Graphics Reference
In-Depth Information
picture, and changing its POC value to zero would change its POC relationship with
other pictures that follow it in decoding order. Note that the BLA picture type is
allowed to be used even though no splicing has occurred.
IDR and BLA picture types may look similar, and converting a CRA picture into
an IDR picture may look like a possibility during splicing. This is certainly possible
but not easy in practice. Firstly, RASL pictures are not allowed to be associated with
IDR pictures, so their presence has to be checked before it can be decided whether
the IDR picture type actually can be used. Alternatively they can be removed but
then it might be necessary to recalculate the buffer parameters. Secondly, the syntax
of an IDR picture slice segment header differs for CRA and BLA pictures. One
example is that POC information is signaled for CRA and BLA pictures but not for
IDR pictures. Therefore the splicer needs to rewrite the slice segment header of the
picture. None of this is needed if BLA is chosen, then changing the NAL unit type
in the NAL unit headers is sufficient.
As shown in Table
2.1
, there are three BLA NAL unit types in HEVC.
BLA_N_LP for which leading pictures are forbidden, BLA_W_RADL for which
RASL pictures are forbidden but RADL pictures may be present, and BLA_W_LP
for which both RASL and RADL pictures are allowed. Even though it is recom-
mended that the splicer checks the subsequent leading picture types and uses the
correct BLA type in the spliced output bitstream, it is allowed for a splicer to always
use BLA_W_LP. By this, the splicer does not need to inspect the NAL units that
follow to check for leading pictures.
2.2.2.10
Sub-layer Reference and Sub-layer Non-reference Pictures
As can be seen in Table
2.1
, each leading picture and trailing picture type has
two type values. The even picture type numbers indicate sub-layer non-reference
pictures and odd picture type numbers indicate sub-layer reference pictures. An
encoder can use the sub-layer non-reference picture types for pictures that are not
used for reference for prediction of any picture in the same temporal sub-layer.
Note that a sub-layer non-reference picture may still be used as a reference picture
for prediction of a picture in a higher temporal sub-layer. A network node can use
this information to discard individual sub-layer non-reference pictures of the highest
sub-layer that it operates on.
Figure
2.10
shows an example where pictures that may use the sub-layer non-
reference picture NAL unit types are indicated by an asterisk (*). These are the
pictures that are not used for reference by pictures of the same temporal sub-layer,
i.e. they do not have an arrow to a picture of the same layer. If HighestTid is
two, pictures B
4
,B
7
,andB
8
may be individually discarded without affecting the
ability to decode other pictures in temporal sub-layers up to that sub-layer, but not
other pictures. If HighestTid is one, picture B
6
could be similarly discarded, and if
HighestTid is zero, picture P
1
could be similarly discarded.
