<s>
The	O
sum	B-Algorithm
of	I-Algorithm
absolute	I-Algorithm
transformed	I-Algorithm
differences	I-Algorithm
(	O
SATD	B-Algorithm
)	O
is	O
a	O
block	B-General_Concept
matching	O
criterion	O
widely	O
used	O
in	O
fractional	O
motion	O
estimation	O
for	O
video	O
compression	O
.	O
</s>
<s>
It	O
works	O
by	O
taking	O
a	O
frequency	B-Algorithm
transform	I-Algorithm
,	O
usually	O
a	O
Hadamard	B-Algorithm
transform	I-Algorithm
,	O
of	O
the	O
differences	O
between	O
the	O
pixels	O
in	O
the	O
original	O
block	B-General_Concept
and	O
the	O
corresponding	O
pixels	O
in	O
the	O
block	B-General_Concept
being	O
used	O
for	O
comparison	O
.	O
</s>
<s>
The	O
transform	O
itself	O
is	O
often	O
of	O
a	O
small	O
block	B-General_Concept
rather	O
than	O
the	O
entire	O
macroblock	B-General_Concept
.	O
</s>
<s>
For	O
example	O
,	O
in	O
x264	B-Language
,	O
a	O
series	O
of	O
4×4	O
blocks	O
are	O
transformed	O
rather	O
than	O
doing	O
the	O
more	O
processor-intensive	O
16×16	O
transform	O
.	O
</s>
<s>
SATD	B-Algorithm
is	O
slower	O
than	O
the	O
sum	B-Algorithm
of	I-Algorithm
absolute	I-Algorithm
differences	I-Algorithm
(	O
SAD	O
)	O
,	O
both	O
due	O
to	O
its	O
increased	O
complexity	O
and	O
the	O
fact	O
that	O
SAD-specific	O
MMX	B-Architecture
and	O
SSE2	B-General_Concept
instructions	O
exist	O
,	O
while	O
there	O
are	O
no	O
such	O
instructions	O
for	O
SATD	B-Algorithm
.	O
</s>
<s>
However	O
,	O
SATD	B-Algorithm
can	O
still	O
be	O
optimized	O
considerably	O
with	O
SIMD	B-Device
instructions	O
on	O
most	O
modern	O
CPUs	B-Device
.	O
</s>
<s>
The	O
benefit	O
of	O
SATD	B-Algorithm
is	O
that	O
it	O
more	O
accurately	O
models	O
the	O
number	O
of	O
bits	O
required	O
to	O
transmit	O
the	O
residual	O
error	O
signal	O
.	O
</s>
<s>
As	O
such	O
,	O
it	O
is	O
often	O
used	O
in	O
video	O
compressors	O
,	O
either	O
as	O
a	O
way	O
to	O
drive	O
and	O
estimate	O
rate	O
explicitly	O
,	O
such	O
as	O
in	O
the	O
Theora	B-Algorithm
encoder	O
(	O
since	O
1.1	O
alpha2	O
)	O
,	O
as	O
an	O
optional	O
metric	O
used	O
in	O
wide	O
motion	O
searches	O
,	O
such	O
as	O
in	O
the	O
Microsoft	O
VC-1	B-Operating_System
encoder	O
,	O
or	O
as	O
a	O
metric	O
used	O
in	O
sub-pixel	O
refinement	O
,	O
such	O
as	O
in	O
x264	B-Language
.	O
</s>
