Many AVC decoders, like vtdec, required the width and height field in the caps. Currently, in RTP pipeline, we are force to add h264parse element in order to gain these field. In the idea that a parser should not generally be needed after the depayloader, we should consider parsing the PPS/SPS in order to extract the width and height field. While doing so, we could complete the caps base on the known field required by downstream (e.g. like setting the profile).