DTS on DIF Cliff Face
by Ben Acord · 08/24/2007 (5:07 am) · 5 comments
Thanks to all of you who provided feedback to my initial post about cliffs and caves.
Over the past few days I've run through a host of tests and want to give update on some findings. The following screenshots are from a very dense mesh (1,600 faces) as a DTS static shape on a Constructor box then exported as a DIF. The base DIF was then copy/pasted several times to form the full cliff face within the mission. While it's not ideal -remember it's a test- I wanted to see how it performed. Mission loading with just this face is horrid because of lighting.
For now I'm not 100% sold on using DIF for cliffs. It seems like a lot of overhead. Keep in mind that I'm talking cliffs not canyons here. The difference is an individual section verses an enclosed side wall and floor combo.
9:30 AM edit: Here are the screenshots for DTS only objects.
Performance sucks big time. It looks like the best thus far is to have a DIF base object with DTS detail. Bring on the "I told you so's". :)
Over the past few days I've run through a host of tests and want to give update on some findings. The following screenshots are from a very dense mesh (1,600 faces) as a DTS static shape on a Constructor box then exported as a DIF. The base DIF was then copy/pasted several times to form the full cliff face within the mission. While it's not ideal -remember it's a test- I wanted to see how it performed. Mission loading with just this face is horrid because of lighting.
For now I'm not 100% sold on using DIF for cliffs. It seems like a lot of overhead. Keep in mind that I'm talking cliffs not canyons here. The difference is an individual section verses an enclosed side wall and floor combo.
9:30 AM edit: Here are the screenshots for DTS only objects.
Performance sucks big time. It looks like the best thus far is to have a DIF base object with DTS detail. Bring on the "I told you so's". :)
About the author
#2
08/24/2007 (1:47 pm)
Stephen is Correct. DTS is not converted into Diffs. With good LOD you should be just fine using DTS shapes.
#3
By linking the DTS and DIF in Constructor I'm testing if a DTS skin will benefit from a custom DIF backbone. One that matches the unique form of the DTS. A TGE in-editor DTS attach would not have the matching DIF shape. DIF keeps all its benefits and the DTS simply rides along for looks. I was surprised that combining DIF and DTS didn't have worse perf than the DTS by itself.
I haven't tried the QuArk terrain module. Maybe later. Right now I'm trying to keep the tool set limited to Blender and Constructor.
It's worth noting that using Constructor's brushes alone would produce excellent mechanical cliffs and caves; like those futuristic, metal type.
08/24/2007 (1:53 pm)
Thanks Stephan. By linking the DTS and DIF in Constructor I'm testing if a DTS skin will benefit from a custom DIF backbone. One that matches the unique form of the DTS. A TGE in-editor DTS attach would not have the matching DIF shape. DIF keeps all its benefits and the DTS simply rides along for looks. I was surprised that combining DIF and DTS didn't have worse perf than the DTS by itself.
I haven't tried the QuArk terrain module. Maybe later. Right now I'm trying to keep the tool set limited to Blender and Constructor.
It's worth noting that using Constructor's brushes alone would produce excellent mechanical cliffs and caves; like those futuristic, metal type.
#4
You are using the .dif from constructor to generate a finer collision box for the .dts embedded? I assume the .dts has no collision box yet?
What about shadow casting? it must be broken in that case. The shadow being generated from the collision box of the .dts...
The point in using QuArk terrain module is that we would certainly come with real .dif structure instead.
I really need to give it a try.
08/24/2007 (2:04 pm)
Correct me if I'm wrong.You are using the .dif from constructor to generate a finer collision box for the .dts embedded? I assume the .dts has no collision box yet?
What about shadow casting? it must be broken in that case. The shadow being generated from the collision box of the .dts...
The point in using QuArk terrain module is that we would certainly come with real .dif structure instead.
I really need to give it a try.
#5
08/24/2007 (8:23 pm)
The terrain in quark is not optimal for Torque. And i am not sure if its directly usable.
Torque Owner Stephan (viKKing) Bondier
It's only an embedded .dts, not a .dif; model structure is different.
Not sure this test is worth anything, better import the .dts directly instead IMHO.
I'm interested in your experience anyway, so thanks for sharing. 8-)
I'm wondering what the "terrain" module within QuArk is producing. Either tried it?