| CPC H04M 3/568 (2013.01) [H04L 12/1822 (2013.01); H04L 12/1827 (2013.01); H04L 12/1831 (2013.01)] | 18 Claims |
|
1. A computer-implemented method, comprising:
extracting a plurality of signal frames from an audio stream of a primary device of a group of co-present devices;
applying a smoothing function to the plurality of extracted signal frames;
determining a first proximity index value for the primary device based at least in part on output of the smoothing function;
receiving respective additional proximity index values calculated by each of one or more secondary devices in the group of co-present devices;
comparing the first proximity index value and one or more of the respective additional proximity index values against a threshold value;
based on the comparison, selecting a particular proximity index value nearest to the threshold value; and
identifying a particular source device from the primary device or the secondary devices as a source of the selected particular proximity index value; and
designating the particular source device as an active speaker device.
|