Hoppa till innehållet

Annual Report 2025/Story: Wikispeech – quality, accessibility and a stable code base

Från Wikimedia
Det här är en översatt version av sidan Verksamhetsberättelse 2025/Story: Wikispeech – kvalitet, tillgänglighet och stabil kodbas. Översättningen är till 86 % färdig och uppdaterad.
Föråldrade översättningar markeras på det här sättet.
Use
2025

In this story, we present an overall picture of what has been done in the Wikispeech project since it was awarded funding from the Swedish Inheritance Fund. We describe both the technical progress made and the organizational and methodological developments undertaken during the period.

Since our Wikispeech initiative was awarded project funding from the Swedish Inheritance Fund in October 2024, a number of important efforts and activities were carried out during 2025, when the project got properly underway. Thanks to this funding, we have been able to drive the development process forward by hiring an additional developer, a project manager, and a new technical project manager. This has in turn led to significantly improved communication with our target groups and project partners, and we have already been able to carry out several of the activities planned within the project.

These activities include reference group meetings, meetings with target groups, staff training in accessibility principles, and testing sessions with Begripsam. All of this work has been important both in ensuring that everyone involved in the project feels included and valued, and in making Wikispeech as relevant and useful as possible for as many people as possible.

The previous version of the player, before feedback from Begripsam.
The new version of the player, after feedback from Begripsam.

På den tekniska sidan har det, tack vare ovan nämnda insatser, skett betydande framsteg. Förutom att ny funktionalitet har utvecklats och den tekniska strukturen stärkts, har även det övergripande arbetet kring tekniken vidareutvecklats, genom nya insikter och fördjupade samarbeten med bland annat vår talsyntesleverantör STTS.

Since the start of the project, a number of technical improvements have been made which, taken together, have strengthened both the functionality and the user experience of Wikispeech. The player has undergone a visual and structural update in which its color, shape, markers, and layout have been reviewed in order to create a more intuitive and accessible user experience (see images).

The development of the player has been iterative and carried out in close dialog with the target groups. Through testing, reference group meetings, and their feedback, we have gained valuable insights into how users engage with and navigate digital content. This has influenced how we have placed buttons, controls, and visual elements, as well as how the interface is structured, in order to better support different user needs.

At the same time, new functionality has been added and brought into use. The player now includes features such as a pause function and an option for users to easily adjust personal settings via a button in the player, such as selecting language, voice, and reading speed. Disability rights organisations highlighted this as particularly important for users who have difficulty taking in written information, so that they have the best possible conditions for using the tool. Work has also begun on enabling Wikispeech to read aloud more parts of an article than just the body text, such as links, infoboxes, lists, images (via alt text), and references. The aim is to create a more comprehensive listening experience for Wikipedia articles.

We have also begun work on generating audio files for pages. Although this function is not yet directly available to users as a button, the technical capacity is now in place. In addition, a script has been developed to pre-generate error messages, ensuring that Wikispeech functions in situations such as when a Wikipedia article is missing or if the text-to-speech service is unavailable.

We have also strengthened accessibility by introducing ARIA labels on the buttons in the player, in line with the EU Accessibility Directive. In addition, the interface has been further developed through a new edit button in the toolbar on Wikipedia in Swedish, as well as the display of relevant information in the lexicon view for users who want to adjust pronunciation or transcription. Finally, the codebase has been modernized, improving its maintainability and long-term development potential.

All these improvements have strengthened the relevance of Wikispeech in relation to other tools on the market, helping make the tool useful to more people.

We look forward to continuing to build Wikipedia's first own text-to-speech solution, to help ensure that more people, regardless of ability or preferences, can access free knowledge on equal terms.