Closed captions vs subtitles vs SDH, and how to make each

What is the difference between subtitles and closed captions?

Who they are for, and so what they carry. Closed captioning assumes the viewer cannot hear the soundtrack, so it writes down the sounds that matter as well as the words. Subtitles assume the viewer can hear and only need the words.

The words differ by country. In the US, closed captions are the accessibility version and subtitles are for translation. In the UK, "subtitles" covers both: the BBC's subtitle guidelines describe closed subtitles, "also known as closed captions".

What does SDH mean in subtitles?

SDH stands for subtitles for the deaf and hard of hearing: subtitles in the same language as the audio that also say who is speaking and what the sound tells you. Netflix's English timed text guide defines them as intralingual subtitles, and writes speaker IDs and sounds in square brackets, in lowercase:

• [phone ringing]
• [applause, cheering]
• [soft music]

The BBC's guidelines do it differently: a sound effect is typed in white capitals on a line of its own, with no brackets, as subject and verb: FLOORBOARDS CREAK. Either way, only the sounds that matter go in: a laugh you can see on a face needs no label.

Are SDH subtitles the same as closed captions?

In content, nearly; in form, no. Closed captions in the US broadcast sense are 608 or 708 data carried with the picture and drawn by the television's own decoder, 608 in a fixed grid 32 columns wide. SDH is a subtitle track or file the player draws in its own style, so it travels where 608 cannot, such as a streaming service's timed-text delivery. A player lists them separately: "English (US)", "English (US) SDH" and "English CC".

How do you make subtitles, closed captions and SDH in Sharpy?

From one transcript, in a few choices:

1. Subtitle tracks, automatically. Every movie Sharpy makes carries the words as a subtitle track for each language, tagged so the player's subtitle menu works and off until the viewer turns it on. A talking head with nothing on the picture still gets one, made from every word through its cuts.
2. Subtitle files. In the window's export panel, tick the captions box for SRT and WebVTT files beside the video. What is an SRT file explains both formats.
3. Closed captions. Ask for closed captions and the first language also goes in as US broadcast 608 captions, pop-on along the bottom rows, with an SCC file beside the movie. Every caption is fitted to 608's 4 rows of 32 columns first, and a character 608 cannot carry, such as an emoji, stops the render before its first frame, named.
4. SDH. Ask for SDH and the first language is written twice in one subtitle choice, as it is and with its sounds. The 608 captions carry the sounds too.
5. Other languages. Each caption is translated whole on your Mac, re-broken into readable lines and kept on its timing, as its own track plus SRT and WebVTT files, labelled as machine translation because it is: "Pin your plugins" once came back in Spanish as "catch your plugins".

Ask the agent in Claude or Codex for any of these when it renders. Every render reads its caption tracks back and fails if the words do not match.

How does Sharpy know which sounds to write?

It listens, and writes nothing it has not measured. Sharpy hears laughter, applause, cheering, music, a phone, a doorbell, a knock, a siren, a dog, a door slam, a car horn and a faint music bed under a voice. Each sound has its own tested rule, set so it claims nothing on recordings of speech alone; under a speaking voice, it heard 28 of 36 planted laughs.

Placement follows the craft. A sound sits where no words are on screen, or as a line above a one-line caption while it plays ("[phone ringing]" over "Who could that be?"). Music Sharpy added is described by its register, such as [upbeat music], and whooshes are left out, as a captioner leaves out texture.

Pick the guide you deliver to and Sharpy writes sounds its way: lowercase in brackets for Netflix and DCMP, capitals on their own line for the BBC. Prime Video's and ESIST's guides say nothing on this, so Sharpy uses its own style and says so.

Which subtitle guidelines should you follow?

The receiver's. Each broadcaster and streamer publishes its own, and they differ:

Sharpy holds an SRT or WebVTT file to the published rules of Netflix, Prime Video, the BBC, DCMP or ESIST, through check_captions, and refuses until no rule is broken. For each caption it says what it measured, what is allowed and whose rule it is; a rule a guide gives no number for is reported as not checked, never borrowed from another guide. Give it the video too and it checks each caption against the picture's cuts: one that starts or ends within half a second of a cut is told to sit on the cut, ending two frames before it, as Netflix's timing guide asks.

What about viewers who cannot see the picture?

They need audio description: a voice that says what the picture shows in the gaps between the words. Sharpy writes the script as WebVTT, in the present tense, paced at or under 160 words a minute, saying what is seen and nothing the sound already says, and labels it a machine's draft for you to read. A typical line: "A man sits at a desk in a room. He wears glasses and a brown shirt."

The agent's narrate tool voices it in Sharpy's own voice, dips the programme under each line by broadcast rules, and writes a movie with the original picture and sound untouched plus a second audio track tagged as audio description.

Frequently asked questions

01

What is the difference between subtitles and closed captions?

Subtitles assume the viewer can hear and give only the words. Closed captions assume the viewer cannot, so they add the sounds and who is speaking, and they can be switched off.

02

What does SDH mean in subtitles?

Subtitles for the deaf and hard of hearing: same-language subtitles that include speaker names and sounds, written in a subtitle format rather than as broadcast captions.

03

Are SDH subtitles the same as closed captions?

They carry much the same information but in a different form. Closed captions are broadcast data in a fixed style; SDH is a subtitle track or file the player styles.

04

What are open captions?

Captions burned into the picture, so nobody can switch them off. Animated captions are open captions.

05

Do I need closed captions for YouTube?

No. YouTube reads subtitle files such as SRT and WebVTT, and lists SCC files of 608 captions among its preferred broadcast formats.

Sources

Checked on 18 October 2026.

• Netflix: English (USA) Timed Text Style Guide
• Netflix: Timed Text Style Guide, General Requirements
• Netflix: Timed Text Style Guide, Subtitle Timing Guidelines
• BBC Subtitle Guidelines, version 1.2.5, March 2026
• DCMP Captioning Key: Sound Effects and Music
• ESIST: Code of Good Subtitling Practice
• Prime Video: captions and timed text
• YouTube Help: supported subtitle and closed caption files