LUMIÈRE
Start free

Lumière · Tools

Synced by the sound itself.

The camera mic heard the room; the recorder heard the voice. Upload the wav or mp3 and the engine lines the two up by the sound they share — on its own track, at the right offset. No clap needed.

Start free See the full suite

How it works

Three steps. No timeline required.

  1. 01

    Record both

    Shoot as you always do — camera rolling, external recorder running, both hearing the same scene. No clap, no timecode rig required.

  2. 02

    Upload the recorder file

    The wav or mp3 straight from the recorder. The engine compares its sound to the video and finds where it belongs.

  3. 03

    Synced, on its own layer

    The recorder track lands on its own audio layer at the right offset. The camera audio stays on its own layer — mute either, keep either.

What you get

Built in, not bolted on.

  • Aligned by sound

    Envelope cross-correlation matches the two recordings by what they both heard. In tests, a known 7.3-second offset was recovered within a millisecond.

  • Honest confidence

    The match is reported as high, medium or low — in words. We will not dress a guess up as a percentage.

  • Nudge it by ear

    When the match is weak, you get nudge controls — half a second or a single frame at a time — to settle it the way editors always have.

  • Better words downstream

    Once the clean recorder audio is on the timeline, transcription and subtitles work from it — better sound in, better words out.

v1 applies a single global offset — it does not stretch for clock drift on very long recordings. And the sync needs the two recordings to share the scene's sound; a recorder that heard a different room has nothing to match.

One suite, every tool

The same engine, other doors.

Each tool here is the same production suite, entered where you need it. Start with one; the rest is already there.

Start free