Cathedral: Week 3
by Greg Findlay · 03/06/2006 (7:53 pm) · 13 comments
Another week has gone by on this project and unfortunately it wasn't the most productive one. I was hoping to be starting the nave (central hall) this week but it looks like I'll only have time to finish the transcept (entrance). Oh well... I knew I was going to have a few slower weeks but having one early is never a good thing. The good news is that almost all the pieces that I've remodelled with more detail have been unwrapped and are ready for texturing. I did a quick test to see what bump mapping could do for the detailing and the result were not bad. I'm thinking I might normal map the whole thing so the test isn't really what the end result is going to be like but it'll be pretty similar. Here is the progress:
About the author
Recent Blogs
• Cathedral: Week 4• Cathedral: Week 2
• Art project: Cathedral
• Plan for Greg Findlay
• Plan for Greg Findlay
#2
03/06/2006 (8:11 pm)
Looks better each week. Just wondering though, how many polys?
#5
03/06/2006 (8:59 pm)
WoW!!!
#6
03/06/2006 (10:53 pm)
Looking good man. I'll definitely be interested to see what it looks like fully detailed out.
#7
Tool of choice is Maya.
03/07/2006 (6:22 am)
Current poly count is around 13k, so a lot. If it's under 100k poly's by the end I'd be very very surprised (actually I wouldn't be completely surprised if it hit 1 million). It's probably going to run like ass in game but the scale is so large that I'll have to see how it goes. You'll never really be able to get more then say 10k polys on screan at once. It's more of a "let's see what Torque can do" then anything.Tool of choice is Maya.
#8
03/07/2006 (6:26 am)
Looks very nice!
#9
1) Torque culls dts objects off the screen if the centroid of the object isn't "on screen". What this means is that basically any time you aren't facing the center of your dts object as a whole, you'll probably see the whole thing dissapear. This can be disabled (don't have the info off the top of my head), and may actually have been done so in 1.4--I'm not sure!
2) While TGE 1.4 has now allowed unlimited collision hulls for dts objects (used to be limited to 8), be verycareful implementing them. My gut tells me that even as complex as this object is, rendering probably isn't going to be your performance hog, collision calculations will be. Instict will probably have you wanting each of those columns individually collide-able, etc., but that's going to lead you down a performance hog path right quick.
FYI, that cathedral looks damned impressive! I had a Torque community member do something very similar for some prototype art on my old project and his work was outstanding--but with your high poly counts (he was told to limit to less <2,500 I think), yours is absolutely beautiful!
03/07/2006 (8:01 am)
Sounds as if this is .dts, so a couple of things you are going to run into very quickly when you get it into Torque:1) Torque culls dts objects off the screen if the centroid of the object isn't "on screen". What this means is that basically any time you aren't facing the center of your dts object as a whole, you'll probably see the whole thing dissapear. This can be disabled (don't have the info off the top of my head), and may actually have been done so in 1.4--I'm not sure!
2) While TGE 1.4 has now allowed unlimited collision hulls for dts objects (used to be limited to 8), be verycareful implementing them. My gut tells me that even as complex as this object is, rendering probably isn't going to be your performance hog, collision calculations will be. Instict will probably have you wanting each of those columns individually collide-able, etc., but that's going to lead you down a performance hog path right quick.
FYI, that cathedral looks damned impressive! I had a Torque community member do something very similar for some prototype art on my old project and his work was outstanding--but with your high poly counts (he was told to limit to less <2,500 I think), yours is absolutely beautiful!
#11
Something I was considering was creating a .dif collision surface and hide it just under the .dts. I haven't looked into it at all (and I think you'd know a lot more about this then I would), it could just be a waste of effort but I know collision detection is faster for .dif objects and with something with this much detail having a little bit more complex collision would be nice. Even with that though there are going to be a lot of areas that will end up having the invisible wall effect which is unfortunate. I also have the option of exporting objects individually. Most objects are divided up and could be repositoned in the game editor to line up properly. That would also mean I could just have an instance of each smaller piece that could be replicated through out the cathedral.
@ Matt
Hadn't seen that VRND project before... it looks cool. It's interesting to see how far things have come since that was released and to see all the ways things have changed for faking what you see.
Thanks for the comments everyone!
03/07/2006 (3:35 pm)
@Stephen Something I was considering was creating a .dif collision surface and hide it just under the .dts. I haven't looked into it at all (and I think you'd know a lot more about this then I would), it could just be a waste of effort but I know collision detection is faster for .dif objects and with something with this much detail having a little bit more complex collision would be nice. Even with that though there are going to be a lot of areas that will end up having the invisible wall effect which is unfortunate. I also have the option of exporting objects individually. Most objects are divided up and could be repositoned in the game editor to line up properly. That would also mean I could just have an instance of each smaller piece that could be replicated through out the cathedral.
@ Matt
Hadn't seen that VRND project before... it looks cool. It's interesting to see how far things have come since that was released and to see all the ways things have changed for faking what you see.
Thanks for the comments everyone!
#12
Looks wonderfull thoe, very impressive.
03/08/2006 (6:13 am)
Im wondering why, if you intended this to be used in torque, you didn't use a .map or .diff editor to make this?Looks wonderfull thoe, very impressive.
#13
03/08/2006 (3:05 pm)
I'm not making .dif or .map because the editors are limited on what you can do with them and I really didn't want to spend the time learning to use an editor for this project. I'm not planning on this becoming something bigger or using this to learn an editor, it's really too complex a project for learning an editor. It's just an art project so I wanted to use the tools I know. Ideally, if I was going to make this for a game it would be a combination of .dif and .dts. The general overall shape could be .dif and the extra detailed objects could be .dts. That's still a possiblity and maybe it will end like that. 
Torque Owner Dracola