Coming up with a new sounds becomes progressively more difficult, not to mention the problems with similar sounds when in a loud area (IE rushing river, the hunt). Some protohuman gets the idea to combine two phonemes. Well, that's going to make things simpler later on, but you already have all these phonemes, you're going to use them.

As people move, they encounter more needs for words (new animals, etc) so each group figures out the phonemes that work best for their combinations (The clicks go out of style) but stagnant groups tend not to change as much. Thus, the phonemes stay in sub-Saharan Africa, and disappear elsewhere.