Sparsity can be used in solving linear inverse problems,
when the setting is naturally sparse, e.g. in source localisation.
We present examples from acoustics using sparsity ideas.
In particular, we answer the old question “can one hear the shape of a room”, a classic inverse problem from acoustics. 
Assume you are blindfolded in a room, snap your finger, and listen to the echoes. Can you ''see'' the shape of the room,
something actually certain blind people are able to do. We show a positive answer, as well as a constructive algorithm
to reconstruct convex rooms from first order echoes. The method is tested in experiments. 
The results are surprisingly precise, showing the robustness of the algorithm. We then solve acoustic indoor localization
using this new algorithm.