As outlined in the Work Package 1 overview, task 1.1 addresses the problems that can arise when a user wishes to pass geometry to a radiosity system that is not prepared for the complexities therein. The consortium partners not only had to overcome the complex nature of much geometry, they also had to overcome the diverse number of sources for such input. Concentrating on a popular set of file formats, as was originally proposed, would have restricted the resulting software to only those formats. Ideally, all geometry would be described in a single, all-encompassing, format where any complexities (e.g., invalid normals, complex polygons, non-polygonal objects) could be dealt with in a pre-process.

• Use of a single, all-encompassing format, means that the pre-processing code only has to deal with a single set of well-defined objects.
• Use of a single all-encompassing format also reduces the problem of handling multiple file formats, down to the problem of writing converters into the one, chosen format.

For these reasons, and others, the consortium each adopted the LightWorks ADS as a front-end. The ADS can handle a very large number of geometric primitives (from polygons to trimmed splines), was available for free to consortium members, and had already undergone many years of use in the market sectors that the consortium was trying to address. Task 1.1 then consisted of LightWork Design expanding the already extensive geometric pre-processing features of the LightWorks ADS, to meet the needs of the consortium, whilst the other partners simply had to use the ADS as a front end, accessing its contents via its well-defined API. It is this work, LightWork Design extending the pre-processing functionality, whilst IGD and GRAVIR hook into that functionality, that has occupied the partners during the last year.

The task has most certainly been a success. The decision to use the LightWorks ADS, rather than write custom software, has had a number of advantages:

• Use of commercial-quality, well-tested software, providing a stable interface.
• A much better starting point than we could otherwise have hoped for (much useful functionality predates ARCADE)
• Allowed excellent distribution of work: LightWork Design extend their code, IGD and GRAVIR utilise the results.

That aside, the high-level decision to remove geometry concerns from the radiosity systems, and address them in a pre-process, is also proving effective:

• The radiosity systems require little alteration, as a consequence
• Extensions to the radiosity systems do not require re-writes of the pre-processing software
• Extensions to the pre-processing code can be easily utilised by the radiosity systems.
• The pre-processing software can be used elsewhere, as GRAVIR are now doing inside the SIMULGEN project

As part of task 1.1, LightWork Design have been working on ...

• Geometry pre-processing enhancements
• Triangulation Enhancements

Restricted information.
© LightWork Design Ltd. All information contained in
this document is the copyright of LightWork Design Ltd. (unless stated
otherwise) and may not be used or reproduced with prior written consent.