Text-Based Video Generation with Human Motion and Controllable Camera