Hacker News Reader

Like it would identify what you're hearing? "Trumpet!" "Wind whistling through oak leaves!" "Male child!" etc?

Ubicoustics [1] would be the closest example to what I am looking for in a FOSS / Commercial offering. Is anyone working on this?