Screen-to-Soundscape turns a screen into a soundscape, for blind and visually impaired users. Traditional screen readers skip images, videos and maps, and offer few voices to choose from. This tool uses open-source computer vision to write detailed alt-text for images and maps, and spatial audio with layered voices, so the position of content on the page is audible and readers can pick the voice they want.
The tool is free and open source. Everything is customisable: how much detail the alt-text carries, which voice reads it, and how the layers sit in space.
Screen-to-Soundscape is supported by the Constant Foundation, The Processing Foundation, and the Stimuleringsfonds.
Collaborators: Alyssa Gersony, Bruno Defalque, Chris Alexandre, Colette Aliman, Dan Xu, Raphaël Bascour, Vincent Leone, Vladimir Nani, Luis Morales-Navarro, Stefan Laureijssen
Read more about Screen-to-Soundscape on www.screentosoundscape.com
Try It: Hear This Page
Enable spatial audio, then hover over any text block below. You'll hear a tone placed at the block's position in 3D space, followed by the text read aloud. Left blocks sound from your left ear, right from your right; top blocks are farther away, bottom blocks are closer. Pick any voice your system has installed. The browser's speech engine plays straight to the output and cannot be routed through the 3D panner, so the voice carries distance as volume while the tone carries the direction. The prototype further down does spatialise the speech itself, by synthesising it as audio data first.
Screen readers often skip images, videos, and maps, leaving blind users with an incomplete picture of digital content.
Our tool uses computer vision to generate rich, descriptive alt-text for images, making visual content accessible through sound.
Spatial audio uses multiple layered voices positioned in 3D space, so content on the left of the screen sounds like it comes from your left ear.
Traditional screen readers offer a single monotone voice. We provide diverse voice options and customizable narration styles.
Screen-to-Soundscape is free and open-source, supported by Constant, The Processing Foundation, and the Stimuleringsfonds.
Best experienced with headphones. Each block has its own tone, positioned in 3D; the voice follows with the block's distance as volume.
Phase 1B Prototype: Wikipedia Soundscape Generator
Enter a Wikipedia article to explore it as a 3D soundscape. Walk through sections with arrow keys, hear singing bowl beacons from each element's position, and listen to spatial text-to-speech. Best with headphones.
Open full-screen for the best experience.