Video Captioning and Subtitles: A Practical Guide

Captions turn the speech and meaningful sounds in a video into synchronized text so people can follow it without hearing it. Subtitles translate the dialogue for viewers who can hear but do not speak the language. Most U.S. organizations need captions first and subtitles second. This guide covers the difference, the formats, how to add both, and what makes a caption track good enough to publish.

Glenwood Media

Make Your Videos Work Harder

Glenwood Media creates videos and photography that help organizations tell their stories and strengthen their digital presence.

Explore Glenwood Media

TL;DR

Captions include dialogue plus important non-speech sounds. Subtitles usually translate dialogue only, so they are not a substitute for captions.
Closed captions can be switched on and off. Open captions are burned into the picture and always show.
Never publish raw auto-captions. Edit them for accuracy, speaker labels and timing before the video goes live.
Keep to two lines and about 45 characters per line, and give viewers time to read.
U.S. rules vary by organization. Section 508 applies to federal agencies, and ADA exposure depends on who you are and what you publish.

Captions vs. subtitles: what each one does
Open captions, closed captions and the "CC" button
Caption file formats: SRT, WebVTT and the rest
How to add captions to a video
Captioning best practices: accuracy, timing and readability
Captions, subtitles and search visibility
What U.S. accessibility rules say about captions
Our perspective: treat captions as part of the shoot
Captioned video from Glenwood Media
FAQ
Sources

Captions vs. subtitles: what each one does

The two words get swapped constantly, and in some countries they mean the same thing. In the U.S. accessibility world they do not.

The W3C Web Accessibility Initiative describes captions as synchronized text for spoken dialogue plus important non-speech audio, made so a video can be understood without hearing it. Subtitles, in the same guidance, generally translate the dialogue into another language for viewers who can hear but do not understand what is being said.

That difference shows up in what the text contains. Section508.gov notes that subtitles usually do not indicate important audio information and rarely identify speakers or nonverbal sounds. It also says they are not an acceptable method for meeting its synchronized-media requirements.

Here is the comparison at a glance.

	Captions	Subtitles
Main audience	Deaf and hard-of-hearing viewers, plus anyone watching on mute	Viewers who do not understand the spoken language
Language	Same as the audio	Usually a translation
Non-speech audio (music, door slam, laughter)	Included	Usually left out
Speaker identification	Included when it is not obvious	Rare
Meets U.S. accessibility requirements	Yes, when accurate and synchronized	Not on their own

A practical rule: if the goal is access, make captions. If the goal is reach into another language market, add subtitles on top of the captions.

Open captions, closed captions and the "CC" button

The phrase "closed captions" comes from the fact that the text is closed in a separate file or data stream until a viewer turns it on. That is the CC icon you see in a player.

There are two delivery types, and the Section 508 site defines both: closed captions are separate text viewers can toggle, and open captions are permanently displayed.

The Federal Agencies Digital Guidelines Initiative adds the production consequences in its accessibility definitions:

Closed captions are stored separately, so they are comparatively easy to edit after publication.
Open captions are embedded in the video and are not easily editable. They are useful when a platform cannot display caption files or when viewers cannot practically turn captions on, but they cannot be turned off.

Which one should you choose?

Choose closed captions as the default. They let viewers control the experience, they can be corrected without re-exporting video, and one video can carry several languages.

Choose open captions when you do not control the player. Think of a looping screen in a lobby, a video played in a conference hall with the sound off, or a social platform where you want the words guaranteed on screen. If you plan a muted loop for a venue, our guide to planning a walk-in loop video covers how to build one, and open captions often belong in it.

When in doubt, deliver both: a clean master with closed-caption files, and a second export with open captions for social and screens.

Caption file formats: SRT, WebVTT and the rest

A caption file is a plain text file with the words and the times they appear. YouTube Help says caption files contain text and display timestamps, and some formats also carry position and style data that helps deaf and hard-of-hearing viewers.

FADGI lists the common sidecar options: WebVTT (.vtt), SubRip (.srt), or XML formats such as TTML, subject to what the media player supports.

In practice:

SRT is the simplest and the most widely accepted. Times and text, nothing else. It is the usual hand-off format.
WebVTT is the format used by HTML5 video players, and it supports more positioning and styling. Choose it for video you host on your own site.
TTML and other XML formats show up in broadcast and enterprise workflows.

A question we hear often is whether YouTube takes SRT or VTT. YouTube's help page confirms you can upload caption files, but it does not spell out each extension, so open the upload dialog in Studio to see what your account accepts. Keep an SRT and a VTT of every final track so you are never stuck when a new platform asks.

How to add captions to a video

There are three routes, and the right one depends on where the video lives.

On YouTube

YouTube's instructions, from its help page on adding subtitles and captions, go like this:

Open YouTube Studio and choose Subtitles.
Select the video and add a language.
Choose Add, then upload a timed or untimed file, type the captions manually, or use auto-sync with transcript text in the video's spoken language.
You can also add captions during the upload itself.

Once a track is published you are not locked in. YouTube lets you change caption text and timestamps in Studio, or download the file, edit it in a plain-text editor and upload it again. That makes captions maintainable content: fix a misspelled donor name or a product name next month without re-publishing the video.

In an editing program

In most editors the workflow is the same: create a caption or subtitle track, play through the timeline, and type or import the transcript, then adjust timing so each caption appears with the speech. From there you can export a sidecar file for closed captions, or burn the text into the picture for open captions. Plan the on-screen space too. If you use name titles, our piece on lower thirds design explains how to keep them from colliding with captions in the lower third of the frame.

Caption editing steps and two export options

On your own website

If you host video on your site, attach a WebVTT file to the player and make sure the player exposes a caption control. Our guide to fast, accessible video embeds covers the markup. For autoplaying hero video, see our homepage video best practices, where muted playback makes captions especially important.

A realistic time estimate

Michigan State University's guide says captioning a video yourself typically takes about 5 minutes per minute of video (that guide dates from 2013, and tools have improved since). A 10-minute video is therefore close to an hour of work if you start from nothing. A draft from automatic speech recognition, followed by a human edit, is usually faster, but only if the edit actually happens.

Captioning best practices: accuracy, timing and readability

The Section 508 team's captions and transcripts guidance is the most useful checklist we know of, and it comes from the U.S. government itself. These are the points that matter most.

Accuracy comes first

Captions must be equivalent to the spoken words and other audio information, including sounds necessary to understand or enjoy the content. Speaker labels and sound descriptions should keep a consistent style throughout. A transcript that leaves out sound information is not an equivalent caption track, which is the W3C's point as well.

Automatic captions are a starting point. W3C WAI says they generally need significant editing and do not meet accessibility needs unless confirmed fully accurate. Section508.gov likewise advises organizations not to rely solely on automatic captioning for prerecorded media, and to use scripts and other text materials in the process.

The errors that most often slip through a machine draft:

Proper nouns: donor names, programs, neighborhoods, product names.
Acronyms and industry terms.
Missing speaker changes.
Missing sound cues, like applause or music that sets the mood.
Wrong punctuation, which changes meaning.

Readability

The same Section 508 guidance recommends:

No more than two lines on screen at once.
No more than 45 characters per line.
Captions centered in the lower third, unless that covers important content.
Enough time on screen to read.
Readable type with strong contrast.
Avoiding scrolling, flashing and distracting animation.
Care with speech above 180 words per minute and overlapping voices, where captions become hard to follow.

The 180 words per minute limit is a script-writing tip as much as a caption tip. If a speaker is rushing, the cleanest fix is often at the shoot or in the edit, not in the caption file.

A quick quality check before you publish

Watch the whole video with sound off and read only the captions. Does it make sense?
Check every name against a written source.
Confirm that music and key sound effects are described.
Look at the longest captions. Are any three lines, or off the screen edge?
Test on a phone, where the text is smallest.

Captions, subtitles and search visibility

Captions give search engines and viewers readable text for content that is otherwise audio. They also keep silent viewers watching, which matters on social feeds where video autoplays muted. If you want to go deeper on how video pages get found, this guide on how to optimize videos for SEO covers presentation and visibility. Treat that as a bonus. Access is the reason to caption, and the SEO benefit comes with it.

On social platforms, specs matter as much as the words. Our LinkedIn video specs checklist shows how to prepare files so captions are not cropped, and our promo video length strategy helps you decide how much text a short video can carry.

What U.S. accessibility rules say about captions

There is no single caption rule that applies identically to every U.S. organization and every video. Who you are and what the video is for decide which rule applies.

Federal agencies and covered technology (Section 508). The regulation text says training and informational video or multimedia that supports an agency's mission, and contains speech or other necessary audio, must be open- or closed-captioned. The standards also require technology that displays synchronized video to process closed captions, and applications to offer caption-selection controls where they offer volume controls, with stated exceptions. The current eCFR text has the details.

Private organizations (ADA Title III). Obligations depend on the organization, the service and the facts. The Department of Justice's settlement with Teachers Test Prep shows the risk: the government treated online courses without closed captions as a Title III issue and required captioning consistent with WCAG 2.0 AA. That settlement binds that organization only, but it shows how enforcement thinks about inaccessible instructional video.

Institutional policy. Many universities and employers set their own bar. UC Berkeley, for example, requires accurate, synchronized captions on public videos posted after December 2, 2022. That is an institutional policy and not a national rule, but if your funder, university partner or corporate client has a similar policy, it becomes your contract requirement.

This is not legal advice. If you are unsure whether your video falls under ADA or a funder's accessibility terms, ask your organization's counsel or accessibility officer, and caption everything in the meantime. It costs far less than a complaint.

What U.S. accessibility rules say about captions — overview diagram

Our perspective: treat captions as part of the shoot

We produce video for nonprofits, corporate teams and event organizers, and the captioning problems we see are rarely about software. They begin earlier.

A fundraising video with a moving story usually has a name, a program title and a number in every other sentence. Those are exactly the words machine captions get wrong, and exactly the words a donor, board member or sponsor will notice. So we write the caption pass into the schedule, not into the last hour before delivery. We collect spellings at the shoot, keep a script or run of show when there is one, and review the track with sound off.

We also build the frame for text. Interviews are composed so the lower third is free, name titles sit apart from caption space, and pacing leaves room to read. A beautifully shot video with captions covering the speaker's name is not finished.

Our advice for any organization: caption every public video, keep a corrected SRT and VTT file with the project archive, and decide up front whether a given export needs open captions. If you are planning a schedule, our video turnaround guide explains where review and caption time fits.

— Bo

Captioned video from Glenwood Media

If you would rather not do the caption pass yourself, we handle it as part of production. Our video production, livestream and photography services cover recap videos, corporate promos and highlight reels, and for live programs our media event coverage is built around the same planning. We do not publish prices, so tell us your video count and where it will be shown, and we will scope it with you.

FAQ

What is the difference between closed captions and subtitles?

Closed captions are a toggleable text track that includes dialogue and important sounds, written for people who cannot hear the audio. Subtitles usually translate the dialogue for viewers who hear it but do not understand the language. Subtitles rarely identify speakers or describe sounds, so they do not replace captions.

Are auto-generated captions good enough?

Not on their own. W3C WAI says automatic captions generally need significant editing, and Section 508 guidance advises against relying only on them for prerecorded media. Use them as a draft, then correct names, terms, speakers and sound cues.

Should I use open or closed captions?

Use closed captions by default, because viewers can turn them off and you can edit them later. Use open captions when the player cannot show a caption file or viewers cannot reasonably turn captions on, such as a muted screen at an event. Open captions cannot be turned off, and they are hard to change once exported.

How many words can a caption hold?

Section508.gov recommends no more than two lines at a time and no more than 45 characters per line. Keep captions centered in the lower third unless they hide something important. Leave each one on screen long enough to read.