Speech recognition from one second audio clips.

image of up