<s>
The	O
macroblock	B-General_Concept
is	O
a	O
processing	O
unit	O
in	O
image	B-General_Concept
and	O
video	O
compression	O
formats	O
based	O
on	O
linear	O
block	O
transforms	O
,	O
typically	O
the	O
discrete	B-General_Concept
cosine	I-General_Concept
transform	I-General_Concept
(	O
DCT	B-General_Concept
)	O
.	O
</s>
<s>
A	O
macroblock	B-General_Concept
typically	O
consists	O
of	O
16×16	O
samples	O
,	O
and	O
is	O
further	O
subdivided	O
into	O
transform	O
blocks	O
,	O
and	O
may	O
be	O
further	O
subdivided	O
into	O
prediction	O
blocks	O
.	O
</s>
<s>
Formats	O
which	O
are	O
based	O
on	O
macroblocks	B-General_Concept
include	O
JPEG	O
,	O
where	O
they	O
are	O
called	O
MCU	O
blocks	O
,	O
H.261	B-Algorithm
,	O
MPEG-1	O
Part	O
2	O
,	O
H.262/MPEG	B-Algorithm
-2	I-Algorithm
Part	I-Algorithm
2	I-Algorithm
,	O
H.263	B-Algorithm
,	O
MPEG-4	B-Application
Part	I-Application
2	I-Application
,	O
and	O
H.264/MPEG	B-Application
-4	I-Application
AVC	I-Application
.	O
</s>
<s>
In	O
H.265/HEVC	B-Algorithm
,	O
the	O
macroblock	B-General_Concept
as	O
a	O
basic	O
processing	O
unit	O
has	O
been	O
replaced	O
by	O
the	O
coding	B-Algorithm
tree	I-Algorithm
unit	I-Algorithm
.	O
</s>
<s>
A	O
macroblock	B-General_Concept
is	O
divided	O
into	O
transform	O
blocks	O
,	O
which	O
serve	O
as	O
input	O
to	O
the	O
linear	O
block	O
transform	O
,	O
e.g.	O
</s>
<s>
the	O
DCT	B-General_Concept
.	O
</s>
<s>
In	O
H.261	B-Algorithm
,	O
the	O
first	O
video	O
codec	O
to	O
use	O
macroblocks	B-General_Concept
,	O
transform	O
blocks	O
have	O
a	O
fixed	O
size	O
of	O
8×8	O
samples	O
.	O
</s>
<s>
In	O
the	O
YCbCr	O
color	O
space	O
with	O
4:2:0	O
chroma	O
subsampling	O
,	O
a	O
16×16	O
macroblock	B-General_Concept
consists	O
of	O
16×16	O
luma	O
(	O
Y	O
)	O
samples	O
and	O
8×8	O
chroma	O
(	O
Cb	O
and	O
Cr	O
)	O
samples	O
.	O
</s>
<s>
This	O
design	O
is	O
also	O
used	O
in	O
JPEG	O
and	O
most	O
other	O
macroblock-based	O
video	O
codecs	O
with	O
a	O
fixed	O
transform	O
block	O
size	O
,	O
such	O
as	O
MPEG-1	O
Part	O
2	O
and	O
H.262/MPEG	B-Algorithm
-2	I-Algorithm
Part	I-Algorithm
2	I-Algorithm
.	O
</s>
<s>
4:0:0	O
,	O
4:2:2	O
,	O
or	O
4:4:4	O
,	O
the	O
number	O
of	O
chroma	O
samples	O
in	O
a	O
macroblock	B-General_Concept
will	O
be	O
smaller	O
or	O
larger	O
,	O
and	O
the	O
grouping	O
of	O
chroma	O
samples	O
into	O
blocks	O
will	O
differ	O
accordingly	O
.	O
</s>
<s>
In	O
more	O
modern	O
macroblock-based	O
video	O
coding	O
standards	O
such	O
as	O
H.263	B-Algorithm
and	O
H.264/AVC	B-Application
,	O
transform	O
blocks	O
can	O
be	O
of	O
sizes	O
other	O
than	O
8×8	O
samples	O
.	O
</s>
<s>
For	O
instance	O
,	O
in	O
H.264/AVC	B-Application
main	O
profile	O
,	O
the	O
transform	O
block	O
size	O
is	O
4×4	O
.	O
</s>
<s>
In	O
H.264/AVC	B-Application
High	O
profile	O
,	O
the	O
transform	O
block	O
size	O
can	O
be	O
either	O
4×4	O
or	O
8×8	O
,	O
adapted	O
on	O
a	O
per-macroblock	O
basis	O
.	O
</s>
<s>
Distinct	O
from	O
the	O
division	O
into	O
transform	O
blocks	O
,	O
a	O
macroblock	B-General_Concept
can	O
be	O
split	O
into	O
prediction	O
blocks	O
.	O
</s>
<s>
In	O
early	O
standards	O
such	O
as	O
H.261	B-Algorithm
,	O
MPEG-1	O
Part	O
2	O
,	O
and	O
H.262/MPEG	B-Algorithm
-2	I-Algorithm
Part	I-Algorithm
2	I-Algorithm
,	O
motion	B-Algorithm
compensation	I-Algorithm
is	O
performed	O
with	O
one	O
motion	O
vector	O
per	O
macroblock	B-General_Concept
.	O
</s>
<s>
In	O
more	O
modern	O
standards	O
such	O
as	O
H.264/AVC	B-Application
,	O
a	O
macroblock	B-General_Concept
can	O
be	O
split	O
into	O
multiple	O
variable-sized	O
prediction	O
blocks	O
,	O
called	O
partitions	O
.	O
</s>
<s>
In	O
an	O
inter-predicted	O
macroblock	B-General_Concept
in	O
H.264/AVC	B-Application
,	O
a	O
separate	O
motion	O
vector	O
is	O
specified	O
for	O
each	O
partition	O
.	O
</s>
<s>
Correspondingly	O
,	O
in	O
an	O
intra-predicted	O
macroblock	B-General_Concept
,	O
where	O
samples	O
are	O
predicted	O
by	O
extrapolating	O
from	O
the	O
edges	O
of	O
neighboring	O
blocks	O
,	O
the	O
predicted	O
direction	O
is	O
specified	O
on	O
a	O
per-partition	O
basis	O
.	O
</s>
<s>
In	O
H.264/AVC	B-Application
,	O
prediction	O
partition	O
size	O
ranges	O
from	O
4×4	O
to	O
16×16	O
samples	O
for	O
both	O
inter-prediction	O
(	O
motion	B-Algorithm
compensation	I-Algorithm
)	O
and	O
intra-prediction	O
.	O
</s>
<s>
A	O
possible	O
bitstream	O
representation	O
of	O
a	O
macroblock	B-General_Concept
in	O
a	O
video	O
codec	O
which	O
uses	O
motion	B-Algorithm
compensation	I-Algorithm
and	O
transform	O
coding	O
is	O
given	O
below	O
.	O
</s>
<s>
It	O
is	O
similar	O
to	O
the	O
format	O
used	O
in	O
H.261	B-Algorithm
.	O
</s>
<s>
The	O
term	O
macroblocking	O
is	O
commonly	O
used	O
to	O
refer	O
to	O
block	O
coding	O
artifacts	B-Algorithm
.	O
</s>
