Generally, a given speech sample is played over the speech path, where it presumably undergoes various types of impairments during encoding, packetization, transmission, and decoding.