Conclusion - Real-Time Rendering: Computer Graphics with Control Engineering - page 109

8 Conclusion
8.1 PERFORMANCE ANALYSIS
In this section, we provide a qualitative and quantitative analysis on the experiment
results of previous research in comparison to that from our proposed framework.
The analysis is primarily based on three characteristics of the rendering perfor-
mance—the frame rate stability, transient response and adaptive tracking capability.
Prior to discussing the analysis, it is known generally that the performance of
different techniques is best compared by applying them in the same test data-set
or environment. However, this cannot be easily accomplished in this research
because we are not simply comparing an improvement to an existing technique or
algorithm but introducing, establishing and validating a novel rendering architec-
ture. First, the subject matter deals with a 3D rendering approach (polygon-based)
which is vastly different compared with other techniques such as image-based and
volumetric rendering. This means that the rendering setup and data format cannot
be shared or used across the platforms. Second, apart from software configuration
certain research spanning interactive 3D rendering techniques surveyed in this topic
relies on specialized hardware [69,82,85,91] or they work on distributed environ-
ments [61,62,67,72] which contrasts greatly with our rendering framework's setup.
Therefore we deem the comparison to be adequate by referencing the qualitative
and quantitative differences (frame rate stability, transient response and adaptive
tracking capability) between the experiment results from previous research and
our work.
8.1.1 f Rame R ate s taBility
One of the key qualitative metric considered in this research which is important in
real-time 3D rendering is frame rate stability. A stable frame rate does not only bring
about steady visual display that allows positive user experience, it also carries the
benefit of optimised resource usage. This can lead to more effective utilisation of
the computer's processor cycles compared to a “best-effort” technique that does not
guarantee a stable frame rate.
From Figure 8.1, it is evident that Pouderoux and Marvie's technique of stream-
ing 3D terrain data using strip masks [78] did not generate persistently stable frame
rates. Gobbetti and Bouvier's multi-resolution technique [88] to control frame rate
produces very coarse results as shown in Figure 8.2. The lack of strong adherence
to target frame rates is probably most apparent in Jeschke et al.'s [77] research on
using imposters as a means to improve frame rates. It is evident from the experiment
data as shown in Figure 8.3 that this type of approach is not adaptive in nature and it
105