Foley and sound design for video: a musician's guide to props, sync, layers and loudness

Foley is the reproduction of everyday sound effects, such as footsteps, clothing rustle, squeaky doors and breaking glass, added to films and videos in post-production. It is named after Jack Foley, who joined Universal in 1914 and helped add sound to the studio's early "talkies" (Wikipedia). If you score short films, ads or YouTube videos, sooner or later a director will ask you for "the sound too". This guide covers the parts a musician can do well at home.

Why video needs Foley at all

Sets and props are usually built from cheap, light materials that look right but don't sound right, and on-set microphones are aimed at dialogue. Fight scenes are staged, so no blow ever lands. Foley puts back the sounds the camera missed, and gives the editor full control over each sound's quality and level (Wikipedia). Done well, nobody notices it; without it, a scene feels unnaturally quiet.

The three categories

Foley work is usually split three ways (Wikipedia):

Recording props

Classic tricks still work with a single mic and a quiet room. Cornstarch in a leather pouch makes snow crunch, a pair of gloves flapped together sounds like bird wings, a thin stick gives a whoosh, frozen romaine lettuce snaps like bone, and balled-up quarter-inch audio tape becomes grass underfoot. Walnuts stand in for ice cubes in a glass because they don't melt (Wikipedia).

Practical habits we'd add: record close and dry, so you can add the room later in the mix; record each sound several times so you have variations; and kill fridges and fans first, because a home mic will hear them. If you are choosing a mic for this, our microphone types guide covers the trade-offs.

Sync: perform to picture

Jack Foley's crew projected the film and performed effects live onto a single track, so their timing had to be exact (Wikipedia). The method hasn't changed much: Foley artists watch the scene, rehearse, then record while the picture plays, much as actors re-record dialogue to picture. What has changed is that each sound can now go on its own track and be nudged into place afterwards.

For a musician, that means importing the video into your DAW, locking the session to the video's frame rate, and performing footsteps and moves against it in passes: one pass per character, one per surface. Perform first, then edit. Hand-placing every footstep on the grid sounds mechanical in exactly the way a quantised drum part does.

Layering

Big sounds are rarely one recording. Ben Burtt built the Star Wars lightsaber hum from a film projector motor mixed with a TV picture tube's hum, and the TIE fighter from slowed elephant calls and cars driving through water (Wikipedia). A useful home rule: one layer for the transient (the attack you see), one for the body, one for the tail or space. Reverb also tells the viewer how big a room is: a large hall rings, a small room barely does (Wikipedia), so match the reverb to the shot, not to taste.

Libraries versus recording

Even on professional films, "specifics" like doors and doorbells are often done more efficiently with stock effects arranged by an editor (Wikipedia). Our split: use libraries for generic, unsynced sounds (traffic, room tone, doors, phones) and record anything performed by a person on screen, because footsteps and cloth have to follow a specific actor's timing and a library can't. Read every library's licence before delivery: free libraries mix licences, and some require attribution in the credits. For how licences work in sync deals, see sync licensing basics.

Loudness for online delivery

Here is where musicians most often get it wrong, by mastering a film mix like a single.

• Broadcast follows standards. EBU R 128 targets an integrated loudness of -23 LUFS, measured over the whole programme, with a tolerance of ±0.5 LU and a maximum true peak of -1 dBTP. Individual broadcasters add their own rules, such as ORF's -3 dBTP limit for data-reduced formats (Wikipedia).
• YouTube publishes no upload target. Its Help Centre says Stable volume is on by default and continuously narrows the range between quiet and loud, but that it is off for YouTube Music and official music videos (YouTube Help). Any "YouTube LUFS" number you see is a third-party measurement; our streaming loudness targets guide sorts the documented numbers from the folklore.

If a client or platform hands you a delivery spec, it wins over any general rule. If not, see what LUFS is and measure the whole piece, not just the loudest scene.

What we think

Musicians are better Foley artists than they assume. Footsteps are rhythm, cloth is texture, and layering an impact is the same skill as layering a kick. The trap is the habits that come with the music side: crushing a mix to sound loud next to other tracks. Video is a dynamic medium, and a dialogue-led short film mastered hot will sound tiring and gain nothing once a platform or broadcaster turns it down. Our default with no spec: mix for dialogue clarity, keep true peaks at or below -1 dBTP, and leave the loudness decision to whoever delivers the final file. And record the human sounds yourself. Libraries are fine for doors, but a footstep from a library never walks like your actor.

FAQ

What is Foley? Everyday sounds recreated and synced to picture in post-production, named after Jack Foley.

Do I need a special studio? No. A quiet room, one mic, a few props and a DAW that plays video are enough for short-form work.

What loudness should I deliver for YouTube? YouTube publishes no target. Follow the client's spec if there is one; otherwise avoid clipping, keep true peaks at or below -1 dBTP, and don't over-compress.

Should I use sound libraries? For generic, unsynced sounds, yes. Check each licence, and perform the sounds that follow an actor yourself.