When sound and image disagree

A calm method for noticing moments where what you hear does not match what you see, and describing that mismatch clearly instead of guessing at intent.

Published by Cutaway Notes editorial desk

Entry checked on 21/09/20265 min read

A quiet screening room where sound and picture can be compared side by side. Illustrative image, generated with AI.

What does it mean when sound and image disagree?

It means the soundtrack and the picture are telling you different things at the same moment. That is all it means at first. The gap might be deliberate, accidental, or simply a technical artifact you noticed, and you usually cannot tell which from inside a single viewing. The useful move is to name the mismatch precisely, then hold off on the explanation.

Mismatch is common in short documentary and independent work, where sync problems, room tone, archival audio and voice-over can layer over each other. Noticing it is not the same as diagnosing it.

How do you notice a mismatch without inventing a story about it?

Watch and listen as two separate streams rather than one blended experience. Ask: is the sound arriving at the moment I would expect from the action? Is the voice coming from a body I can see? Does the ambient room match the space shown?

A simple discipline helps. Describe only what is observable. "The footsteps continue after the walker stops" is observable. "The filmmaker is lying to me" is a conclusion. Staying on the observable side keeps the moment open long enough to be useful, and it pairs well with the habit described in Noticing without taking notes.

One framing worth borrowing comes from the National Film Board's classroom guidance, which suggests discussion questions such as "Whose voices are missing?" rather than questions that invite students to guess at intention. You can read the full set of tips at the NFB education blog.

What kinds of mismatch are worth naming?

Most mismatches fall into a few recognizable families. Sorting the moment into one of them gives you vocabulary without pretending to know the reason behind it.

What you notice	What it probably involves	What to say out loud
Sound continues after the visible action stops	Sync or edit decision	"The sound outlasts the picture here"
A voice has no visible source	Voice-over, phone call, or off-screen speaker	"The voice is not attached to a body on screen"
Room tone does not match the space	Recorded separately or replaced later	"This room does not sound like that room"
Music contradicts the mood of the image	Deliberate counterpoint, or reused track	"The music sits against the picture"
A repeated sound returns at a different volume	Mix choice or level drift	"Same sound, different level the second time"

This is a description table, not a diagnosis table. It gives you a sentence you can actually say in a group without overreaching.

Why does describing the mismatch beat interpreting it?

Interpretation closes the moment. Description keeps it available for other people to check. If you announce that a sound was added to manipulate you, and someone else heard it as an on-location recording, the conversation stalls on a disagreement you cannot resolve.

If instead you say what you heard and when, others can confirm or contradict the observation itself, which is a much more productive disagreement. This is the same reason a question about a film can miss what is actually there: the companion piece on missed questions walks through how a framing question can slide past the material. The NFB's suggestion to allow processing time before discussion supports the same patience.

Does a mismatch always mean something went wrong?

No. Separated sound is a normal part of how films are built, and much of what reaches your ears was recorded away from the picture. The NFB film Listen, a portrait of composer R. Murray Schafer directed by David New and produced for the 2009 Governor General's Performing Arts Award, is built around the idea of the soundscape as a collection of sounds, and the film's own production credits list separate sound recording, sound design and mix roles, which is a useful reminder that sound is assembled through several stages. The film is listed at nfb.ca/film/listen.

What that means practically: a mismatch is a fact about the track, not a verdict on the work. Treat the floor as neutral. For a broader vocabulary of how sound is organised, see Music, noise and silence.

How do you raise a mismatch in conversation without derailing it?

Ask it as an observation with a question attached. "Around the middle, the traffic noise kept going after the shot changed. Did anyone else hear that, or was it just me?" That phrasing invites confirmation, correction, or a shared shrug, all of which are fine outcomes.

Two habits help. First, let others answer before you explain what you think it meant. Second, if someone supplies a technical explanation you cannot verify, accept it as a possibility rather than a fact. A neutral reporting style also makes mismatches easier to compare later, which is the spirit behind Two viewings, one question.

A short decision checklist for the moment you notice something:

State the observation in one sentence, without a motive.
Give the rough location in the film, not an exact timecode you cannot check.
Ask whether anyone else registered it.
Accept the answer, including "no."
Write nothing down as a conclusion.

When should you just let the mismatch go?

Most of the time. A single odd sound cue rarely carries the weight of the whole film, and chasing every one turns viewing into auditing. If the mismatch did not change your attention or your feelings, it is fine to let it pass.

Save the energy for the mismatches that genuinely shifted something: the moment you stopped trusting the scene, the moment the space stopped feeling like the space, the moment a voice seemed to come from nowhere. Those are the ones worth a sentence. A calm, precise sentence about a sound you heard is a better contribution than a confident theory about a sound designer's motives, and it leaves the next viewer free to check you.

How does this connect to a wider viewing habit?

Treating sound and image as two related but separate tracks is a small adjustment with a large effect. It slows down the instinct to explain, and it makes your observations easier for other people to trust. Start with one film, one mismatch, one sentence. The method does not require equipment or expertise, only the willingness to describe before you interpret.

If you want to build the habit around a question rather than a technique, One question before you press play sets out a starting point that works alongside this piece.