A method of video conference communication between N terminals of N users is described. The method can be implemented by one of the N terminals, and can include receiving, from a processing device, N audiovisual streams respectively transmitted by the N terminals, N items of voice activity information of the N users, respectively associated with N corresponding user identifiers, each of the N items of information assuming a first or a second value respectively representing the presence or absence of voice activity. The method can also include determining, for at least one of the N users, whether or not the information is at the same value from a certain time. The method can also include requesting, from the device upon determining that the information is at the first value from this time, a stream associated with the user as the main stream to be displayed, receiving, and displaying the main stream.