Audio descriptions for videos, read by your screen reader.
This project is designed for people who regularly use both audio descriptions and screen readers.
Available Video List. Subscribe to Updates on Substack.
To use audio descriptions on this site, you will need:
To get started, try the music video "Tonight, Tonight" by the Smashing Pumpkins. After clicking on that link, you will find a YouTube video. As the video plays, your screen reader will begin to speak descriptions. As you play and pause, the descriptions will play and pause alongside you. You can jump around the video, and the descriptions will update appropriately.
We recommend against adjusting the playback rate. The descriptions are designed to be read by a screen reader at a certain rate. If you increase the playback rate, the screen reader may not be able to read the text fast enought to keep up with the video.
As you may already know, audio descriptions are usually an alternative audio track used in movies or on streaming sites, like Netflix or Amazon Prime. These tracks will have an extra narrator describe the visuals on screen, which helps people with blindness or low vision follow the story. Studios produce their audio descriptions with the help of script writers, voice actors, and audio engineers. That number of skills can produce a polished, professional product.
The Screen Reader Description Project believes that pro-voiced Audio Description is the best form of audio description. Unfortunately, pro-voiced AD isn't always cost-effective, especially for shorter videos commonly found on the web. This project provides a cheap alternative method for producing AD, making it easier for members of the community to contribute new descriptions and make more media accessible to more people.
The Screen Reader Description Project is an open source repository of audio descriptions (and the technology to play them via the screen reader). Our audio descriptions are a collection of timestamps, description text, and some metadata. We play YouTube videos, and track where the time is on the video. When the video's time overlaps with an audio description, we tell the screen reader to say the description text. This instruction is done using a "live region", leveraging the "aria-live" attribute in HTML. As far as users aware, they just play the video, and their screen reader starts describing alongside the video.
For now, the project only works with youtube videos, embedded on our site.
If you'd like to see how the code works with the youtube embed, check out the youtube_embed.html page in our GitHub repo.