<s>
Temporal	B-Algorithm
anti-aliasing	I-Algorithm
(	O
TAA	O
)	O
is	O
a	O
spatial	B-Algorithm
anti-aliasing	I-Algorithm
technique	O
for	O
computer-generated	O
video	O
that	O
combines	O
information	O
from	O
past	O
frames	O
and	O
the	O
current	O
frame	O
to	O
remove	O
jaggies	B-Algorithm
in	O
the	O
current	O
frame	O
.	O
</s>
<s>
Prior	O
to	O
the	O
development	O
of	O
TAA	O
,	O
MSAA	B-Algorithm
was	O
the	O
dominant	O
anti-aliasing	O
technique	O
.	O
</s>
<s>
MSAA	B-Algorithm
samples	O
(	O
renders	O
)	O
each	O
pixel	O
multiple	O
times	O
at	O
different	O
locations	O
within	O
the	O
frame	O
and	O
averages	O
the	O
samples	O
to	O
produce	O
the	O
final	O
pixel	O
value	O
.	O
</s>
<s>
This	O
makes	O
TAA	O
faster	O
than	O
MSAA	B-Algorithm
.	O
</s>
<s>
In	O
parts	O
of	O
the	O
picture	O
without	O
motion	O
,	O
TAA	O
effectively	O
computes	O
MSAA	B-Algorithm
over	O
multiple	O
frames	O
and	O
achieves	O
the	O
same	O
quality	O
as	O
MSAA	B-Algorithm
with	O
lower	O
computational	O
cost	O
.	O
</s>
<s>
TAA	O
and	O
FXAA	B-Algorithm
both	O
sample	O
each	O
pixel	O
only	O
once	O
per	O
frame	O
,	O
but	O
FXAA	B-Algorithm
does	O
not	O
take	O
into	O
account	O
pixels	O
sampled	O
in	O
past	O
frames	O
,	O
so	O
FXAA	B-Algorithm
is	O
simpler	O
and	O
faster	O
but	O
can	O
not	O
achieve	O
the	O
same	O
image	O
quality	O
as	O
TAA	O
or	O
MSAA	B-Algorithm
.	O
</s>
<s>
Using	O
motion	O
vectors	O
from	O
the	O
game	O
engine	O
to	O
perform	O
motion	B-Algorithm
compensation	I-Algorithm
before	O
blending	O
.	O
</s>
<s>
Nvidia	O
's	O
DLSS	B-Device
operates	O
on	O
similar	O
principles	O
to	O
TAA	O
.	O
</s>
<s>
Unlike	O
TAA	O
,	O
DLSS	B-Device
does	O
not	O
sample	O
every	O
pixel	O
in	O
every	O
frame	O
.	O
</s>
<s>
DLSS	B-Device
uses	O
machine	O
learning	O
to	O
combine	O
samples	O
in	O
the	O
current	O
frame	O
and	O
past	O
frames	O
,	O
and	O
it	O
can	O
be	O
thought	O
of	O
as	O
an	O
advanced	O
TAA	O
implementation	O
.	O
</s>
