Computer image generators have reached a level of maturity in terms of the image realism they can achieve. The recent addition of advanced hardware texture to surfaces has provided unsurpassed realism, allowing the creation of much more effective and diversified visual cues. As a result, CIG systems can now be applied to a broader class of training missions.

Because the scene realism is now adequate for many training missions, the CIG data base content problem is becoming increasingly important. New training tasks such as mission rehearsal and nap-of-earth flight require data bases with an increasing geographic extent and density of three-dimensional objects. To meet the new requirements, many data base management and compression schemes are being devised and new hardware capabilities will be created to support them. The problem for the user then becomes one of determining which of these approaches meets his needs.

This paper will discuss and show through example the need for data base management and compression. In addition, it will discuss the trade-offs that occur because of compression schemes. Last and most important, this paper will suggest several methods of specifying data base requirements that take into account such parameters as geographic extent, specific content, density and correlation. The goal of the specifications is to allow the user to more readily understand the data base processing performance of a computer image generator and to reduce the confusion caused by the various management and compression schemes.