Page 42 - ITU Journal, ICT Discoveries, Volume 3, No. 1, June 2020 Special issue: The future of video and immersive media
P. 42
ITU Journal: ICT Discoveries, Vol. 3(1), June 2020
Fig. 2 shows an example of an overlay in an
omnidirectional system. The background
omnidirectional video is rendered on the spherical
surface and the overlay video is rendered on a plane
inside the spherical volume. The user viewing the
omnidirectional video is assumed to be present at
the center of the spherical volume.
MPEG OMAFv2 defines an overlay as a visual media
(video, image, or timed text) that is rendered over
an omnidirectional video or image or over a
viewport. In the presence of an overlay, the visual
media on which it is overlaid is referred to as
background visual media. In the present context, the Fig.2 – Example of a 2D overlay at a depth from the user
background visual media can be considered as viewing point in an OMAFv2 omnidirectional system.
OMAF omnidirectional video or image content. The
next section describes the different features of
overlays as defined in OMAFv2.
3.1 Overlay features in OMAFv2
OMAFv2 provides a framework to indicate the
number of overlays present and also active in the
omnidirectional system. Additionally, it provides
the capability and flexibility to control different
overlay features. These features can be broadly
categorized into the following four types:
a) Spatial property;
b) Temporal property; Fig. 3 – Example of a spherical overlay on the background
video in an OMAFv2 omnidirectional system.
c) Interactivity property;
d) Inherent property.
3.1.1 Spatial property
Every overlay in the OMAFv2 system has a defined
spatial position. The standard defines the following
three possible positions for an overlay:
1) the overlay could be positioned at a depth
from the user viewing position, as shown in
Fig. 2;
2) the overlay may be positioned over the
background video without any gap between
the two, as shown in Fig. 3;
Fig. 4 – Example of a viewport-relative overlay in an OMAFv2
3) the overlay may be positioned on the users’ omnidirectional system.
viewing screen. This is also called viewport- The standard also enables the flexibility to signal
relative or viewport-locked overlay; it is the rotation angles (yaw, pitch, roll) (for 2D
always present on the users’ viewing screen overlays only) and the size of each overlay. In
and is independent of the users’ current addition, it also allows us to signal the layering
viewport. An example is shown in Fig. 4. order of overlays when multiple overlays are
present at the same depth in the omnidirectional
system.
20 © International Telecommunication Union, 2020