Live captions are generated while speech is happening. Closed captions are a text track that viewers can turn on or off. Transcription focuses on creating a written record of spoken content. These terms describe different aspects of text-based speech access, so the same content can sometimes fit more than one category.
The terms live captions, closed captions, subtitles, and transcription are often used as though they describe four completely separate technologies.
The distinction is not always that simple.
Describes when the text is created: while the speech is happening.
Describes whether the viewer can control whether the captions are visible.
Describes the process or result of converting spoken content into a written record.
Often refer to translated dialogue, although terminology varies between countries and platforms.
A live television broadcast, for example, may provide captions as the program happens while still allowing viewers to turn them off. Those captions are both live and closed.
Understanding these differences makes it easier to choose the right type of text support for a video, meeting, phone call, classroom, or face-to-face conversation.
What Are Live Captions?
Live captions convert speech into text while the speech is happening.
They may be created by:
- A professional real-time captioner
- A Communication Access Realtime Translation provider
- Automatic speech-recognition software
- A system combining human and automated processes
Live captions are used for television broadcasts, online meetings, lectures, public events, phone calls, and everyday conversations.
Because the words must appear without waiting for the event to end, live captioning operates under time pressure. The captioner or recognition system has limited opportunity to review the complete sentence, correct names, or reconsider earlier wording.
This can create a balance between speed and accuracy. Text that appears quickly may change as more context becomes available. Text that waits for additional context may be more stable but arrive later.
The Web Accessibility Initiative's guidance on captions explains that live captions can be produced in person or remotely. If a live event is later posted as a recording, its original captions may need editing before the recorded version is published.
Automatic live captions are influenced by audio quality, background noise, accents, speaker distance, overlapping speech, and specialized vocabulary. Our guide to live caption accuracy explains these factors in more detail.
What Are Closed Captions?
Closed captions are captions that viewers can choose to display or hide.
The word closed does not mean the captions are unavailable. It means they exist as a separate selectable track rather than being permanently visible in the video image.
A viewer may turn them on through a television, streaming service, video player, mobile app, or meeting platform. The familiar "CC" symbol is commonly used to indicate that a closed-caption track is available.
Closed captions are different from open captions. Open captions are permanently included in the visible video and cannot be turned off.
The Federal Communications Commission defines closed captioning as a visual display of the audio portion of video programming. Captions may include more than spoken dialogue. They can also identify speakers and represent meaningful sounds.
For example:
MAYA: I thought you had the keys.
[door locks]
[alarm sounding]
Speaker labels and sound descriptions help someone understand what is happening without depending on the audio.
Closed captions may be prepared and edited before a recorded video is released. They may also be generated during a live broadcast. The defining feature is that the user can turn them on or off, not whether they were created live.
What Is Transcription?
Transcription is the process of converting spoken content into written text. The resulting document is called a transcript.
A transcript usually presents a larger, more complete written record than captions shown on a screen. It may be formatted as paragraphs, divided by speaker, organized by topic, or marked with occasional timestamps.
Transcripts are commonly created for:
- Interviews
- Podcasts
- Meetings
- Research recordings
- Legal or medical documentation
- Lectures and webinars
- Customer-service calls
- Recorded videos
Unlike captions, a transcript does not always need to appear at the same moment as the corresponding speech. A reader may use it before, during, or after listening to the original content.
Transcripts can also be edited in different ways. A verbatim transcript attempts to preserve every spoken word, including repetitions and unfinished sentences. A cleaned transcript may remove filler words, correct grammar, or reorganize speech for easier reading.
The Web Accessibility Initiative's transcript guidance notes that transcripts can include headings, links, speaker identification, and important visual information. Timestamps are optional and may be added only where they are useful.
Captions and transcripts can be created from the same text. The main difference is how that text is organized and used. Captions are synchronized with a moment in the audio. A transcript is designed to be read as a document or record.
Where Subtitles Fit In
In common American usage, subtitles usually display a translation of spoken dialogue.
For example, a movie spoken in Spanish might provide English subtitles for viewers who can hear the dialogue but do not understand Spanish.
Traditional subtitles may focus mainly on spoken words. They may not include background sounds, music, tone, or speaker identification because they assume the viewer can hear those parts of the soundtrack.
Captions are generally designed to provide access to the audio itself. The National Institute on Deafness and Other Communication Disorders explains that captions may identify speakers and describe sound effects that are important to understanding the program.
The terminology still varies:
- Some countries and platforms use subtitles for both translated and same-language text.
- Intralingual subtitles use the same language as the spoken audio.
- Interlingual subtitles translate speech into another language.
- Subtitles for the deaf and hard of hearing, often labeled SDH, may include speaker names and meaningful sounds in much the same way as captions.
This means a live captions vs. subtitles comparison depends partly on how the platform uses the terms. The safest approach is to check what information the text includes, whether it is translated, and whether it can be turned off.
Key Differences
These categories overlap, but they emphasize different features.
| Feature | Live Captions | Closed Captions | Transcription | Subtitles |
|---|---|---|---|---|
| Created in real time? | Yes | Sometimes | Either live or afterward | Usually prepared beforehand, but can be live |
| Synchronized with speech? | Yes | Yes | Not always | Usually |
| Includes non-speech sounds? | May include them | Usually includes important sounds | Depends on purpose | Often focuses on dialogue |
| Creates a lasting record? | Not necessarily | Stored with media or delivered as a stream | Usually | Usually stored with media |
| Can be edited? | Limited before display | Often edited for recorded media | Usually | Usually |
| Can the viewer turn it off? | Depends on the system | Yes | Not applicable as a document | Usually, unless open |
| Primary use | Access during current speech | Access to a media soundtrack | Reading, reference, or documentation | Understanding dialogue, often in another language |
The most important distinction is that these labels do not all answer the same question.
- Live answers: When is the text created?
- Closed answers: Can the user control whether it is visible?
- Transcript answers: Is the text intended to function as a written record?
- Subtitle often answers: Is the dialogue being represented or translated for a media viewer?
Which One Is Used in Everyday Conversation?
Everyday conversations generally use live captions.
The purpose is to make speech available as text while people are still talking. The user needs the words during the exchange, not several hours later as a completed transcript.
Examples include:
A conversation may also produce a transcript if the system saves the text. That does not change the original function of the live captions. The captions supported the conversation as it happened, while the saved transcript provides a later record.
Saving conversational transcripts introduces privacy and consent considerations. Users should understand whether audio or text is stored, where it is processed, how long it is retained, and whether everyone involved should be informed.
For important medical, financial, legal, or safety information, live captions should not automatically be treated as an official transcript. Critical details may still need to be confirmed in writing.
How Wearable Captions Differ From Video Captions
Video captions are tied to a defined piece of media. The audio has a timeline, and each caption can be synchronized with a particular section of that timeline.
Wearable captions operate in a less controlled environment. Speech may begin without warning. Speakers can move, interrupt one another, turn away, or change topics. Background noise may also change from one moment to the next.
A wearable caption system therefore needs to:
- Capture speech from the surrounding environment
- Process it quickly enough for conversation
- Present text within a limited display area
- Update the text as the conversation continues
- Keep captions readable without blocking too much of the user's view
Unlike prepared closed captions, wearable captions usually cannot be reviewed by an editor before the user sees them. They may also focus primarily on spoken words rather than identifying every sound in the environment.
The purpose is different as well. Video captions help someone follow media displayed on a screen. Wearable captions are intended to provide a visual reference while the user remains engaged with people and activity around them.
CaptionLens is developing wearable caption technology around this everyday communication use. The goal is not to replace professionally prepared captions or create an official transcript. It is to make spoken words available as an additional visual source of information during live conversation.
The underlying text may still be called captions, but the environment, timing, display, and user needs are different from those of a television program or prerecorded video.
