Terrain texture dump
by Konrad Kiss · in Torque Game Engine Advanced · 07/18/2008 (2:25 pm) · 4 replies
Hi
Is there an easy way / implemented function to dump the blended terrain texture to a file for legacy terrains?
Same for the terrain lightmap?
I am creating a tool that lets people edit missions through their web browsers, and I need some way to actually show where they are editing an npc for example. I'm not worried about statics / interiors and such, so the terrain texture would be just enough for now.
A better solution would be to create a snapshot from above with fog off, 0 volume fogs, large visibility distance and an ortho view, but I had trouble setting the right fov for that, and that sounds complicated enough. And then I'd want the best LOD on everything, so trees don't look like a line from above with their last billboard LOD kicking in.. it could get pretty complicated, so for now, I'd just try a texture dump.
Where should I dig to find this texture object?
Any help would be very much appreciated!
Thanks,
Konrad
Is there an easy way / implemented function to dump the blended terrain texture to a file for legacy terrains?
Same for the terrain lightmap?
I am creating a tool that lets people edit missions through their web browsers, and I need some way to actually show where they are editing an npc for example. I'm not worried about statics / interiors and such, so the terrain texture would be just enough for now.
A better solution would be to create a snapshot from above with fog off, 0 volume fogs, large visibility distance and an ortho view, but I had trouble setting the right fov for that, and that sounds complicated enough. And then I'd want the best LOD on everything, so trees don't look like a line from above with their last billboard LOD kicking in.. it could get pretty complicated, so for now, I'd just try a texture dump.
Where should I dig to find this texture object?
Any help would be very much appreciated!
Thanks,
Konrad
About the author
http://about.me/konrad.kiss
#2
Yes, getting the heightfield seemed relatively easy. I've worked on porting your CommanderMap Hud to 1.7.1, and that helped me to understand a lot about top-down terrain.
I want to present my editors some web eye-candy. That preserves enthusiasm more, and I need them to be excited about the project, so the heightmap just won't do it.
What I'll do then is I'll create a top-down cam similar to the one in your resource, except I'll take out volumetric fog, make the waterblocks less see-through, and fix the lods for objects rendered there to be the best. Once I have a correct ortho view, I'll probably automate the cam's positions to go through the entire map and get it by chunks. Once it iterated through the map, I will exit the script.
I'll work on how to put the chunks together, and finally create a map that's usable for my idea or even in-game. Once that's done, I'll update your ported resource to work from the map file instead of using a cam to take out the 2nd rendering overhead.
Well, that looks like a nice plan for the week. Thanks again for your help.
07/19/2008 (1:49 am)
Thanks BenYes, getting the heightfield seemed relatively easy. I've worked on porting your CommanderMap Hud to 1.7.1, and that helped me to understand a lot about top-down terrain.
I want to present my editors some web eye-candy. That preserves enthusiasm more, and I need them to be excited about the project, so the heightmap just won't do it.
What I'll do then is I'll create a top-down cam similar to the one in your resource, except I'll take out volumetric fog, make the waterblocks less see-through, and fix the lods for objects rendered there to be the best. Once I have a correct ortho view, I'll probably automate the cam's positions to go through the entire map and get it by chunks. Once it iterated through the map, I will exit the script.
I'll work on how to put the chunks together, and finally create a map that's usable for my idea or even in-game. Once that's done, I'll update your ported resource to work from the map file instead of using a cam to take out the 2nd rendering overhead.
Well, that looks like a nice plan for the week. Thanks again for your help.
#3
07/19/2008 (5:17 pm)
Sounds like a good plan. That would make a great resource, I'm sure a lot of people would get mileage out of it.
#4
So far I have managed to set up the cam, take screenshots through the terrain with it, and assemble the shots into one big map with pixel correct positions.
I still have to crop the final image, make it automatically find the correct pixel / unit ratio on the shots so it works with any terrain, I have to skip water reflection renders, and I have not yet poked object LODs. There's still much to do, but it's coming along nice.
I created a new gui element called guiMapDump that is a guiTSCtrl descendant. The resource will probably have to introduce a separate gui for the whole thing. I will try to make it plug&play.
Later on I will also extend the resource to be able to stream these small shots into a bitmap on the screen, so eventually it'd be possible to use cached chunks of images for a new mini map (commander map). This would probably be a lot more faster than rendering the scene again - even if only a few object types are masked in.
I'll post my progress in this thread.
07/20/2008 (1:14 pm)
Sure, I'll post it soon enough, I just need to make it resource worthy first. :)So far I have managed to set up the cam, take screenshots through the terrain with it, and assemble the shots into one big map with pixel correct positions.
I still have to crop the final image, make it automatically find the correct pixel / unit ratio on the shots so it works with any terrain, I have to skip water reflection renders, and I have not yet poked object LODs. There's still much to do, but it's coming along nice.
I created a new gui element called guiMapDump that is a guiTSCtrl descendant. The resource will probably have to introduce a separate gui for the whole thing. I will try to make it plug&play.
Later on I will also extend the resource to be able to stream these small shots into a bitmap on the screen, so eventually it'd be possible to use cached chunks of images for a new mini map (commander map). This would probably be a lot more faster than rendering the scene again - even if only a few object types are masked in.
I'll post my progress in this thread.
Associate Kyle Carter