Image Processing Reference
In-Depth Information
Fig. 3.9 Example of tile compatible format [
23
]
about how decoders should handle their output according to the used frame-
packaging scheme. The various packing formats include: Checkerboard, column-
based interleaving, row-based interleaving, top-bottom, temporal interleaving, and
2D. In the last amendment of H.264/AVC the tile format [
23
] was added, which is
able to handle two HD frames (1280 per 720 pixels) to be packed into a full HD
frame (1920 per 1080 pixels) using a tiling method, where different regions of one
view are tiled with the other view.
As seen in Fig.
3.9
, the left view is located at the top left corner of the full HD
frame and the right view is divided into three regions, which are placed in specific
regions of the resulting full HD frame. While maintaining the backward compati-
bility with legacy 2D devices, as it requires only a 720p crop to obtain a 2D version
of the content in HD resolution, the full resolution of the original frames is
maintained in all dimensions, as there are no downsampling operations involved.
Also in H.265/HEVC, the spatial and temporal frame packing arrangement is
supported by SEI messages.
3.3.1.2 Simulcast with Hybrid Schemes
Despite the higher compression efficiency of the most advanced video coding
algorithms, the providers must take into consideration the capabilities of the
existing receivers and the transition plan, when introducing new services. For
example, there are still many terrestrial broadcasting systems based on MPEG-2.
Also several US cable providers rely their systems on mix of MPEG-2 and H.264/
AVC. Thus, rather than using the video encoder algorithm for the second view, it is
also possible to consider the use of advanced codecs to encode the video required to
form the 3D program. This is referred to as a hybrid solution, since two different
codecs (e.g., MPEG-2, H.264/AVC and H.265/HEVC) would be used to represent a
single 3D program [
24
].
With the advent of H.265/HEVC it is likely that this high efficiency video coding
algorithmmay be preferred to encode the second view, independently of the residue
resulting from some kind of inter-view or temporal prediction, as depicted in
