← Articles / How to use the MV Studio to make a lyric video
How to use the MV Studio to make a lyric video
A walkthrough of the studio on this site: load audio, add lyrics, time them, pick a style, and export a video plus a poster frame.
2026-08-18 · updated 2026-09-14 · 8 min read
What the studio does
The MV Studio at /tools/lyrics-mv/studio/ turns an audio file and a block of lyrics into a lyric video. Everything runs in your browser: the audio is decoded locally, the visual is drawn on a canvas, and the export is captured from that canvas together with the audio track. Nothing is uploaded to a server for processing.
It exists because the auto-generated video export from a music platform is a static image with compressed audio, and that's rarely what you want on a channel you care about. Here you control the palette, the typography, the timing and the aspect the text arrives in.
Step 1 — load the audio
- Use the file input for a track you already exported — the lossless file if you have it.
- Or paste a public song URL; the page fetches it through the site's own proxy.
- Wait for the waveform-free timeline and duration readout to appear; that confirms decoding finished.
- Play a few seconds to check the audio is the take you intended before you invest time in timing.
Step 2 — get the lyrics in
Paste the lyrics into the textarea. Two modes work. If your lines already carry timestamps in [mm:ss.xx] form, the studio uses them directly and your timing is done. If they're plain text, lines are distributed evenly across the track's duration as a starting point.
Even distribution is fine for a steady song and obviously wrong for anything with an instrumental intro or a long outro. If you need real timing, the lyrics-timestamp tool on this site is built for that: play the track, tap each line as it lands, and export LRC. Paste that back here and everything snaps into place.
[00:12.40] The kitchen light is on at four
[00:16.10] You left your coat beside the door
[00:19.80] I count the cars that never stopStep 3 — choose a look
| Control | What it changes | Advice |
|---|---|---|
| Style preset | Gradient palette and overall mood | Pick one that matches the track's tone, not your favourite colour |
| Font size | Text scale on the canvas | Bigger than feels right — phone screens are small |
| Position | Where lines sit in frame | Centre or lower third; avoid edges that get cropped |
| Animation mode | fade, slide or karaoke highlight | Fade for ballads, karaoke for anything with a clear pulse |
Step 4 — preview honestly
- Play from the start, not from the middle; intros are where drift shows first.
- Scrub to each chorus and confirm lines land with the beat, not after it.
- Check the longest line still fits the frame at your chosen font size.
- Watch the last thirty seconds — outros are where auto-distribution breaks down.
- Look at the preview at a small window size to approximate a phone.
- Fix timing in the lyrics, not by changing the animation to hide it.
Step 5 — export
Export produces a WebM captured from the canvas with the audio track muxed in, plus an optional poster PNG. Grab the poster while you're there — it's the natural thumbnail and it's already in the right style.
Two practical points. First, capture happens in real time, so a four-minute song takes about four minutes to export; leave the tab focused and don't let the machine sleep. Second, WebM is widely accepted by video platforms, but if a destination insists on MP4 you can convert the file afterwards without touching the audio quality.
Tips that make it look deliberate
- Leave the frame empty during instrumental sections rather than holding the last line on screen.
- Keep lines short — the same discipline that makes lyrics singable makes them readable.
- Use one preset for a whole release so the videos read as a set.
- Export the poster at the same time and use it as the cover everywhere.
- Save your lyrics-with-timestamps file; it's reusable if you ever remake the video.
- If the text is hard to read over a busy gradient, change the preset rather than adding a shadow.
The whole thing is fifteen minutes of work for a result that doesn't look automated. That's the entire pitch.