I run a small music and video production studio in Seoul. Some of you might know my previous tool, Vilm Care (a free Windows optimization tool for audio work, kindly received here and on the RME forum). My work has always been about helping creators, and somewhere along the way that turned into building programs for them. This one is for singers and musicians who put their own performances on screen.
If you've ever released a live take, a cover, or a lyric video, you know the chore: your captions have to match what you actually sang, word for word. Auto-captions don't. Singing breaks them: stretched-out notes, ad-libs, humming, mixed languages. They rewrite the words the artist actually wrote. So this tool works the other way around. Your lyrics are always the subtitle text; speech recognition is only used to find the timing, never to rewrite. It separates vocals from the accompaniment first, so it stays stable on live recordings. Paste your lyrics, open the audio or video, and you get an SRT with each line placed where it's sung.
What it covers:
- Your original lyrics in, timed SRT out (text and order always preserved)
- Vocal separation before alignment, so it holds up on live recordings
- Mixed Korean/English lyrics
- Whole file, or just a drag-selected range
- Auto mode for a fast first pass, Manual mode for per-line control
- Windows 10/11 and Apple Silicon macOS standalone app, plus an optional DaVinci Resolve Studio panel (standalone works with any editor that imports SRT, including Resolve Free)
GitHub: https://github.com/banjuman/vilm-lyrics-aligner
Download: https://www.voiceandfilm.com/product/so ... s-aligner/
Honest caveats: it's an early release, so it isn't code-signed (Windows) or notarized (macOS) yet, and you may see a security warning on install. First setup pulls a few GB into its own isolated runtime, so it won't touch your other environments. And singing is expressive, so long sustains and ad-libs can still need a small manual touch. It's meant to cut hours of timing work down to minutes, not to be zero-edit.
Hope it helps any singer or musician who's ever timed lyric subtitles by hand, or given up on auto-captions halfway through a song. Feedback and suggestions for what to add next are welcome.