Physically Accurate Lighting Simulation in Computer Graphics Software

A.B. Khodulev, E.A. Kopylov


Prev Contents    Introduction    Lightscape System    Specter System    Radiance System    Scenes for Comparison    Experimental Results    Conclusion    Acknowledgments Next

6. CONCLUSION

The results of our analysis are summarized in the table below. In this table we use signs "+" or "-" to denote presence or absence of some feature; "+/-" means presence with restrictions. Numerical values: 1, 2, 3 mean relative order of systems from the viewpoint of some criterion (3 - the best, 1 - the worst).

criterion LVS Specter Radiance
Comprehensiveness of local model
light sources 1 2 2
material attributes 1 2 3
Completeness of global light propagation model
specular to diffuse - + +/-
long light paths * +/- + +/-
accuracy 1 2 3
visual image quality 1 3 2
User interface features
result presentation 1 3 2
interactivity 3 2 1
total ** 9 18 15

Table 8. Overall comparison.

Comments:

* "Long light paths" means capability of simulating arbitrary long sequences of ray-surface events. This capability seems to be restricted in LVS (otherwise we could expect more accurate results for cube) while in Radiance it is available under user control; this is the reason for "+/-" estimates.

** We calculated the total by summing up all the values in columns counting "-" as 0, "+/-" as 1, and "+" as 2. The greater the resultant value the better the system. Of course, reader may assign different (non-uniform) weights to rows of the table according to his/her preferences thus obtaining different total scores.


Prev Contents    Introduction    Lightscape System    Specter System    Radiance System    Scenes for Comparison    Experimental Results    Conclusion    Acknowledgments Next

Copyright 1996 Andrei B. Khodulev, Edward A. Kopylov.- All Rights Reserved [Home]