at home from a USB stick, the MP4 video is all you need. If you run karaoke nights with dedicated software, MP3+G (in the Pro Edition) fits straight into your library.

The whole app on the first 60 seconds of any song, up to three songs a day — including karaoke videos and the party — so you can judge the quality for yourself before buying.

Commercial use needs the Pro Edition licence. Please note that the licence covers the software; the rights to the music you process are your own responsibility.

Not at the moment — Vocal Split is made for Windows 10 and 11.

The interface is in English. The installer is available in fourteen languages.