Method and apparatus for providing a minimum acceptable quality of service for a voice conversation over a data network

摘要

A method and apparatus for transmitting delay-sensitive data over a packet-based network involve converting the delay-sensitive data into two versions for transfer through the network with one version of the data being used to supplement the other version of the data in the event that packets are delayed or lost. In a preferred embodiment, real time voice conversation data is compressed using two different compression algorithms, where one version is more highly compressed and of a lower quality than the other. The highly compressed data is sent before corresponding packets from the less compressed data and the highly compressed packets are buffered at the receiving device to be utilized to regenerate any data segments from the less compressed packets that are lost or delayed during transmission. Sending dual versions of the same data allows the lower quality voice segment to be used as a backup in places where the high quality voice segments are lost.