Hello. We are working on a low latency virtual reality project. Currently we are encoding the stereo content stored side-by-side using NV_ENC_CODEC_H264_GUID and NV_ENC_PRESET_LOW_LATENCY_HQ_GUID and it works through the whole pipeline.
However H.264 supports MVC for use with e.g. stereo content to achieve better quality/lower bitrate. We would like to use that. nvEncoder seems to support this through NV_ENC_H264_PROFILE_STEREO_GUID but we are fighting with lack of documentation here. When we set profileGUID=NV_ENC_H264_PROFILE_STEREO_GUID, encodeCodecConfig.h264Config.enableStereoMVC=true, and encodeCodecConfig.h264Config.stereoMode=NV_ENC_STEREO_PACKING_MODE_SIDEBYSIDE, the call to NvEncoderD3D11::EncodeFrame() hangs in WaitForCompletionEvent() in NvEncoder::GetEncodedPacket().
Can anyone give us a hint/an example on how to properly use NV_ENC_H264_PROFILE_STEREO_GUID?
We are able to generate the stereo content in side-by-side, top-bottom, or succeeding surfaces.