I really like this combination of multi-point interactions and speech. In the demonstration the operator utilizes a specific gesture to enable speech input. Another nice touch is audio feedback when an interaction is recognized.

Source: Microsoft Research