What is sound design in film? Who designs, edits and mixes a soundtrack, the layers it's built from, how the lightsaber was made, and what Atmos means.
Method. Checked November 18, 2026 against ScreenSkills job profiles, Netflix's Studio Partner sound mix specification, Dolby's professional Atmos pages, Walter Murch's Transom.org essay, interviews with Murch and Ben Burtt reproduced on FilmSound.org, and the Academy's 1980 awards record; where only an encyclopaedic source was available we name the books it cites. Vinc makes an AI film studio.
Looking for on-set recording and mix terms (boom, ADR, LUFS)? See [Film Sound Terms](/blog/film-terms-sound). For the types of effects with recordings, see [Types of Sound Effects](/blog/sound-effects-explained).
Design, editing and mixing are three different jobs
Sound design in film means creating the sounds that can't simply be recorded and, in its broader meaning, shaping the whole soundtrack. Sound post splits into three jobs: designing sounds that don't exist yet, editing recorded and designed sounds into sync with the picture, and mixing all of it into one balanced soundtrack. On a small film one person may do all three; on a large one each is a department. Terms like ADR, walla, stems and M&E are defined in Film Terms, Visualised: Sound.
The supervising sound editor
ScreenSkills, the UK screen skills body, describes supervising sound editors as the people who "manage the team that looks after each part of the sound of a film or TV drama". On lower-budget films they start at picture lock; on bigger ones, before shooting.
After picture lock the supervisor runs a spotting session with the director, deciding the overall feel ("naturalistic or stylised") and going through every effect and line to see what's needed. They hold the budget, run the workflow "from sound editorial, Foley recording, ADR sessions, pre-mix to final mix", and oversee deliverables such as the M&E.
Sound editors: dialogue, effects, Foley
Sound editors, "sometimes known as track layers because of the task they perform" (ScreenSkills), work through the film sequence by sequence, laying dialogue, effects, Foley and music in sync and filling gaps so the mixer can make smooth transitions.
On larger films the work splits by layer. Dialogue editors make location dialogue "more understandable, as well as smoother"; effects editors draw on sound-effects libraries; Foley artists perform sounds to picture. ScreenSkills gives the textbook substitutions: a guillotine is a sliced cabbage, breaking bones are snapped celery.
The re-recording mixer
The re-recording mixer, called a dubbing mixer in the UK and Europe, "brings together all the different sound elements of a production to create the final mix", in ScreenSkills' words. The trade-off: "if the effects are too loud you might not be able to understand the dialogue; if they are too quiet the film may lose much of the excitement."
The traditional mix starts with premixes: each family of sound is balanced on its own, then the premixes are combined in the final mix.
The sound designer
The sound designer creates sounds that can't simply be recorded. Ben Burtt, who made the sounds of Star Wars (1977), explained why he used the title: "I called myself a sound designer because I wasn't really functioning as a production recordist." As he put it, "my job was to coordinate all you heard in the final soundtrack of the film."
Randy Thom, on FilmSound.org, says the term "has come to be associated simply with using specialized equipment to make 'special' sound effects", narrower than that original whole-soundtrack meaning.
A sound that couldn't simply be recorded. Its maker says it is a rubber belt rubbed on guitar strings, then "low pitched, distorted, inverted" and given reverb."GenericUnknownTypeMonsterRoar" by Ovkovko, Freesound, CC0.
The order of work ScreenSkills describes for a supervising sound editor: spotting after picture lock, editorial, Foley and ADR, premixes, final mix, then deliverables such as the M&E.
Who coined "sound designer"
The title is usually traced to Walter Murch and Francis Ford Coppola, in two accounts. Wikipedia's Murch article says that "for his work on Apocalypse Now, Murch was the first person to receive a credit as 'Sound Designer'" (unfootnoted, though it cites Caldwell's Production Culture for the title being "a professional designation that Murch is widely attributed to have coined"). Its "Sound design" article says Coppola first gave Murch the title for The Rain People (1969), citing Ondaatje's The Conversations (2003, p. 53).
The official record: at the 52nd Academy Awards (1980) the Sound Oscar for Apocalypse Now went to Walter Murch, Mark Berger, Richard Beggs and Nat Boxer.
The layers of a soundtrack
A finished soundtrack is built from a handful of families of sound. Netflix's sound mix specification says that in Atmos home mixes for series, beds and objects "must be separated and groups labeled as 'Dialog', 'Narration', 'Group ADR', 'Music', 'Effects', 'Backgrounds', or 'Foley'". For films, "DME separation must be maintained whenever possible" (dialogue, music and effects kept apart).
Layer
What it contains
Where it usually comes from
Dialog
Production dialogue and ADR
Location recording, ADR stage
Narration
Voice-over
Recording booth
Group ADR
Crowd voices, walla
Loop group in post
Music
Score and licensed songs
Composer, music supervisor
Effects (hard effects)
Specific sync sounds: doors, guns, cars, impacts, designed sounds
Effects editors, libraries, sound designers
Backgrounds
Ambience and room tone that set the place
Location recordings, libraries
Foley
Footsteps, cloth, props performed to picture
Foley stage
Group labels from Netflix Studio Partner's Sound Mix Specifications; the "contains" and "comes from" columns summarise ScreenSkills' role descriptions.
Our illustration of building a scene in layers: someone walks up a street, finds their keys and goes in. Each version adds one family at the same mix level, so the only change is the new layer.Mix by Vinc of recordings by Liljendal, FoleyMakers, TRP and hisoul, Freesound, CC0.
Two backgrounds, also called ambience or atmos. Neither has a sound you'd sync to picture; each tells you where you are.Recordings by Liljendal and NoBody112, Freesound, CC0.
Each layer becomes a stem at the end. Without the dialogue stem, the rest becomes the M&E for foreign-language dubs, and a separate music stem lets a cue be replaced without remixing (licences: Music Licensing Explained).
What a real premix looks like
A premix is one family of tracks balanced into a single master before the final mix. In "Dense Clarity – Clear Density" (Transom.org), Murch describes the Apocalypse Now helicopter attack, which had "over a hundred and seventy-five separate soundtracks", grouped into "six premixes of film, each one a six-channel master", listed "in order of importance":
Premix
Order of importance
Dialogue
1
Helicopters
2
Music (The Valkyries)
3
Small arms fire
4
Explosions
5
Footsteps and other Foley-type sounds
6
How many layers can an audience follow
Murch's rule of thumb came from footsteps on THX 1138: "our minds can keep track of one person's footsteps, or even the footsteps of two people, but with three or more people our minds just give up". With three robots walking, "nothing had to be in sync." He called it his "Law of Two-and-a-half": a layer stays clear when there are "not quite three layers of something." The same thinking runs through Walter Murch's Rule of Six.
Perspective: where the listener is
Sound perspective is how close or far a sound seems, and from whose position the audience hears it. A whisper in close-up should sound close; across a car park in a wide, it usually shouldn't. The main lever is the balance of direct sound to room reflections.
Randy Thom argues it is also a storytelling tool: "what they hear can give us lots of information about who they are and what they are feeling." That is the case for subjective sound, such as muffling the world after an explosion.
A J-cut: the next scene's sound arrives before its picture, so the audience hears the new space before they see it.
Worldizing
Worldizing is playing a clean recording through a loudspeaker in a real location and re-recording it, so it picks up that place's acoustics. FilmSound.org defines it as "manipulating sound until it seemed to be something that existed in real space" and credits Walter Murch with the concept.
The standard example is the radio in American Graffiti (1973). Murch "played it back on a Nagra in a real space - a suburban backyard," recording from "fifty-or-so-feet away" while moving the microphone. Blending dry and worldized tracks let him set the dry-to-reflected balance shot by shot: "This creates the sonic equivalent of depth of field in photography."
Our illustration of perspective through reverb. We ran the dry steps through an impulse response, a recording of how a real building hall answers a sharp sound, and mixed it with the dry signal. Listen to the tail after the last step.Footsteps by FoleyMakers; hall impulse response by Hal9000; both Freesound, CC0. Processing by Vinc.
Sound design examples: how famous sounds were made
The best-documented origins come from the original Star Wars, as FilmSound.org summarises Ben Burtt's work:
Sound
How it was made, as documented
Lightsaber
"Burtt blended the sounds of his TV set and an old 35 mm projector"
Blaster
"The sound of a hammer on an antenna tower guy wire"
R2-D2
"50 % of the droid's voice is generated electronically; the rest is a combination and blending of water pipes, whistles, and vocalizations by Burtt"
Chewbacca
"Wookie sounds are constructed out of pieces of walruses and other animal sounds"
TIE fighter
"a drastically altered elephant bellow"
The lightsaber, step by step
The lightsaber was built in three steps that make a good model for any designed sound, as Burtt describes them in a laserdisc interview:
A base tone. USC projectors' interlock motors made a "wonderful humming sound."
The missing texture. "What was missing was a buzzy sort of sparkling sound," which he found in a television set.
Movement. Moving a microphone past a speaker playing the sound, "you get a pitch shift in the sound" (the Doppler effect), worldizing again.
The Wilhelm scream
The Wilhelm scream is a stock scream effect that sound editors keep reusing as an in-joke. Wikipedia, citing Wired, The Times and Rinzler's The Sounds of Star Wars, says it "originates from a series of sound effects recorded for the 1951 film Distant Drums". Burtt "named the scream after Pvt. Wilhelm, a minor character from The Charge at Feather River" (1953) and used it in Star Wars.
The scream as it sits on the session reel, "S38-01 Man eaten by alligator; screams". Craig Smith restored it from the Sunset Editorial collection held at USC and published the whole session on Freesound in 2023.Recording via craigsmith (USC / Sunset Editorial collection), Freesound, CC0.
Object-based mixing: Dolby Atmos
A channel-based mix sends each sound to a fixed speaker. An object-based mix stores each sound with position metadata, and the playback system picks the speakers. Dolby Atmos combines a channel-based "bed" with objects placed anywhere in the room.
Channels in miniature, best on headphones. The same street recording in stereo, then folded to mono by us. In stereo the passing cars move between left and right; in mono everything comes from one point."Day Traffic - City Life" by Liljendal, Freesound, CC0.
Dolby's published figures:
Spec
Figure
Source
Total audio tracks (cinema)
Up to 128
Dolby Professional, Dolby Atmos for cinema
Bed (cinema)
9.1 channel-based bed
Dolby Professional, Dolby Atmos for cinema
Objects (cinema)
Up to 118
Dolby Professional, Dolby Atmos for cinema
Speakers in a theatre
Up to 64, each powered independently
Dolby Professional, Dolby Atmos for cinema
Common bed for content creation
7.1.2 (7.1 plus two overhead channels)
Dolby Professional, Dolby Atmos content creation
Dolby Atmos Renderer
"up to 128 inputs (including audio beds and objects with metadata)"
Dolby Professional, Dolby Atmos Renderer
A 10-channel bed (9.1) plus 118 objects is 128 tracks.
For streaming, Netflix asks for the Atmos home mix to be made "in a 7.1.4 room at minimum", hitting its −27 LKFS dialogue-gated loudness target, with a "Dolby Atmos BWAV ADM file" print master at 48 kHz, 24-bit.
An Atmos home-mix chain as Dolby and Netflix describe it: beds and objects rendered by the Dolby Atmos Renderer, exported as an ADM BWF master, then wrapped into IMF for streaming.
Applying this on an indie or AI-assisted project
Hold a spotting session, even if it's just you. List, scene by scene, what each layer needs.
Keep the layers apart from day one. Name tracks by family, the way Netflix labels them, so stems and an M&E come free.
Premix before you final-mix, then decide which family leads each scene.
Respect two-and-a-half. Sync the first two of anything carefully and let the rest blend.
Give perspective a real room. Worldize key sounds.
For generated audio, ask for layers separately. Google's Veo prompt guide, quoted in our sound glossary, recommends describing audio "in separate sentences".
How we compiled this
Roles come from ScreenSkills' job profiles and delivery language from Netflix's sound mix specification, read on November 18, 2026; Atmos figures are from Dolby. Historical claims come from Murch's and Burtt's own words on Transom.org and FilmSound.org, and where only Wikipedia was available we name the sources it cites. The practical checklist is ours.
Aspect Ratios in Film Explained: 1.33 to 2.39 and 9:16, With the History and Delivery Specs
Aspect ratios in film are width divided by height. See where 1.375, 1.85, 2.39, 16:9, IMAX and 9:16 came from, and the containers cinemas and Netflix expect.