To better understand HD voice, let us compare it with what we already have, the voice transmitted during a PSTN conversation, that is a tranditional landline phone conversation. In this normal system, 8000 sound samples are taken in one second, thereby purposely eliminating a large part of the full spectrum of sounds that a human ear can hear in a conversation. Despite this, PSTN voice has been for many years praised as being the best of all communication means.
HD Voice improves that considerably, first by doubling the number of samplings in one second, taking 16000 of them. This considerably increases the amount of sound 'details' you can hear in a conversation and subsequently makes the conversation feel more real. Also, the frequencies at which the sound is transmitted can reach up to 7 KHz, compared to 50 Hz for PSTN, which is manifold.
What is even more interesting with HD Voice is that it does all this extra with less requirements. The bandwidth required for HD Voice is only 32 kbps, half of what is required for PSTN.
This is achieved by a codec called G.722. Other codecs commonly used are the G.722.1 and the MPEG-4 AAC Low Delay codec. These codecs work by compressing the voice captured thus making the transmission speed optimal. HD Voice normally comes with the hardware used to set up VoIP system, with the necessary codecs installed on the devices.
