benwaggoner
15th May 2013, 19:16
I've got some interesting 3D content lying around and thought I'd take a whack and encoding and playing it. I had encoded a bunch of stereo 3D with Expression Encoder for the Silverlight stereo 3D player several years ago, but I've never done it with x264.
First question: does --frame-packing 3 only set a metadata flag, or does it have some impact on the encoding process itself?
It occurs to me that it may be preference to encode each eye's image as an individual slice, to reduce the risk of prediction crossing the frame border and causing some weirdness around the edge. Also, a decoder could then decode left/right in parallel.
To that end, is there any way in x264 to specify the shape of the slices? I presume setting --slices 2 would give me a top/bottom split as long as the frame height is mod32. But how about a left/right split? I see lots of options to control slice size etcetera, but none that seem to apply to location.
First question: does --frame-packing 3 only set a metadata flag, or does it have some impact on the encoding process itself?
It occurs to me that it may be preference to encode each eye's image as an individual slice, to reduce the risk of prediction crossing the frame border and causing some weirdness around the edge. Also, a decoder could then decode left/right in parallel.
To that end, is there any way in x264 to specify the shape of the slices? I presume setting --slices 2 would give me a top/bottom split as long as the frame height is mod32. But how about a left/right split? I see lots of options to control slice size etcetera, but none that seem to apply to location.