Computer graphics processing and selective visual display system – Computer graphics processing – Attributes
Reexamination Certificate
2001-08-31
2004-03-09
Bella, Matthew C. (Department: 2672)
Computer graphics processing and selective visual display system
Computer graphics processing
Attributes
C345S582000, C345S591000, C345S593000, C345S426000
Reexamination Certificate
active
06704025
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to computer graphics, and more particularly to shadow mapping in a graphics pipeline.
BACKGROUND OF THE INVENTION
A major objective in graphics rendering is to produce images that are so realistic that the observer believes the image is real. A fundamental difficulty in achieving total visual realism is the complexity of accurately representing real world visual effects. A scene can include a wide variety of textures, subtle color gradations, reflections, translucency, etc.
One important way to make images more realistic is to determine how objects in a scene cast shadows and then represent these shadows in the rendered image. Shadows enhance the realism of an image because they give a two-dimensional image a three-dimensional feel.
The addition of shadows to the rendering of a complex model can greatly enhance the understanding of that model's geometry. The human visual system uses shadows as a cue for depth and shape. Consequently, shadows are very useful for conveying the three-dimensional nature of rendered objects and for adding realism to computer generated scenes.
Prior Art
FIG. 1
illustrates a scene to be rendered to include shadows. As shown in Prior Art
FIG. 1
, a particular scene
100
includes a light source
104
, an object
102
, and a surface
106
. Due to the relative orientation of the light source
104
, object
102
, and surface
106
, a shadow should be shown on the surface
106
where the object
102
obstructs light from the light source
104
. Shadow mapping is a technique by which a graphics pipeline may determine which portions of the surface
106
should be rendered with a shadow.
Prior Art
FIG. 2
illustrates a prior art technique
200
of shadow mapping. As shown, a light depth value (Z
light
) is collected for every portion of the surface
106
that is to be rendered. Such light depth value (Z
light
) represents a distance between the light source
104
and the object
102
which may (or may not) be obstructing the light source
104
at a particular portion of the surface
106
that is to be rendered. Traditionally, the light depth value (Z
light
) is gathered from a shadow-map that is generated by rendering during a first pass.
Further calculated is an eye depth value (Z
eye
) that is taken from a perspective of an eye or camera
202
and transformed into the light's perspective. Such eye depth value (Z
eye
) represents a distance between the light source
104
and the portion of the surface
106
that is to be rendered.
As is now readily apparent, when the eye depth value (Z
eye
) is greater than the light depth value (Z
light
), it is assumed that the particular portion of the surface
106
should be rendered with a shadow.
A particular problem arises during the foregoing shadow-mapping procedure due to the fact that the light depth value (Z
light
) and the eye depth value (Z
eye
) are sampled from two different viewpoints, which potentially produce slightly different depth values that may cause incorrect shadow results. Specifically, the light depth value (Z
light
) is obtained from a point in space sampled from the light source viewpoint
104
of the scene
100
, while the eye depth value (Z
eye
) is obtained from a close-by point in space sampled from the point of the eye
202
, and then transformed into the viewpoint of the light source
104
. The two points under comparison are close in space, but still may cause a large enough depth difference with respect to the light source
104
, which produces unwanted shadow results.
Prior art
FIG. 2A
illustrates this shadow-mapping problem.
FIG. 2A
shows a single ray from the light source hitting a horizontal surface and the corresponding texture element (i.e. texel) in the shadow map contained between point ‘a’ and ‘b’.
FIG. 2A
also shows three rays (
1
,
2
, and
3
) from the eye
202
hitting the same surface
106
, which index the single shadow-map texel (all the points inside the [a, b] interval index the same shadow-map texel). Three points on the surface
106
result from the eye viewpoint sampling, and three depth values (Z
1
, Z
2
, and Z
3
) are computed with respect to the light source
104
. The comparison of those three depth values with respect to the depth value sampled from the light source point of view illustrates the problem. Clearly, ray number
2
hits the surface exactly at the same point where the light sampled the surface
106
and the comparison of Z
2
with respect to the light depth value (Z
light
) produces the correct shadow result. However, rays number
1
and
3
may produce opposite results. Ray number
1
produces a z-value (Z
1
) which is smaller than the light depth value (Z
light
) and ray number
3
produces another z-value (Z
3
) greater than the light depth value (Z
light
). The consequence of this difference is that points to the left of the texel center may cause one shadow result, whereas points to the right of the texel center may cause the opposite result, within a given tolerance around the texel center.
In addition to sampling differences between the light and the eye, another problem may also occur. Suppose, in prior art
FIG. 2A
, that Z
2
and the light depth value (Z
light
) were produced using different perspective-interpolation methods (i.e. one with pre-perspective divide and the other with post-perspective divide). In such situation, even though the rays from the light and from the eye sample the exact same point on the surface, the use of different perspective-interpolation methods would cause slightly different depth values, which in turn could cause incorrect shadow results.
Unfortunately, these sampling and interpolation problems result in defects in the shadow-mapping algorithm. In particular, portions of the rendered scene that should be shadowed are not, and vice-versa. A common term for describing such defect is “shadow acne.”
Various solutions have been proposed to compensate for such problem. For example, one technique (a.k.a. Woo Shadow Mapping Algorithm) utilizes the conventional shadow-mapping methodology, but with an altered light depth value (Z
light
). In particular, such algorithm utilizes an average light depth value (Z
light-AVE
), which is positioned halfway between the object
102
and the surface
106
that is to be rendered. (Z
light-AVE
) is calculated from a first light depth value (Z
light-1
) and a second light depth value (Z
light-2
). In regions where a single surface is available, the algorithm uses the farthest depth value in the scene (or its average with the single surface available) as the altered depth value.
Prior Art
FIG. 3
illustrates the various values
300
utilized to carry out the aforementioned shadow mapping algorithm. Table
1
illustrates the equation that may be used for calculating the average light depth value (Z
light-AVE
).
TABLE 1
Z
light-AVE
= (Z
light-2
− Z
light-1
)/2
While this technique works well in theory, there are many practical complications that preclude the present shadow mapping algorithm from being implemented in a hardware-implemented graphics pipeline. For example, the second light depth value (Z
light-2
) is traditionally not available, or incalculable in a graphics pipeline. Without such second light depth value (Z
light-2
), the present shadow mapping algorithm can not be feasibly implemented in a graphics application specific integrated circuit (ASIC), or graphics accelerator.
DISCLOSURE OF THE INVENTION
A system and method are provided for improved shadow mapping in a graphics pipeline. Raw depth values are initially collected from two depth layers in a scene to be rendered. Shadow-map depth values are then calculated utilizing the raw depth values. The scene is then shadow mapped utilizing the shadow-map depth values in order to improve the appearance of shadows in a rendered scene. The various steps are carried out by a hardware-implemented graphics pipeline, which may include texturing or shadowing mapping hardware.
In use, the raw depth values may be collected during an i
Bastos Rui M.
Everitt Cass W.
Kilgard Mark J.
Bella Matthew C.
NVIDIA Corporation
Silicon Valley IP Group PC
Tran Tam
Zilka Kevin J.
LandOfFree
System and method for dual-depth shadow-mapping does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for dual-depth shadow-mapping, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for dual-depth shadow-mapping will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3213425