Plan for Adrian Tysoe
by Adrian Tysoe · 08/24/2005 (1:16 am) · 12 comments
It's been a while since I posted anything, Over the last couple of months I've been learning the basics of TGE and working with someone on and off on a possible new art path.
In the past week we have been discussing integrating the art pipeline into a new Leadfoot Productions Game rather than spending a lot of valuable time working on a tech demo. We started the game in earnest just over a week ago. Most of my time has until now been spent on producing simple meshes that test various material properties, and trying to trouble shoot bugs found in the loader when loading old art assets into TGE from previous projects like Fun racer and Aerial Antics.
I thought I'd share a couple of early pictures of the first Arena for the new Leadfoot game "King of the mountain".
Here's a couple of screenshots of the Arena in TGE, I threw in some hastily rendered lightmaps as one of the nice features of the new art pipeline is that it enables TGE to do 8 multitexture layers with various blendmodes and multiple UV's outside of TGE, so you can do lightmaps and multipass effects using 3rd party tools in this case I used 3dsmax.
As you can see, still pretty rough but already looking quite good.
Future plans in consideration are to redoing Aerial Antics in TGE and release for Mac and Linux and possibly creating A sequel that makes the best of the new tools and the superior networking that torque offers over our previous engine.
In the past week we have been discussing integrating the art pipeline into a new Leadfoot Productions Game rather than spending a lot of valuable time working on a tech demo. We started the game in earnest just over a week ago. Most of my time has until now been spent on producing simple meshes that test various material properties, and trying to trouble shoot bugs found in the loader when loading old art assets into TGE from previous projects like Fun racer and Aerial Antics.
I thought I'd share a couple of early pictures of the first Arena for the new Leadfoot game "King of the mountain".
Here's a couple of screenshots of the Arena in TGE, I threw in some hastily rendered lightmaps as one of the nice features of the new art pipeline is that it enables TGE to do 8 multitexture layers with various blendmodes and multiple UV's outside of TGE, so you can do lightmaps and multipass effects using 3rd party tools in this case I used 3dsmax.
As you can see, still pretty rough but already looking quite good.
Future plans in consideration are to redoing Aerial Antics in TGE and release for Mac and Linux and possibly creating A sequel that makes the best of the new tools and the superior networking that torque offers over our previous engine.
About the author
#2
Looking fatastic by the way. I haven't even seen it in that form just yet ;)
08/24/2005 (1:38 am)
Yeah, the collisions are in as well as some optimizations.Looking fatastic by the way. I haven't even seen it in that form just yet ;)
#3
08/24/2005 (4:27 am)
Nice work, Adrian. :)
#4
The original polysoup implementation that we tried was too slow in Multiplayer.
Then we tried DIF's which had the speed, but converting the GLB exported .map files from max via map2dif to TGE was problematic with slightly offset verts. So we would have to wait for Matt to release a new version.
Today the lead programmer on the loader handed us a new polysoup collision system to try which were sorting out as I write this.
One interesting this we discovered this past week was how much difference the compiler can make. Jeremy used the standard Visual studio 2003 originally, after updating the compiler to the pro version wegained a whopping extra 30FPS!!!!
08/24/2005 (5:32 am)
Hey thanks guys :). Tom, the collisions and transposing coordinates from D3D to GL have been the tricky parts. The original polysoup implementation that we tried was too slow in Multiplayer.
Then we tried DIF's which had the speed, but converting the GLB exported .map files from max via map2dif to TGE was problematic with slightly offset verts. So we would have to wait for Matt to release a new version.
Today the lead programmer on the loader handed us a new polysoup collision system to try which were sorting out as I write this.
One interesting this we discovered this past week was how much difference the compiler can make. Jeremy used the standard Visual studio 2003 originally, after updating the compiler to the pro version wegained a whopping extra 30FPS!!!!
#5
Have you considered going with some simple primitives placed under a special node in the scene as the collision data? Seems like that would work just as well, be quite flexible, and give you a substantial improvement in performance over pure polysoup solutions.
08/24/2005 (5:44 am)
Well the standard edition of VC++ does not include the optimizing compiler... so in that case it's not too strange.Have you considered going with some simple primitives placed under a special node in the scene as the collision data? Seems like that would work just as well, be quite flexible, and give you a substantial improvement in performance over pure polysoup solutions.
#6
08/24/2005 (7:17 am)
Yeah, what's messed up is that they give the optimizing compiler away for free anyway! I submitted a web resource for the free upgrade ... should get up sometime soon. Still, who would think the compiler could give you 35 fps ... like the compiler it comes with would be that useless. These 35 fps weren't going from 300 - 335 ... we're talking 40 - 75 fps. So almost a 100% increase in a certain instance. I think it's a bit ridiculous. Especially, when you are made to think you need to spend almost a grand to get it. 
#8
--Mike
08/24/2005 (1:08 pm)
this does look verrrrry nice though... even though it's a wip, congrats on some nice looking graphics...--Mike
#9
Never going to get round to trying it in earnest at this point though :)
08/24/2005 (2:25 pm)
I still say that matching a dts style polysoup mesh with a DIF based collision rig would be the ideal solution. The DIF would do all collisions quicky, but the DTS would have all the pretties.Never going to get round to trying it in earnest at this point though :)
#10
There are some advantages to polysoup collisions like having single sided open concave surfaces with far fewer triangles than the dozens of small closed covex hull shapes required to produce the same result.
08/24/2005 (2:51 pm)
You can mix and match collision types as you please. There's nothing stopping you from using one method or another or both together. Just depends on what tools your using and which you prefer or is more convenient in a given situation. DIF is pretty easy to work with in Max at least, so long as you buy the commercial release of GLB. Easier than ever now that Matt's new map2dif parses the GLB .map files without having to use a CSG editor to convert to valve220 format :) There are some advantages to polysoup collisions like having single sided open concave surfaces with far fewer triangles than the dozens of small closed covex hull shapes required to produce the same result.
#11
08/24/2005 (7:31 pm)
What many have been looking for is DTS with multiple UVs and no limit on the number of collision primitives (currently i think it's 9). While i agree that one could make a collision only DIF and some mesh shape and combine them to get a final result, i don't think that makes for a quick and simple art path. Ideally one could work in Max/Maya completely when working with something like what you've created here.
#12
08/24/2005 (9:54 pm)
Ah, there was already a resource for it. I didn't find it seaching this site when I had the problem ... but cool that it's in there.
Associate Tom Spilman