Captioning Music and Sound Effects So Deaf Viewers Can Follow

Watch a thriller with the sound off and the captions on, and count how often the screen goes blank while something important happens. A floorboard creaks behind the lead. A phone buzzes on the table. The score drops to a single low note. For a deaf viewer, a caption track that only transcribes dialogue turns all of that into silence, and the scene stops making sense.

Captioning music and sound effects is the part of the job most creators skip, and it is the part that separates a transcript from real captions. The good news is that it follows a small set of rules, and once you learn them, adding non-speech information to a ten minute video takes about fifteen extra minutes, not an extra afternoon.

What Counts as Non-Speech Information

Professional caption guidelines group everything that is not spoken dialogue under one label: non-speech information. In practice that covers four kinds of sound.

• Sound effects that drive the story, such as a knock, a gunshot, a car pulling up or a notification chime
• Music, both the score and any song a character can hear
• Speaker manner, such as whispering, shouting or a sarcastic tone the words alone do not show
• Off-screen voices and sounds whose source is not visible

The test for each one is simple. If a hearing viewer would react to it, or would understand the scene differently without it, it belongs in the captions. If it is background texture, like general traffic in a city scene, it usually does not.

Formatting Sound Effect Captions

The widely used convention is square brackets around a short description, written in lowercase or sentence case: [door slams], [phone buzzing], [distant thunder]. Keep the description to two or three words where you can. Viewers read captions in under two seconds, and our guide to caption reading speed and line length explains why a long bracket steals time from the dialogue that follows.

Describe the sound, not the picture

Write [glass shatters], not [the vase falls off the shelf]. The viewer can see the vase. What they cannot get is the sound and how loud or sudden it was.

Use precise verbs

[Footsteps] tells a viewer very little. [Footsteps approaching] or [slow footsteps upstairs] tells them where the threat is. One well chosen word does more than a sentence of description.

Place it where it happens

Time the caption to the start of the sound, not a second before. A [gunshot] that appears early spoils the jump for deaf viewers in a way the audio never would for hearing ones.

Captioning Music Without Overdoing It

Music is where caption tracks most often go wrong, in both directions. Some ignore it completely. Others fill every scene with [music playing], which tells the viewer nothing.

The better approach is to describe what the music does. Compare these:

For songs with lyrics that matter, caption the lyrics themselves and mark them with the eighth note symbol at the start and end of each line, which is the standard broadcasters use. If the lyrics are background filler under a montage, a single [soft acoustic ballad playing] at the start of the cue is enough. Re-describe the music only when it changes, swells or stops.

When You Should Leave Sounds Out

A lot of accessibility advice tells you to caption every sound you can hear. I think that advice causes real harm, and I push back on it with every editor I train.

A cooking video with [sizzling], [knife chopping], [pan clatters] and [oven beeps] stacked across thirty seconds buries the presenter's instructions under a pile of brackets. Deaf viewers have told me again and again that over-captioned video is more tiring to watch than under-captioned video, because they cannot tell which sounds actually matter.

Caption the sound that changes what the viewer understands. In that cooking clip, [oven timer beeps] matters because the presenter turns and says "right, it's ready". The sizzle does not, because the picture already shows it. A useful limit is no more than one sound caption per four or five seconds unless the scene genuinely depends on sound.

Speaker Manner and Off-Screen Sound

Tone is easy to forget because the words look complete on the page. "Great, thanks" can be warm or furious. When the delivery flips the meaning, add a short tag before the line: [sarcastically] Great, thanks. Use it sparingly, perhaps three or four times in a ten minute piece. If every line has a tag, none of them stand out.

For off-screen voices, name the speaker if the viewer knows who it is, and add a location if it matters: [Mum, from downstairs] Dinner's ready. On podcasts and interviews where several people talk over each other, the same naming habits apply, and we cover them in detail in labelling speakers in captions for a video podcast.

A Fifteen Minute Workflow for Any Video

Most editors add sound captions after the dialogue pass, which is the right order. This routine works in any caption editor that shows a waveform.

1. Finish and proofread the dialogue captions first.
2. Mute the picture and play the audio only at normal speed. Note every sound that made you look up.
3. Now watch the picture with audio muted. Note every moment where the screen action implies a sound you cannot see explained.
4. Compare the two lists. Sounds on both lists are your captions. Sounds on only one list usually are not.
5. Add the brackets, aligning each one to the start of the waveform spike.
6. Watch the full video once with sound off and captions on, the way your deaf audience will.

On a ten minute vlog, that usually produces eight to fifteen sound captions. On a short horror film, it can be fifty or more, and that is fine because the genre lives on sound.

Frequently Asked Questions

Do auto-generated captions include sound effects?

Some platforms now add basic tags such as [music] or [applause], but they are generic and often mistimed. Treat them as a starting point you will rewrite, never as finished non-speech captions.

Should sound captions be in capitals?

Older broadcast styles used all capitals, but sentence case or lowercase inside brackets is easier to read and is now more common. Pick one style and keep it consistent across your channel.

How do I caption silence?

Only caption silence when it is dramatic, for example [music stops] or [silence] after a loud scene. A quiet moment that is not meaningful does not need a caption at all.

The Practical Takeaway

Pick one video you have already published, run the mute test from the workflow above, and add only the sounds that appear on both lists. Keep each bracket under four words, describe what the music does rather than that it exists, and tag tone only when it flips the meaning. That single pass will make your captions useful to the viewers who rely on them most.