Viewing new LOD...

[Guest]

I created some new LOD 12 mesh terrain, but the new terrain doesn't appear until I'm almost on top of it. The transition between default and the new mesh is too abrupt. Is there a way to make the new mesh appear when I'm further away from it(further than 4 or 5 miles say)?

[Guest sgreenwood]

Nope! Try recreating it at LOD10 and compare screenshots.Steve

[Guest]

What LOD is most of the default scenery at anyway?Will I see LOD10 before LOD11 and LOD11 before LOD12 etc.? Are there specific distance cut-off points?

[Guest sgreenwood]

>What LOD is most of the default scenery at anyway?My list of FS2002 mesh may help: www.fs-traveler.com/cgi-bin/fs2002-default_mesh.zip>Will I see LOD10 before LOD11 and LOD11 before LOD12 etc.? Higher LOD has priority. Higher source resolution has priority if LOD is the same. LOD12 and above are not used if fs2002.cfg TERRAIN_MAX_VERTEX_LEVEL is the default value of 19 (needs to be 21).>Are there specific distance cut-off points?Higher LOD has smaller radius. So if you have the same area covered with LOD12, LOD11, and LOD10 mesh you should see the mesh as a set of concentric circles, with the highest res LOD12 under the aircraft, ... (I don't know the exact distances, but LOD13 is VERY disappointing!)This also applies for the several LOD levels of default mesh. And is probably why MS supplies at least two LOD levels worldwide.I suggested the LOD10 test because I don't think the small differences justify the much larger file size, performance demands and reduced radius of the higher LOD values. If they increase the radius in the next FS version, however ....

[Guest]

Here's another question since I have your ear, and you seem like a real expert. To your knowledge, is it possible to transform the 10 meter USGS dems to a Lat/long form that would be compatible with fs2002. I tried using MicroDem and resampling my data from UTM to latlong and I get what I suppose is a binary file that Microdem can view. Then I tried saving this file as a USGS ACII dem using MicroDem. I get what looks like a dem file but read_dem.exe can't read it. I have no way of transforming the binary or the new dem into something resample.exe can use.Is any of this making sense? If I use read_dem.exe on the original downloaded dem and then use resample.exe on the resulting binary, the bgl I view has those thin wedges on all four sides where no data was read. The viewer is looking for lat/long to line up with the x and y axes not the UTM coordinates.I'm really banging my head against the wall with this thing. I could really use some advice. I want to put those 10 meter dems to use.

[Guest sgreenwood]

Definitely not an "expert", but I have created quite a bit of mesh. >is it possible to transform the 10 meter USGS dems to a >Lat/long form that would be compatible with fs2002. Yes. I usually save UTM data in DTED format (DT1 files) and then merge them. The result is a MicroDem Dem file based on the geographic grid. This is also required if you want to merge data of different resolutions - merged files must be the same resolution. You specify the resolution in arcsecs when saving the DTED files. See the Help file for more info on merging UTM data.Be sure you have your Microdem Options-Maps set to use WGS84 as the Primary Datum. From the help file----------------------Saving and datums:Saving as DTED will always transform to WGS84. Saving as GEOTIFF will save the image of the current map in the current primary datum. Saving a DEM in MICRODEM format will save in the current primary datum. ---------------------------->I have no way of transforming the binary or the new dem>into something resample.exe can use.Use BIGBSQ (available at http://www.terrainmap.com/ - interesting site) to convert the Microdem format to one the SDK can use. Same concept as read_dem.exe, but for a different source format.BIGBSQ has little documentation, and I don't use it, but I believe it works best if you put everything in the root directory or use a short path and DOS 8.3 filenames. Just run the program and provide the information it requests.

[Guest]

When I bring up the data manipulation window and click on export per the instructions in the help files, I can see no "DEM in DTED" format selection in the export drop down menu. I have "ASCII XYZ,FULL DEM", "Random XYZ", and "DXF Contours" but no "DTED". It's mentioned in the documentation, but I can't find a reference to it anywhere in the program.Actually I think there must be something wrong with my install, because I had to bring up the quick view window even to be able to select "data manipulation".Could you hook me up with an earlier version of MicroDem? Maybe one that isn't so buggy?

[Guest sgreenwood]

It's the Help file that is "buggy", but it is improving. This application is used as an educational tool so the primary audience probably has access to other resources: instructors, related text books, ... No need to Export the data. Just open your data in the program and then use File-Save Dem-DTED to save as a DTED file.If you are unable to view your source data, I'll need to know what format you are using: SDTS, BIL, Tiff,...The Regular menu should have both the In/Out toolbar icon and a File-Manipulation menu option. I'm using the April 2003 and May 2003 versions. (The May version no longer supports the SRTM hgt files. He now only supports the Tiff version of the data available at the NED seamless server. This is a better approach, but I have all the data in hgt files.)

[Guest]

Yep, I'm finding it all. I cleaned up my hard drive and did a reinstall on MicroDem. "Data Manipulation" is back where it should be, but the DTED selection on the export drop down menu is still not there as the documentation says it should be. Strange. But then you said it's the "help" files that are buggy, which would explain things. Apparently, according to the MicroDem web site, a lot of the people who maintained this program are no longer available. Ah well, I guess we just have to do a little reading between the lines.I'm saving my files in DTED format. I'll merge them and then convert them using BIGBSQ, then see if resample.exe can handle the bsq. I'll let you know if it works.In the data manipulation window there is a "resample" drop down menu that apparently allows you to resample data as Lat/Long or DTED. Is this the same as using resample.exe in the fs SDK? How is resampling as DTED different from saving as DTED in MicroDem?In your previous message you mentioned that you never really use BIGBSQ. What is your proccedure then for converting the merged dem?Thanks for all the leads, by the way. You've been a real help. Now all I have to do is master writing inf files for resample.exe. That's still a work in progress.

[Guest]

Did you mean that you don't use BIGBSQ "documentation" or the BIGBSQ program itself? I ask because I got no documentation at all. All I have are the prompts the program provides.

[Guest sgreenwood]

>In the data manipulation window there is a "resample" drop >down menu that apparently allows you to resample data as Lat/Long >or DTED. Is this the same as using resample.exe in the fs SDK?No. In Microdem, "Resample" means change the spacing of the data. It is more like using different LOD values in the SDK Resample. The SDK resample does several other tasks as well. Such Resampling is necessary, for example, before merging 10m data with 30m data, to fill in holes, or whatever. All merged data must be at the same resolution or it will be omitted during the merge process.>How is resampling as DTED different from saving as DTED in MicroDem?Probably no difference - the Help file suggests that it is still saved as a dted file. I believe it can only resample using degrees and not meters. You get the same opportunity to specify the output resolution either way. >In your previous message you mentioned that you never really >use BIGBSQ. What is your procedure then for converting the>merged dem?Well... OK. I've written my own program to convert the DEM directly to a bgl file, or to process it in steps. Since I load the data set into an array, I can do additional custom processing before converting it. I have all the header information on screen, so it is easy to use this directly to create the inf and batch files with code. Problems encountered with various forms of data have resulted in the following, and a bit more. The list keeps growing. :(I routinely check for missing values and extreme values. The total changed and missing value counts are updated on-screen during processing, giving me a feel for the quality of the data. I can change the parameters that determine what is "extreme", because locale and resolution of data matter here.I can interpolate data across merged dem boundaries when they do not align correctly (or anywhere, but this is the most common use). You will probably see cliffs in these situations when you merge your own data. They can be seen in Microdem if you Zoom in enough. A bit of good news here. I've processed all the SRTM data outside the US and have not noticed this effect anywhere. When they release the cleaned up version without the holes, life will be much easier.Within rectangles, I can change any given elevation to any other value, all values to another value, or set all values above a cutoff point so a new value.I have an Edit screen, which works much like the cell array in Microdem, but I can move about more easily in it.Recently added feature: When I specify the LOD value, it shows me the size of the LOD cells and the border coordinates of the final mesh. Useful for assuring complete coverage before constructing the bgl.There is a bit more, but you get the idea. Next on the list, maybe: I would like to be able to use the LOD information to trim the data to just enough to cover the area desired, without having to create subsets in Microdem in order to avoid overlapping bgl files.>Thanks for all the leads, by the way. You've been a real help. You're welcome.>Now all I have to do is master writing inf files for resample.exe.>That's still a work in progress.That won't take you long, then it just becomes tedious.

[Guest]

>I can interpolate data across merged dem boundaries...Yeah, I wish I could do that. I thought I had it all together. I got my four quads merged. I converted the resulting dem to bsq. I got my destination boundries precisely defined according to the lat. and long. boundries in the SDK for LOD12. I resampled and when I viewed the resulting bgl in the tmfviewer, the boundries of the image were exactly the boundries I was asking for. Hooray!! Unfortunately, right along a merged boundry in my resampled area is a very noticable seam or perforation. This is even more noticable at the intersection of the four merged files. It reminds me of the perforations you might see in a block of stamps, but not quite so pronounced. I didn't notice it in MicroDem but it really shows up in tmfviewer, and it would really show up in a flyby.Big disappointment! Will the picky little probs. never cease? Is it possible to use MicroDem to clear up this prob.?It's like the wild west out in this cybermapping wasteland. No standards, what seems like a thousand different formats. If it weren't for this forum and people like yourself, I'd really be feeling around in the dark.Microsoft sure could have made things go a lot smoother with SDK documentation that was more comprehensive.The next great hurdle: getting rid of those seams. After that: making lake elevations match up with my new mesh terrain. I can already see they're a bit off.You should upload this program of yours to the file library.

[Guest sgreenwood]

>Big disappointment! Will the picky little probs. never cease?With a little experience, creating mesh becomes fairly straigtforward. But even with experience and good tools, it can take a great deal of time to turn some data into good mesh. At the very least, once suspicious areas are identified with Microdem, they must be verified by flying over them in the sim. And once corrected, they must be flown over again to confirm that the corrections are satisfactory. And sometimes these boundary problems extend for 2, 3, or more degrees!With 10m mesh, where these problems are common, I fly over all 15 second quad boundaries, and this can take a very long time when the areas are as large as the entire states of New York or New Jersey!>Is it possible to use MicroDem to clear up this prob.?Only some of it. With your data, converting to lat/long DTED format before merging (at least before merging data that crosses UTM Zones) is important. Either during the merge or after, you should have Microdem fill some of the remaining holes with the Edit-Fill Holes menu command, then Edit-Missing data to sea level. I don't know if other GIS software can do a better job with this task; perhaps others can shed some light here.>I didn't notice it in MicroDem but it really shows up in tmfviewerUse the Modify-Map Area-Zoom In menu command to increase the magnification of the image. Then click the Info menu item to determine the screen pixel size. When each pixel represents an area of about 250m or less, you should be able to see most of the problems that may need correction. After a while you develop a feel for what will show up in the sim and what may not.Once you have corrected the data, you may want to repeat the procedure to see what your repairs look like. You can use the Subset & Zoom toolbar button to magnify small areas.I work with Options-Maps-default display set to Reflectance view, with the Green/Blue color option. I also set the Missing Values color to bright Yellow. These settings reduce distractions in the image and make identifying problems much easier for me.

[Guest]

Oh man! Where do I start?First of all, let me tell you how I did my merge and see if it meets with your approval. I downloaded 6 10 meter USGS dems in SDTS format. At first I did the merge per your instructions: convert to DTED then merge. This produced the many holes along the merge borders as I described in my last message. Then I decided to use the resample UTM to Lat/long feature. I converted each of the 6 original SDTS quads to lat/long form in this way, and then did the merge. The results were much better. In the whole 6 quad merge, I could see only one small hole in what was otherwise a perfectly clean tmfviewer image of the bgl file. I like using tmfviewer because the defects really pop out at you. I was encouraged--thought I had it made. When I did a fly through everything looked great. What detail! Gradually, I noticed trees that were floating where they shouldn't be, rivers that were running along the walls of their basins instead of inside them, lakes shifted over to one side of the basins that should contain them, and cliffs--everywhere cliffs on the "outside" borders of my resampled area(nowhere in the interior merge borders mind you, just on the outside borders). It's as if my whole resampled area has been shifted about a tenth of a mile in what looks like a northwest direction.I'm assuming that the default roads and lakes and rivers are right on the money as far as long/lat positioning goes, and that "I'm" the one that's off target. I expected some problems with lake "elevation" disagreements and problems meshing with differing elevations in the default terrain, but I didn't expect a lat/long disagreement.Do you think the problem is in the quality of the merge or just the quality of the original data itself? I can't believe the original data could be off by so much.Could the problem be in my entries in the inf file? I'm thinking in terms of how I calculate x and y cell dimensions. I'm not sure I'm doing that quite right.Enough said for now. I'm going to do some more fly throughs and see what's up. I think my priority now is just to get everything lined up properly. I can worry about lines and holes and other artifacts later.Hope to hear from you.

[Guest sgreenwood]

This discussion is probably getting far too specific to continue on-line. e-mail me at support@fs-traveler.com and we can continue from there. Steve