Alt Text vs Captions: What Should Accessible Ebooks Include?
An image in an ebook can carry two kinds of text that readers often confuse. Alt text is a text alternative associated with the image so assistive technology can communicate its information or function. A caption is visible content presented with the image for everyone. They can overlap in meaning, but they do different jobs.
Treating one as a substitute for the other can leave readers with missing information or needless repetition. Accessible publishing starts by asking why the image is present, what a reader must understand from it, and whether that information is already available in nearby text.
The short answer
Use alt text to provide an appropriate text alternative for readers who cannot perceive the image in the same way. Use a caption when visible context, identification, explanation, attribution, or numbering benefits all readers. Some images need both. Decorative images may require a null alternative rather than a spoken description.
The wording should respond to the image's purpose in this specific book. The same photograph may need a different alternative in a travel memoir, a botany guide, and a design textbook because each context asks the reader to notice something different.
What alt text is meant to do
Alt text communicates the essential information or function that would otherwise be unavailable when an image is not seen. A screen reader may announce it, a reading system may expose it in another way, or it may appear when an image cannot be displayed. It belongs to the digital structure rather than the visible page design.
Good alt text is concise enough to follow but specific enough to be useful. It does not need to inventory every colour, object, or background detail. It should convey the information the author expects the reader to obtain from the image.
For a functional image, the alternative describes the action or destination, not merely its appearance. For an informative photograph, it describes the relevant content. For a chart or diagram, a short alternative may identify the graphic while a longer description or nearby text communicates the important data and relationships.
What a caption is meant to do
A caption is visible editorial content. It can identify a person, place, event, object, artwork, or stage in a process. It can explain why an image matters, state a figure number, connect the image to the surrounding argument, or help readers scan a chapter.
Captions are read by people who see the image and may also be read by assistive technology as ordinary text. They should therefore make sense in the reading order. A caption placed visually under an image but stored elsewhere in the ebook structure can become confusing.
A caption is not the same as a credit line. A credit line deals with creator, source, permission, or rights information. It may appear beside the caption or in a separate list, depending on the book's design and agreements. Accuracy in one does not replace accessibility in the other.
When an image needs both
An informative image often benefits from both forms. The visible caption might identify the subject and explain its role in the chapter. The alt text can describe the visual information required to understand it. The two should cooperate rather than repeat the same sentence word for word.
For example, a caption might name a historical building and date the photograph. The alt text could describe the architectural feature being discussed in the paragraph. A reader using a screen reader then receives the identification from the caption and the relevant visual detail from the alternative.
If the caption already communicates all essential information, the alt text can be shorter and avoid duplication. Authors and ebook developers should test the combined reading experience instead of reviewing each field in isolation.
Decorative images need a different decision
An image used only for atmosphere, ornament, or visual separation may not need a spoken description. W3C guidance recommends a null text alternative for images that are purely decorative. This allows assistive technology to skip them rather than announcing unhelpful file names or repetitive descriptions.
Decorative is a functional classification, not an artistic judgment. A flourish between scenes may be decorative if the scene change is also encoded structurally. The same symbol is not decorative if it is the only signal that time, viewpoint, or subject has changed.
Authors should not omit alternatives simply because an image is attractive or because nearby readers can infer its purpose. The question is whether removing the image would remove information or functionality.
Complex images need more than a short phrase
Charts, maps, family trees, timelines, equations, and instructional diagrams can contain more information than a compact alternative can carry. A useful approach is to provide a brief alt text that names the image and states its main point, plus a longer explanation in the surrounding text or an accessible linked description.
The long description should communicate the relationships needed for the author's argument, not merely list visual elements. For data graphics, include the relevant values, trend, comparison, and conclusion. If readers need the full dataset, consider a properly structured table.
Do not hide essential analysis only inside an image. Search, text resizing, translation, quotation, and assistive reading all work better when important content is genuine text.
Avoid these common alt-text mistakes
Do not start every alternative with phrases such as image of when the reading system already identifies the element as an image. Mention the medium when it matters, such as a map, painting, screenshot, or diagram. Avoid file names, marketing keywords, and generic labels like photo or graphic.
Do not describe protected personal characteristics unless they are relevant to the context and handled respectfully. Do not guess a person's identity, emotion, disability, gender, ethnicity, or relationship. Describe observable, relevant information without adding unsupported interpretation.
Do not use the same alt text for every image in a gallery. Repetition signals that the description was treated as a formality rather than meaningful content. Conversely, do not make every description long. Brevity is valuable when it preserves the image's essential purpose.
Print and ebook editions need coordinated content
A print book displays captions but does not expose embedded alt-text fields. If the image carries information that every reader needs, ensure the visible prose or caption also supports the print edition. Accessibility planning should not create two editions with different intellectual content.
The ebook can add machine-readable alternatives and flexible descriptions without changing the visible page. During conversion, verify that image files retain the intended alternatives and that captions remain adjacent in the reading order. A design that looks correct in a preview can still be structurally wrong.
Fixed-layout ebooks require particular care because visual placement does not guarantee logical reading order. Reflowable ebooks need images, captions, and descriptions that adapt without manual positioning tricks.
How to write better alternatives
Begin with the sentence around the image. Identify what the author wants the reader to learn or do. Then state the image's contribution in direct language. Include text visible inside the image when that text is important and not supplied elsewhere.
Use the book's terminology consistently. If the chapter calls a component the pressure valve, the alternative should not rename it the small metal control. For a sequence, distinguish each step and make sure the alternatives follow the same order as the instructions.
Ask a reviewer to read the chapter without looking at the images. The reviewer should understand the argument, know why each informative image is present, and be able to complete any required task. That test will not answer every accessibility question, but it exposes missing information quickly.
A prepublication accessibility check
Create an image inventory with the file name, location, purpose, alt text, caption, credit, and long-description requirement. Mark decorative images deliberately rather than leaving fields blank by accident. Confirm that repeated icons have consistent functions.
Inspect the exported EPUB, not only the source manuscript. Check reading order, heading structure, navigation, image alternatives, captions, links, language metadata, and accessibility metadata. Use automated checking tools where available, then perform human review because software cannot decide whether a description conveys the right meaning.
Test with more than one reading environment when possible. Listen to representative pages with a screen reader or text-to-speech workflow and notice repetition, missing context, or descriptions announced far from their images.
Choose based on purpose, not appearance
Use alt text when the image conveys information or function that needs a text alternative. Use a caption when visible editorial context helps all readers. Use both when they carry complementary information. Use a null alternative only when the image is genuinely decorative and the same function or meaning is available without it.
Accessible image treatment is not extra copy added at the end. It is part of editing, rights management, design, conversion, and quality assurance. Planning it early produces clearer prose and a more dependable ebook for everyone.
Make every image earn its place
Frequently asked questions
Is alt text the same as an image caption?
No. Alt text is a structured text alternative for an image, while a caption is visible editorial content presented for all readers.
Does every ebook image need descriptive alt text?
Every image needs an accessibility decision. Informative images need an appropriate alternative, while purely decorative images may use a null alternative.
Can a caption replace alt text?
Sometimes a complete nearby caption reduces what alt text must repeat, but the image still needs correct structural treatment and contextual testing.
Should alt text begin with image of?
Usually not, because assistive technology often identifies the element as an image. Name the medium only when it adds useful context.
How should authors describe charts and diagrams?
Use a short alternative for identity and purpose, then provide the important data, relationships, and conclusion in nearby text or a longer description.
What makes an ebook image decorative?
It is decorative when removing it would not remove information or function. An ornament that is the only scene-break signal is not purely decorative.
Should alt text include words shown inside an image?
Include important visible text when it is not already available nearby. Avoid repeating text that readers will immediately encounter in the caption or body.
How can authors test ebook image accessibility?
Inspect the exported EPUB, review reading order and alternatives, use automated checks, and listen to representative pages with assistive reading tools.
Comments