Did you know that adding words to a learning video can make it much easier to use—but only if they are the right kind of words? A child who wants to follow a teacher’s explanation while watching, another who needs a translation, and a third who wants to reread the explanation may all be asking for different things.
The answer lies in three related terms. Captions present speech and important sounds in synchrony with a video. Subtitles often present a translation of speech, although many platforms and regions use the words captions and subtitles differently. A transcript presents the audio content as separately readable text; a descriptive transcript also includes visual information needed to understand the video. Knowing which one a learner needs changes what a parent, student or teacher actually prepares.
An invented Secondary 1 Science lesson about water
Picture an 80-second class demonstration. A dish of water stands by a window. The narrator says, “Some liquid water changes into water vapour and moves into the surrounding air.” A quiet fan hums while an animation shows arrows moving upward from the water’s surface. The word evaporation appears on a diagram. These details are entirely invented to teach an accessibility choice; they are not a description of a school programme or actual learner result.
If a pupil has difficulty hearing the narration, accurate captions can display its words at the moment they are spoken and identify meaningful sounds. If another pupil is learning English, appropriately translated subtitles may help with the explanation, while a trusted glossary supports new Science vocabulary. If a pupil wants to look up evaporation while revising, a transcript makes the sentence easier to search. And if someone cannot see the arrows, simply typing the speech does not explain everything the diagram conveys.
Five words that solve five different reader jobs
- Captions: written words synchronised with a video that present speech and relevant non-speech audio. For example, a caption can show the narrator’s sentence and identify “[fan humming]” if that sound matters to the lesson.
- Subtitles: words displayed during playback, often a translation of dialogue. Some interfaces call same-language text “subtitles”, too, so check what the specific track contains.
- Transcript: written audio content available for reading separately from the moving picture. It may include speakers’ names, meaningful sound information or timestamps and can support searching and review.
- Descriptive transcript: a transcript that includes essential visual information, such as what a silent chart shows, which object an arrow points to, or what label appears on screen.
- Audio description: spoken narration that explains important visual information. This is an audio access tool and is not the same as captions.
- Synchronisation: alignment of text with the correct moments of speech and sound. Captions that arrive too early or late can distract or mislead.
- Automatic speech recognition: software that predicts spoken words from audio and produces text. Its output needs checking, particularly for unfamiliar terms, numbers and names.
Watch the word families develop: transcribe → transcription → transcript; translate → translation → translated; describe → description → descriptive; synchronise → synchronised → synchronisation. Natural collocations include closed captions, translated subtitles, time-aligned captions, searchable transcript and describe important visuals. That vocabulary helps a pupil ask a precise question instead of the vague “Can I have the text?”
Four text outputs from the same twelve seconds
1. Synchronous captions
At 00:12, as the video shows water in the dish, the on-screen caption reads: “As the liquid warms, some water evaporates into the air.” When a meaningful sound occurs, a relevant short cue could be added. The viewer can connect what is heard with what is seen at that moment. Accurate timing helps the sentence follow the changing illustration.
2. Translated subtitles
The same sentence appears at roughly the same time, but in another language selected for the learner. A sound translation preserves the distinction between evaporation and boiling. It should not invent a temperature, a new explanation or a certainty that the speaker never gave. A translated subtitle track may not contain all the sound cues that a full caption track provides.
3. A basic transcript
Below the video, a reader finds a separate line: “Narrator: As the liquid warms, some water evaporates into the air.” That text can be highlighted, searched, reread or annotated at the learner’s own pace. Unlike captions, the paragraph does not have to appear in the same second as the spoken line. A pupil preparing notes may find this format especially convenient.
4. A descriptive transcript
The transcript adds a factual description: “On screen: blue arrows rise from the water surface toward the air; a label identifies the process as evaporation.” Those extra words share what the visual communicates when the narrator never says it aloud. The description should report what appears, not embellish an animation with invented scientific conclusions.
Why “words on screen” is not the same as access
Suppose a video displays English captions that repeat every spoken sentence, while a silent diagram names four parts of a flower. A person relying only on those captions may still miss the information in the diagram. Conversely, a detailed transcript with diagrams explained may help with review but cannot provide the moment-by-moment simultaneous experience of watching with captions. The format should fit the barrier, not just satisfy a vague instruction to add text.
The World Wide Web Consortium’s media accessibility guidance on captions explains that captions represent speech and relevant non-speech sound in synchrony with video. Its guidance on transcripts distinguishes basic transcripts and descriptive transcripts, including their value for learners who need text access to audio and important visual information. These are accessibility principles, not a claim that a specific classroom or examination mandates one particular production workflow.
A subject-word error can survive beautifully formatted captions
Automatic speech recognition might write concentration when a science speaker actually says condensation. Both words are real; they refer to different concepts. Similarly, evaporation and evaluation may look plausible to a rushed student yet would change the meaning. A neat font and perfect timing cannot correct an inaccurate term. Ask pupils to check important subject words against the video, lesson material and a reliable glossary.
The existing eduKateSG article on checking automatic captions before turning videos into study notes owns the specific task of correcting auto-generated study text. This article addresses the different decision that happens first: which caption, subtitle or transcript format a learner needs.
A practical choice guide for a real lesson
- A pupil needs to read spoken words while the scene changes: select accurate, synchronised captions in a language the pupil can use.
- A pupil needs help understanding speech from another language: consider appropriate translated subtitles, while checking subject terminology rather than trusting a machine translation blindly.
- A student needs to search, copy short study phrases, review at a slower pace or follow along separately: provide a readable transcript where available and permitted.
- An important chart, gesture, title or picture is never explained aloud: ensure its meaning is accessible through appropriate visual description or a descriptive transcript.
- A student needs the sound explained in text, such as an alarm or a meaningful door slam: captions should communicate that non-speech cue when it matters.
- A platform labels a track “auto-generated”: verify names, numbers, units, negative words, speaker changes and scientific vocabulary before relying on the result.
Teacher judgment matters. Some learners benefit from more than one tool at once, and individuals should be asked what actually works for them rather than having a support format assigned on assumptions about age, disability or language background. A Primary English learner using text as a reading scaffold has a different reader job from an adult needing a complete descriptive record of a public meeting.
Three low-cost tests for parents, tutors and teachers
The sound-off test
Mute the demonstration and turn the captions on. Can a learner follow the spoken explanation without guessing? Are the non-speech sounds that actually matter represented? If the captions say “the figure”, but the picture contains several figures, a clear visual description may also be needed.
The picture-off test
Read the transcript without looking at the video. Does it convey the steps and any essential diagram information, or does it say only “as you can see”? If the video relies on arrows, colour, numbers or gestures, consider a descriptive transcript or a suitable audio description.
The science-word test
Ask the child to find the two most important terms in the text—say, evaporate and water vapour. Can they explain each one with the right subject meaning? Would a mistranslation change the claim? For Science and Mathematics, even one missed unit or a wrongly placed negative can make a sentence misleading.
Language development across Primary, secondary and adulthood
In Primary English, students can use accurate same-language captions to notice unfamiliar words, sound-letter relationships and sentence structures; they can then practise using the words independently without the video. A PSLE-age pupil should treat this as language enrichment rather than assume captioned video replaces reading, speaking or the official examination tasks.
In secondary G1/G2/G3 lessons, students may encounter videos in English, Science, Geography or Design and Technology. A clear transcript can support revisiting a challenging passage; a well-described diagram supports access to conceptual content. The necessary support follows the individual and the learning goal, not a generalisation about a student’s subject level. Later the same terms matter in workplace induction, public information, film editing, software tutorials and meetings.
Mini diagnostic and retrieval practice
Offer four learner jobs: A wants speech displayed at the moment it is said; B wants the speaker’s words translated; C wants a text document to search later; D needs to know what a silent arrow on a diagram indicates. Suitable first answers are captions for A, translated subtitles for B, transcript for C and descriptive visual information for D. Then ask where one output may be insufficient, such as when spoken sound and visual evidence both matter.
For spaced retrieval, ask students the following day to explain the difference among caption, subtitle and transcript without looking. A week later, transfer the task to a History documentary, a cooking demonstration or a school laboratory safety video. Can they choose the right format and identify what must be checked for accuracy? This requires application rather than simply recalling three labels.
Frequently asked questions
Are captions and subtitles always different?
No. Usage varies by region and product. Some services use “subtitles” for both same-language and translated text. Check the actual language, sound cues and timing rather than infer features from a menu label alone.
Can a transcript replace captions for everyone?
No. A separate transcript is excellent for review and some access needs, but time-synchronised captions support following speech while viewing. Providing both may be appropriate, and learners’ preferences and needs differ.
Can automatically generated captions be used without editing?
They may be a useful starting point, but important words, names, numerical values and negative statements need checking. A wrong scientific term is not made correct by appearing quickly.
Does every educational video need translated subtitles in every language?
Not automatically. Useful access depends on the learners and intended publication context. Prioritise accurate, relevant language support and other necessary accessible formats rather than promising every possible track.
Vocabulary is powerful because it enables a meaningful request: “Please provide time-synchronised English captions for the speech and a descriptive transcript for the arrows.” That is clearer than “The video needs words.” Explore related reading and listening pathways in the eduKateSG vocabulary hub and Parenting 101 English. All children and video details here are fictional teaching examples.