Problems with Maui mesh in the AVSIM File Library

This data is of no value as mesh when used in the sim. However, it does present an opportunity to illustrate several mesh-related issues. Disclaimer: I e-mailed the developer the day this mesh was uploaded to AVSIM. I pointed out the problems described below and was told that he would have the file removed from the Library. That was over a week ago and there have been over 1150 downloads of this 31MB file to date. (My apologies to the over-enthusiastic developer; no disrespect is intended. Perhaps we can derive some benefit from your efforts.)1) What mesh data looks like:http://forums.avsim.com/user_files/41993.jpgThese images are views of two mesh (bgl) files, displayed with tmfviewer.exe (included in the MS FS2002 SDK). The application status bar provides a quick reality check, providing latitude, longitude, and elevation information as the cursor is moved over the mesh. All developers should check their mesh with this tool as a part of their quality checking, preferably even before testing the mesh in the sim. Files should be checked individually and adjacent files should be checked together, to confirm that there are no gaps between them. Others should also consider using this tool to check out their add-on mesh, especially when they are unfamiliar with the work of the developer.The image on the left is the mauimesh.zip data. The red areas represent negative elevations (most of it thousands of meters below sea level). The white areas represent terrain above sea level (up to 100,000 feet!).The image on the right is typical of any 10m10 Maui mesh (my own, in this case). This is what mesh should look like. Blue water at sea level and gradual transitions of color/elevation from sea level to higher elevations.This simple check would have been enough to indicate there are serious problems with the data. The project should have gone no further before resolving them.2) Mesh file size:Interpreting mesh file sizes does require some experience with mesh. Increasing the LOD by 1 increases the number of elevations points by a factor of 4, although file compression will reduce this factor somewhat. The expanded size of the Maui data is over 43MB. Valid LOD10 mesh constructed from 10 meter data produces an 800KB file. Clearly the Maui mesh was constructed using a much higher LOD. (0.8MB * 2^3 = 64MB, suggesting an LOD of 13)3) Mesh rendering:Mesh constructed with an LOD greater than 10 is not used by the sim if the TERRAIN_MAXIMUM_VERTEX_LEVEL (in the TERRAIN section of fs2002.cfg/fs9.cfg) is left at the default value of 19. This mesh does not show up in the sim with the default setting, further evidence that the claimed LOD of 10 is incorrect. (This is probably why I have not noticed any other comments about this data. I assume those who have tried it out saw only the default mesh.)When the value of this parameter is raised to 20, the mesh is displayed. The following images were obtained using that setting. 4) LOD and mesh radius:http://forums.avsim.com/user_files/41994.jpgThis image is a close-up of the GPS screen in the next image. We are heading SE towards Maui and are about 5 nm offshore. The blue circle represents the area covered by the "10m10" mesh under the aircraft. The remaining mesh is the default mesh included with the sim. At this point, the circle has just reached the coastline, as seen in the next image. (The "circle" of mesh is composed of rectangles and so is not actually round - but you get the idea.)A brief diversion: LOD determines the radius of the mesh which is actually visible at any one time. Higher LOD mesh is rendered with a smaller radius. LOD13 mesh has a radius of about 5nm, the approximate distance from the aircraft to the sliver of mesh rendered on the coast ahead. This is further evidence that 13 is probably the true LOD for this mesh. (LOD10 mesh has a radius of about 30-40nm, and would cover much more of the island with the aircraft at this location.)5) Finally, one clear indicator of faulty mesh:http://forums.avsim.com/user_files/41995.jpgHere we see the mesh (the pillar rising into the clouds) just starting to show up. It is present under the aircraft, but is flattened by the water, giving us a unique opportunity to demonstrate this effect.http://forums.avsim.com/user_files/41996.jpgAs we move a bit closer, the circle of mesh under the aircraft covers a bit more of the island, revealing more of the faulty mesh.http://forums.avsim.com/user_files/41997.jpgAt this point, we are far enough inland to be surrounded by mesh on all sides. The aircraft is now at 100,000 ft, so you can see a bit of the world beyond. At lower elevations, the aircraft is located at the bottom of a very deep black hole. Note that the pattern visible in the GPS Terrain view reflects the pattern seen using tmfviewer.exe.I hope this information will be helpful for both new mesh developers and users of mesh.If you would like to compare this mesh with true LOD10 mesh constructed from 10 meter source data, stop by my site and download my freeware mesh for the Hawaiian Islands (including Maui). Only 5.4MB for all the major islands.www.fs-traveler.com/products-f.shtmlRegards,Stevewww.fs-traveler.com