| 摘要 |
A method and apparatus for providing unified conferencing services in a telecommunications system is described. A DSP chip in the system contains an internal CPU and on-chip memory storage device. The CPU is programmed such that when a conference has been established, the time slots or channels of PCM information are collected and stored. Channels are sub-sampled at a pre-determined optimal time interval. The energy for each sub-sample, which is a mathematical summation over time utilizing a sum of squares equation, is calculated for each channel. These calculated energies are stored in an associated array. The energies are surveyed to select the N loudest channels. The input signals (the voice information) for those N loudest channels are used to form the conference. More specifically, the sum of the N loudest channels is generated. This sum forms the basis for a broadcast output. In the case of an individual participant who is not one of the N loudest, the broadcast output is written out in the correct PCM format to that conferee. In the case of an individual participant who is actually one of the N loudest, its own sample is subtracted from the total sum and the result is written out in the correct PCM format to the particular conferee. The broadcast outputs are also written out in required PCM formats for half-duplex listen-only connections to the conference.
|
