We are a research team in University of Washington and have some questions about the Video Coding Engine technology, more specifically, VCE-DEM technology.
- What is the detailed format of V-ES output type?
- We chose V-ES as the output format and tried to decode the output using VA-API hardware acceleration. However we failed using both gstreamer, libav*, and ffmpeg libraries. Our decoder reported “no frame” for every frame sent. We searched online and believe it might be related to loss of SPS/PPS/other header informations but we are not sure about this.
- What is the average/typical latency of VCE-DEM using “Low latency mode” ?
(only taking into account the latency of encoding and capturing, not decoding)
- We are also considering using AV-TS instead of V-ES but failed to demux the mpegts stream using libavformat library. We might need to fix the demuxer by ourselves and would like to learn more details about Transport Stream format.
- Is there any recommended software/library that could utilize VA-API or other hardware accelerations to decode the output of VCE-DEM?
- What’s the latest version of the AMD Media SDK? It seems to us that SDK 1.1 beta is the latest version but the VCE library is at V3.0 which is pretty weird. Why does VCE library have more revisions than the AMD Media SDK?