Abstract: For use in multiport digital conference arrangements wherein speech samples of selected speakers are summed for distribution to the conferees, the disclosed method and apparatus controls the level of speech represented by information samples to be included in an output sample for distribution to the ports and equalizes the speech level between speakers to reduce speech level contrast heard by the conferees. In addition, a speech detector for each port and microprocessor-controlled switching hardware also adjust the signal level represented by samples received on the ports to effect speaker selection. Furthermore, gain coefficients for a port may be incrementally adjusted during a predetermined period of time to avoid noticeble signal level changes when implementing speaker selection.

Abstract: For use in a multichannel time division multiplexed conference arrangement where the samples in one time frame of simultaneous speakers are summed for distribution to each of the conferees, the disclosed arrangement reduces the amount of speech that is often not heard (clipped) when a conferee first begins to speak. The arrangement features a delay unit which stores all the speech samples from one time frame for a period of time greater than the processing time required to detect and determine whether to include speech samples from a new speaker in the distributed sample. By first storing the samples from one time frame and then summing the stored samples, any samples from a new speaker may be included in the resulting output sample to reduce speech clipping.

Abstract: For use in multiport digital conference arrangements adapted to sum speech samples of selected speakers, the disclosed method and apparatus improves the quality of communication by holding over selected ports for varying time intervals depending on the interval of speech above a threshold. Furthermore, a minimum number of ports is always enabled to provide background noise when all the conferees are silent.

Abstract: For use in multiport digital conference arrangements wherein the speech samples of selected speakers are summed for distribution to the conferees, the disclosed arrangement features processor-controlled switching hardware responsive to a control signal from a conference control processor. Generated in response to a request from a port for priority treatment, this control signal designates particular conferees as having assigned priority treatment such as attendant and broadcast priority in the speaker selection process. In addition, the arrangement also includes a speech detector for each port. By evaluating the conferee's assigned priority treatment, the conferee's speech level, and other factors, each conferee is dynamically assigned an adjusted priority level by a conference processing program. The speech samples of only the highest priority level conferees are summed to control various types of distortion and to permit designated conferees to be always heard by the other conferees.

Abstract: For use in multiport digital conference arrangements wherein the speech samples of selected speakers are included in an output sample for distribution to the conferees, the disclosed arrangement features a speech detector for each port and processor-controlled switching hardware. By evaluating the conferee's speech level, the current state of the port, and other factors, each conferee is dynamically assigned an adjusted priority by a conference processing program. The speech samples of only the highest priority conferees are summed to control noise level, singing caused by echo, and other types of distortion.