Searching speech should be as easy as searching text. To a search engine, untranscribed audio is invisible audio. With text, every keyword is a breadcrumb, leading new audiences to discover that material online. Yet with audio, these same keywords are hidden, making recorded voices inaccessible to the audiences they deserve.
Pop Up Archive uses the latest in automatic speech recognition to make audio visible to the web. We do this both with tools directly for creators at Pop Up Archive, and for wider audiences, through our project Audiosear.ch: a full-text search and recommendation engine for podcasts and radio.
This blog highlights the archival audio uploaded by our partners at Pop Up, and innovative audio projects from around the world.