<p>The present invention relates to a method of communicating a stream of voice data in packet mode comprising the steps: of chopping a frame of data to be transmitted into at least two sub-frames of data as a function of a classing of the bits of the frame according to a criterion of importance of the bits in the frame; of coding the sub-frames according to respective modes of coding with a robustness which is dependent on the importance of the bits of the sub-frame in the data frame to be transmitted; and of transmitting the coded sub-frames.</p>