With the fixed size of the video bubbles - you lose a lot of visual detail on the faces of the people you are speaking to. This limits Freeflows ability to serve the one on one use case - or any situation in which you really need to know how the other person is feeling in that moment.
Suggestion: Adaptively scale the size of the videos, depending upon the number of people on the call.
For example, if it's just 2 - It would ideally be as close to the total available screen area possible.