Image Processing Reference
In-Depth Information
3.3.1 Stereoscopic Video Coding
The initial approach for 3D video coding and transmission involved the use of a
couple of cameras to acquire two offset video sequences separately to the left and
right eye of the viewer. This is still the most common way to deliver 3D video
content, by means of a simulcast or a frame compatible approach. In this section we
will describe briefly how these systems use the state-of-the-art video coding
standards, H.264/AVC [ 20 ] and H.265/HEVC [ 21 ], to implement such solutions.
3.3.1.1 Frame Compatible
Allowing the transmission of 2D-compatible formats over current networks enabled
the rapid deployment of 3D video delivery services and applications. Broadcasters
embraced this solution in the initial phase of services, maintaining the use of legacy
video decoders with 3D-ready displays already popular in the user market.
Therefore, a set of frame compatible formats have been defined, which allow to
transmit stereoscopic video without the need to modify the coding algorithms used
for single video compression, like H.264/AVC or H.265/HEVC. For a stereo video,
it consists of packing pixels from left and right views into a single frame of
compressed stream, at the cost of reducing the spatial or the temporal resolution.
In the former, the spatial resolution is affected due to the column or row pixel
downsampling in each view, while in the latter the views are interleaved as
alternating frames or fields, as can be seen in Fig. 3.8 .
The left and right views are identified in the compressed stream via specific
signaling. 3D frame compatible formats encoded with H.264/AVC use the supple-
mental enhancement information (SEI) messages, as defined by the subclause
D.2.25 in the standard [ 20 , 22 ]. These messages are used to convey information
a
b
Left
Right
Left
Right
Left
Right
Time
Fig. 3.8 Example of frame compatible formats. (a) Side-by-Side; (b) temporal interleaving
Search WWH ::




Custom Search