It has been stated frequently that the soundscape as perceived and appraised by the user of a space, extends beyond the physical stimulus. We argue that, when introducing to human-factor in analyzing a sonic environment, the sounds that people hear play an important role. This holds in particular for rather quiet and infrequent disturbance of park soundscapes. Auditory attention mechanisms are essential in the process. Attention can be drawn by saliency elements such as changes in time and frequency, but it can also be outward oriented and voluntary. These mechanisms could explain the special role of natural sounds in distracting attention from mechanical background hum in a park environment. These theoretical concepts have now been implemented in measuring equipment that allows estimating how often particular sounds will be heard by a human listener. The methodology includes biologically inspired feature extraction, learning based on co-occurrence of features and saliency, attention focusing, and inhibition of return. Extension to binaural measurements increasing the unmasking effect is also discussed.