Method and system for synchronizing audio and video streams in media relay conferencing

摘要

A novel technique allows synchronizing a plurality of audio and video streams received at a receiving media relay endpoint via an intermediate node, such as a media relay multipoint control unit. The received streams were generated by a plurality of transmitting media relay endpoints and been relayed to the plurality of receiving media relay endpoints via the intermediate node, but are seen as being using a single time domain, in terms of wall clock and manipulated timestamps, while preserving the correct capture time.

主权项

1. A method, comprising:
receiving, at an intermediate device that is communicatively coupled between a plurality of media relay endpoints, a plurality of input streams from the plurality of media relay endpoints, wherein each stream comprises input data chunks; parsing the input data chunks; converting a timestamp field of each input data chunk into a manipulated timestamp, reflecting a single time domain; and calculating a parameter corresponding to each input stream for converting the timestamp field of the input data chunks into the single time domain, comprising:
estimating a difference between a clock of a transmitting media relay endpoint of the input stream and a clock of the intermediate device; anddetermining the parameter as a function of the estimated difference, a timestamp clock rate associated with the input stream, a clock field and a timestamp field in a sender report corresponding to the input stream, a clock value of the intermediate device, and a timestamp value of the intermediate device,wherein the clock value of the intermediate device and the timestamp value of the intermediate device are obtained at a same time.