Audio Match
Drop an audio or video file — find scenes that sound like it.
Some footage is findable by what it sounds like and nothing else — the room you were in, the music that was playing, the machine humming off-camera. Audio Match takes a reference file, lets you trim to the exact passage you care about, and returns the scenes whose acoustic signature lines up. It matches sound, not words: for dialogue, use transcript search instead.
Getting started with Audio Match
- 1.
Point Orcah at a folder
Indexing computes an acoustic fingerprint for every scene alongside the visual one, so audio search works across the same library with no extra pass.
- 2.
Drop in a reference file
An MP3, a WAV, or a video you already have — Orcah reads the audio track either way. The reference never needs to come from your library.
- 3.
Trim to the passage that matters
Scrub the waveform and select a region, or match the whole clip. Narrowing to the few seconds that carry the sound is what sharpens the results.
What it can do
- Any source: MP3, WAV, or the audio track of a video file you drag onto the window.
- Region selection: match the whole clip or trim the waveform down to the passage that matters.
- Music and ambience: melody, instrumentation, room tone and background noise, rather than spoken words.
- Ranked results: every scene carries its match percentage, source file and timecode.
- Runs on-device: the fingerprints and the matching both stay on your Mac.
Pro Tip: matching a few seconds of clean room tone is the fastest way to group every take shot in the same space — even when the picture looks nothing alike.
Questions worth asking
- Is this the same as searching the transcript?
- No, and the difference matters. Transcript search looks at words; Audio Match looks at how the audio actually sounds. If you want the scene where someone says “pricing”, use transcript search. If you want every scene recorded in the same echoey stairwell, or scored with the same track, this is the one.
- What counts as a match when the sound is just a room?
- The fingerprint captures spectral texture — reverb, noise floor, the frequency balance of the space. Two shots recorded in the same room with the same mic will sit high in the results even with completely different dialogue over them, which is exactly what makes it useful for grouping takes.
- Can I match against music I don't own?
- You can use any file as a reference — it's only read locally to compute a fingerprint, never uploaded or stored. What comes back is your own footage, so nothing about the reference ends up in the result.
- Does it work on footage with no dialogue at all?
- That's the case it's best at. Drone shots, b-roll and action-cam clips usually have no speech to search, which leaves audio texture and the visual index as the only two ways in.
Find the moment. Not the file.
Download immediately, models included. Your footage stays on your Mac.