7. Advanced Shading (17/17)

Search in book...
Toggle Font Controls
Create new playlist

Name your new playlist

Playlist description (optional)
Sign In

Email address

Password

Forgot Password?

or

Continue with Facebook

Continue with Google
Sign Up

Full Name

Email address

Confirm Email Address

Password

or

Continue with Facebook

Continue with Google

7.9. Combining Lights and Materials 281

Since the light shader is separate from the surface shader, each shader can

be used for a longer string of objects, improving eﬃciency [1410].

Another important advantage has to do with programming and assets

management costs. With the traditional approach, the vertex and pixel

shader programs retrieve each light’s and material’s parameters and com-

pute the eﬀect of one on the other. The deferred shading method of ren-

dering allows a strong separation between lighting and material deﬁnition.

In our example application, a deferred shading approach would need 11

shaders (one for each material type to compute the eﬀect of the material

and store the relevant parameters in the G-buﬀers, plus one for each light

source type) instead of the 25 shaders required by a multipass lighting

approach or the 1050 required by an ¨ubershader approach. The separa-

tion of lighting and material should not be undervalued. This split makes

experimentation simple, as instead of numerous combinations, only one

new shader needs to be added to the system for a new light or material

type [691].

Another advantage of deferred shading is that the time-consuming pro-

cess of determining exactly which lights aﬀect which objects is avoided. In

deferred shading, the model is established in space and lights are applied to

it. Shading can be optimized to be computed only in areas where the light

source has any possible eﬀect. Local light sources aﬀect only small areas

of the screen, and various geometric entities and rendering modes can be

used to determine which pixels to shade. For example, point lights can be

treated as spheres and spotlights as cones to determine what pixels they

aﬀect. Only those pixels where surfaces are inside of the volume of eﬀect

need to be shaded [503, 1288]. Even a directional light (e.g., the sun) can

be optimized by marking the stencil buﬀer on the initial draw to shade

only pixels aﬀected by it [1174]. Local light sources can be culled based on

area of eﬀect or amount of contribution, or can be combined if found to

aﬀect similar areas of the screen [1018]. Fifty or more dynamic lights in a

scene can easily be handled by deferred shading; a traditional renderer can

handle only a fraction of that number at the same frame rate [691].

Deferred shading does have some drawbacks. The video memory re-

quirements and ﬁll rate costs for the G-buﬀers are signiﬁcant, especially

when programming for consoles.

The most signiﬁcant technical limita-

tions are in the areas of antialiasing and transparency. Antialiasing is not

performed and stored for rendered surfaces. To overcome this limitation,

Shishkovtsov [1174] uses an edge-detection method for approximating edge

coverage computations. Supersampling to perform antialiasing is also pos-

sible, but the increase in ﬁll rate makes this approach expensive [1018].

Transparent objects must be drawn using screen-door transparency or ren-

For example, the Xbox 360 has 10 megabytes of render-to-texture memory.

282 7. Advanced Shading

dered into the scene after all deferred shading passes are done, since only

one object per pixel can be stored in the G-buﬀer. From an eﬃciency

standpoint, if the lighting computations desired are relatively simple, over-

all speed can be worse because the costs of writing and reading the render

targets outweigh any gains from deferring lighting computations.

Even with these caveats, deferred shading is a practical rendering

method used in shipping programs. It can handle large numbers of complex

light sources and naturally separates lighting from materials. In deferred

shading, the whole process of rendering becomes a series of post-processing

eﬀects, once the G-buﬀers are established.

Variants on the G-buﬀer approach are possible. One problem with

deferred shading is the limitation on what can be stored per pixel. Even

with the greater number of render targets available with newer GPUs,

there is an imbalance in how much is needed at each pixel. Some surfaces

may need only 6 values stored to evaluate their shader formulae, while

others could use 20 or more. To compute and store 20 values for all visible

surfaces is then a waste of bandwidth. One hybrid approach is to save

out some limited set of values, e.g., depth, normal, diﬀuse, and specular

contributions. Apply lighting, then render the geometry in the scene again,

this time with shaders that draw upon these common results. In this way,

some of the performance increase from deferred shading can be achieved,

while also providing a wider range of shader choices.

For more ﬂexibility and little additional rework to existing code, an-

other approach is to perform an early z pass. Applications with complex,

expensive shaders and high depth complexity can beneﬁt by performing an

initial pass and establishing the Z-buﬀer [484, 946, 1065, 1341]. After the

Z-buﬀer is created, the scene is rendered normally with full shading. In this

way, there is only one shader evaluation performed per pixel, for that of

the closest surface; no time is wasted evaluating shaders for fragments that

are hidden by surfaces rendered later. The cost is that the scene’s geom-

etry is rendered an additional time, though the prepass z-depth rendering

is so simple that it can be extremely rapid. There are some limitations

on pixel shaders for this optimization to work. See Section 18.3.6 for de-

tails. Obtaining and storing the z-depths also turns out to be useful for

other rendering techniques [887, 1341]. For example, fog and soft particle

techniques, discussed in Chapter 10, use the z-depth.

An interesting variant is presented by Zioma [1410]. Instead of ﬁrst

evaluating the geometry, the areas of eﬀect of local light sources are ren-

dered to and stored in buﬀers. Depth peeling can be used to handle storing

the information for overlapping light sources. The scene’s geometry is then

rendered, using these stored light representations during shading. Simi-

lar to traditional deferred shading, eﬃciency is improved by minimizing

switching between shaders. One major advantage that this approach has

7.9. Combining Lights and Materials 283

over deferred shading is that surface materials are not limited by what is

stored in the G-buﬀers. Also, storing light source descriptions can take

considerably less space than storing a full material description, so memory

and ﬁll rate costs can be reduced.

Table of Contents for 7. Advanced Shading (17/17)

Create new playlist

Sign In

Sign Up

Table of Contents for
7. Advanced Shading (17/17)