Graphics Reference
In-Depth Information
v
1
Δ
t
v
1
g
r
p
,
s
p
h
θ
q
,
φ
q
[
]
[
]
Surface map
View map
→
d
3
v
1
v
3
v
1
v
1
(
x
3
,
y
3
)
v
1
v
1
s
1
,
t
1
(
)
v
1
v
1
s
3
,
t
3
(
)
→
y
1
v
1
v
1
x
1
,
y
1
(
)
v
1
v
1
Δ
t
(
x
2
,
y
2
)
→
→
d
2
v
1
z
1
→
d
1
v
1
v
1
v
2
v
1
v
1
s
2
,
t
2
→
(
)
x
1
v
2
Δ
t
v
2
g
r
p
,
s
p
h
θ
q
,
φ
q
[
]
[
]
Surface map
View map
→
d
3
v
2
v
3
v
2
v
2
(
x
3
,
y
3
)
v
2
v
2
(
s
1
,
t
1
)
v
2
v
2
(
s
3
,
t
3
)
v
2
v
2
(
x
1
,
y
1
)
→
x
2
Δ
t
→
v
2
v
2
d
2
v
2
(
x
2
,
y
2
)
→
d
1
v
2
v
1
→
z
2
v
2
v
2
v
2
→
y
2
(
s
2
,
t
2
)
v
3
Δ
t
v
3
→
g
r
p
,
s
p
h
θ
q
,
φ
q
z
3
[
]
[
]
Surface map
View map
→
d
3
v
3
v
3
v
3
(
x
3
,
y
3
)
v
3
v
3
v
3
→
y
3
(
s
1
,
t
1
)
→
x
3
v
3
v
3
(
s
3
,
t
3
)
v
3
v
3
(
x
2
,
y
2
)
Δ
t
→
d
2
v
3
v
3
v
3
→
d
1
v
3
(
x
1
,
y
1
)
v
1
v
2
v
3
v
3
(
s
2
,
t
2
)
Figure 5.20
Surface maps and view maps for a surface light field. Being vertex based, there is one term
for each triangle vertex. The spatial value for a term in the approximation is stored in a
texture map called a “surface map” (middle column), obtained from the surface coordinates
(
r
,
s
)
at the evaluation point. The directional value is obtained by projecting the view
direction in the local coordinate system at a vertex (the axes are labeled
x
i
,
y
i
,
z
i
)andthe
value is obtained from the “view map” texture (right column).
(From [Chen et al. 02]
c
2002 ACM, Inc. Included here by permission.)
The surface light field is rendered over each triangle by looking up values in
the surface and view maps for the vertices of the triangle, as illustrated in Fig-
ure 5.20. The left column of this figure shows the (view-dependent) rendering
geometry; the surface and view maps are shown in the middle and right columns,
respectively, corresponding to the particular vertex in the left column. The sur-
face map is a straightforward texture map lookup: the position on the triangle is
converted to texture coordinates and the value of
g
is obtained from the matching
texel. Lookup in the view map is slightly more complicated. Unlike the surface
map, which stores the value of
g
directly as a function of position, the view map
gives the value of
h
at the vertex. The view direction at the vertex is computed
