Libraries and media centres: audio in the service of reading accessibility

A library doesn't only lend books: it organises reading access for a whole community, including people whom printed text keeps at a distance. That accessibility mission has a legal basis, and it has limits. Synthetic audio can play a useful role in it, provided you see clearly which role, because it replaces neither an adapted collection nor the work of professionals.

What the law already provides, and what it doesn't cover

Most countries have a legal exception for accessible formats. In France it's the so-called "disability exception," in the Intellectual Property Code; across the EU and beyond, the Marrakesh Treaty framework lets authorised bodies produce and share accessible copies (braille, large print, audio) of published works for people who can't read print, without asking each rights holder for permission. This is what makes possible the adapted catalogues that some libraries offer, often through dedicated national platforms.

We need to be precise, because the confusion is easy and costly. That exception governs the adaptation of published, protected works, within an authorised framework. It does not concern the content a library produces and publishes itself: the institution's website, the cultural calendar, the practical pages, the news, the reference guides written in house. You own those texts, and nothing stops you from offering a spoken version of them to all your users. That's exactly the ground where quality speech synthesis belongs, as a matter of the accessibility of your own pages.

Two distinct needs, two answers

The first need is access to works for people who can't read print. It falls under the legal framework above, adapted collections and specialised schemes. Synthetic audio can contribute, but this isn't where you decide alone: it goes through authorisation and the established channels.

The second need is access to the library's own information. A dyslexic, low-vision, older, or simply on-the-move user must be able to find your opening hours, your programme, your search guides. There, you own the content, and an audio version of every page removes a real barrier. We look at these two audiences in our articles on reading and dyslexia and on low vision: in both cases, hearing the text isn't a comfort, it's sometimes the only route.

What audio concretely brings a media centre

On your own content, audio changes three things. It makes your pages usable by people who wouldn't otherwise read them, which serves your first mission. It also helps rushed users, who listen to a workshop announcement while packing their things. And it gives your institution a coherent stance: a public body that talks about reading access gains from having its own pages as readable as they are listenable.

The demands of language are an asset here. An author's name, an exhibition title, a shelf mark, an address: these are exactly the places where badly tuned synthesis stumbles, giving a proper noun an English colour or misreading an acronym. Reading these elements correctly isn't a detail for a library, it's the credibility of the voice itself. That's the subject we treat at length on our French text to speech page, and it's what sets a tool tuned for the language apart from a general-purpose engine.

Where to start, without over-investing

There's no need to open everything at once. Begin with the pages your users consult most (opening hours, programme, how to register) and add listening there. A good order is to start with the two or three pages that carry the most everyday traffic, then move outward to reference guides and dossiers only once you see real listening. The audio player fits onto those pages without rebuilding the site, and you then measure what is actually listened to before extending. That measured, page-by-page approach also keeps the service honest: you extend where an audience shows up, not where you assumed it would. Treated this way, accessibility stops being a box to tick and becomes what it always was for a library, making the collection reach the person who needs it.

Read next

• Integrating schema.org markup for a narrated article: using AudioObject well
• Audio, ADHD and attention: what the format actually helps with
• Audio for accounting and consulting firms: where listening serves a practice