Subjective effects of transmitting speech over packet networks

Packet switching is suitable for the transmission of most types of data, but it is not a natural medium for speech and at present there are no agreed standards on the way in which this should be done. To help determine some of the acceptable limits for such standards a series of experiments have been carried out. These examine some of the subjective effects of using packet networks as carriers of speech. Delays when sending/receiving packets over these networks are not usually critical for data, but they are not acceptable for speech. If speech packets are delayed for too long they must be assumed to be missing. Different methods have been proposed to cope with this loss of information. Listening tests have been run to determine the subjective effects of some of these methods, and to find what percentages of missing packets are found to be both detectable and objectionable to the listener