Video Tutorials: Say the Step the Camera Shows

A maker films a short tutorial on assembling a display stand. The overhead camera captures everything: a long tab, a short tab and the slot that takes the long one. On the soundtrack, the maker says, “Put this in here.” Someone who cannot see the shot knows an action took place, but cannot tell which tab to pick up or where to put it. The hardest part of following the tutorial was left to the camera.

The simplest fix can happen before recording. In this example, the maker could say, “Slide the longer tab into the slot in the base. It should sit flush with the underside.” Those two sentences name the part, locate the slot and explain how to check the result. The footage can still show the move. The sound now carries the information a listener needs to make sense of it.

Plan the description while the script is still open

The World Wide Web Consortium’s Web Accessibility Initiative describes this approach in its guidance on creating audio and video. Its guide to describing visual information calls this integrated description: the essential visual detail is spoken as part of the main script. It distinguishes that from adding a separate description for a video whose soundtrack leaves important information out. The choice depends on what the audience needs to understand and how much you can convey naturally in the main audio.

In an instructional video, start with the task the person is trying to complete. Give each critical shot a job. Does it show which control to use, where a piece fits or what confirms success? Those details deserve room in the script. A decorative plant in the background rarely needs a mention. But a red marker on one of three otherwise identical containers may matter. If colour is the only cue, decide whether a printed label, a position or a physical feature can identify the container more usefully.

For a screen recording, the same principle shifts attention from pointer movements to named actions. Imagine a creator teaching someone to export a practice file. “Click the button over there” is hard to use without the picture. A more useful line names the control, says what it will open and describes the option that matters. If the presenter intends the viewer to use a particular folder or file type, say so when that choice occurs. Listing all the controls on the screen would make the narration harder to follow and still might miss the decision.

Listen for choices, checks and changes

After editing, play the clip without looking at the picture. As a quick check, ask a colleague who has not watched it to mark where they lose track. This exercise can reveal gaps; it cannot establish how accessible the video is for every viewer. Where practical, involve people who use the format and act on their feedback. Use a short, representative task rather than asking whether the narration “sounds clear”. Someone can enjoy the presentation and still have no way to choose the correct tab.

• Check the choice. Can a listener tell which of two similar parts, files or buttons the presenter used? Give it a stable name and use that name consistently.
• Check the instruction. Does the sound say what moves, where it goes and when the viewer should act? Keep the description close to the relevant step. If listeners need time to act, leave a pause instead of crowding the next instruction.
• Check the result. If the presenter relies on a change in colour, position or displayed text to confirm success, say what changed and what that means for the next step.
• Check the final cut. Listen again after changing timing, music or screen footage. A line that made sense beside one shot can point to the wrong thing when the edit moves it.

Keep the text version useful too

Integrated description and captions solve different problems. Description makes important visual information available through sound; captions represent speech and relevant sounds as text. If you make a transcript, fold in any essential visual information that is still missing. MaryChuks.com’s guide to writing down the step the camera shows is useful when the footage already exists. If you are planning a new series, it is cheaper in editing effort to try a better spoken instruction in a rough take than to rebuild a tightly cut explanation after recording.

AI can help you review a script for phrases such as “here”, “this one” and “as you can see”. Treat its suggestions as questions to check against the footage and the materials. An assistant that guesses which tab fits can make the narration worse. The creator or editor should confirm the detail, then listen to the revised line at the actual point where it will play. For an existing clip with dense dialogue and crucial silent action, consult the W3C description guide before deciding whether a separate described version would serve the viewer better.

For the next tutorial, choose the moment at which a viewer could most easily do the wrong thing. Write the spoken instruction for that moment before filming the polished take. If a listener can identify the part, carry out the step and tell whether it worked, the narration is doing some of the teaching that previously happened only on screen.

Sources

• W3C Web Accessibility Initiative: Audio Content and Video Content and Description of Visual Information, part of its media accessibility guidance updated in September 2024.