One scenario that I want to implement is when decoding a clip from file but the user wants to jump forward to a later frame. So in say a 100 frame clip, the clip starts but the user wants to jump to from 68 by sliding a slider widget.
In this case would I have to determine if frame 68 is an I-Frame or not? If it is just render it, if not would I have to seek back in the file to the most previous I-Frame, then step forward and process any P and B-Frames up through frame 68 in the VC1Decoder?
Would I have to read the VC-1 specification to understand enough about the data layout or should I use a VC1Splitter instance?
VC1Splitter produces VC-1 frames without additional information about frame type. So, it need touse output of VC1Splitter and parse several first bytes of each frame according to spec (see 7.1 Chapter of SMPTE 421M VC-1 standard) to determinate frame type.