Single-input-multiple-ouput transcoding for video streaming

In this work, a single input multiple output (SIMO) transcoding architecture is proposed. SIMO will benefit the mobile edge computing (such as HTTP Live Streaming requiring multiple copies of the video streams at different quality levels) without resorting to the legacy transcoding that video stream is completed decoded and encoded multiple times without exploring the compressed information. Leveraging the information encoded in the existing video streams, we could reduce the search candidates when transcoding the high quality bitstream to other versions with reduced quality level. As the first step, we have demonstrated the SIMO idea with bit rate shaping (i.e., bit rate transcoding) only scenario. It has shown more than 2x complexity reduction without quality loss using the common test conditions.