uss comes with a hierarchy of sound classes it tags by default, their example image is impressive, it acts like it is separating and labeling every individual sound from a mono recording. thinking some on converting it to triangulation