| View previous topic :: View next topic |
| Author |
Message |
Adar
DeleD PRO user
Joined: 29 Aug 2007
Posts: 233
|
 Posted: Sun Feb 03, 2008 4:38 pm Post subject: Deled is slow when using the Union tool..? |
 |
|
Trying to use the union tool on not so low poly models takes forever !! sometimes I have to wait for 5-7 minutes..
I have a lot of geometry in my scene but there isn't 1 object which contains lots of geometry.. I tried hiding parts, splitting them and what not.. it takes forever !!  |
|
| Back to top |
|
 |
Jeroen
Site Admin
Joined: 07 Aug 2004
Posts: 5332
Location: The Netherlands
|
| Posted: Sun Feb 03, 2008 6:07 pm Post subject: |
|
|
The Constructive Solid Geometry tools build socalled Binary Space Partition trees first. This process can take a lot of time on complex models. For reference; Quake levels are also precompiled into BSP trees and compiling a single level can sometimes take hours. It's the nature of the technology, really.
Just out of curiousity, how many polygons per model are we talking here? |
|
| Back to top |
|
|
Adar
DeleD PRO user
Joined: 29 Aug 2007
Posts: 233
|
| Posted: Sun Feb 03, 2008 8:08 pm Post subject: |
|
|
| Jeroen wrote: |
The Constructive Solid Geometry tools build socalled Binary Space Partition trees first. This process can take a lot of time on complex models. For reference; Quake levels are also precompiled into BSP trees and compiling a single level can sometimes take hours. It's the nature of the technology, really.
Just out of curiousity, how many polygons per model are we talking here? |
Uniting an object with 11 polygons to an object with 2400 polygons takes almost 10 minutes of my life.. is that reasonable? 
I need to add like 10 of those  |
|
| Back to top |
|
|
Starnick
DeleD PRO user
Joined: 28 Jul 2007
Posts: 611
|
| Posted: Mon Feb 04, 2008 1:48 am Post subject: |
|
|
I don't use the CSG tools a whole lot let alone on complex objects so I wouldn't know if that was abnormal or not...but I just threw a cube in arbitrarily that intersected the middle part of a 1500 poly plaza area, and it took about 2 and a half minutes (calculating the partition trees was relatively fast). I then tried it again, this time pretty much subdividing the entire complex with a rectangle which intersected a lot of interesting geometry (ie stairs, railing, etc). That took just about 3 and a half minutes to complete.
Not sure if that's comparable of course  |
|
| Back to top |
|
|
jwatte
DeleD PRO user
Joined: 26 Apr 2006
Posts: 513
|
| Posted: Mon Feb 04, 2008 7:56 pm Post subject: |
|
|
| It shouldn't need to subdivide the entire mesh -- it only needs to extract the parts of the mesh(es) whose bounding boxes overlap. Perhaps there's some optimizations to be had there? |
|
| Back to top |
|
|
Adar
DeleD PRO user
Joined: 29 Aug 2007
Posts: 233
|
| Posted: Mon Feb 04, 2008 8:47 pm Post subject: |
|
|
Well I wouldn't know either 
It's slow as hell even when optimizing, anyway most of the time I can't use optimize on complex objects cause polygons tend to dissappear when I do that. |
|
| Back to top |
|
|
Jeroen
Site Admin
Joined: 07 Aug 2004
Posts: 5332
Location: The Netherlands
|
| Posted: Mon Feb 04, 2008 9:06 pm Post subject: |
|
|
| jwatte wrote: |
| It shouldn't need to subdivide the entire mesh -- it only needs to extract the parts of the mesh(es) whose bounding boxes overlap. Perhaps there's some optimizations to be had there? |
Yes, I definately think the CSG routines can be optimized, especially in situations when complex objects are being manipulated. I thought about extracting only the affected parts of the model myself and I will experiment with that a bit if time allows.
@Adar: I would advise to use the Optimize tool often. Even on less complex objects. If you optimize an object early, it'll produce better results when you optimize again sometime later in the modeling process. If you wait until the object is already very complex (and unoptimized), the Optimize algorithm might wrongfully assume it is ok to remove parts of the geometry while in fact, it isn't.
I also think I will be working on optimizing parts of a model, and not just the whole model at once. |
|
| Back to top |
|
|
Adar
DeleD PRO user
Joined: 29 Aug 2007
Posts: 233
|
| Posted: Mon Feb 04, 2008 10:03 pm Post subject: |
|
|
| Quote: |
| Yes, I definately think the CSG routines can be optimized, especially in situations when complex objects are being manipulated. I thought about extracting only the affected parts of the model myself and I will experiment with that a bit if time allows. |
How about a progress bar too? so we'll now how much time to wait 
| Quote: |
@Adar: I would advise to use the Optimize tool often. Even on less complex objects. If you optimize an object early, it'll produce better results when you optimize again sometime later in the modeling process. If you wait until the object is already very complex (and unoptimized), the Optimize algorithm might wrongfully assume it is ok to remove parts of the geometry while in fact, it isn't.
I also think I will be working on optimizing parts of a model, and not just the whole model at once. |
Actually it's enough that I unite 2 objects together and then optimize for the optimization to go crazy.. even if both objects are optimized at first and fairly simple.. |
|
| Back to top |
|
|
Jeroen
Site Admin
Joined: 07 Aug 2004
Posts: 5332
Location: The Netherlands
|
| Posted: Tue Feb 05, 2008 7:06 am Post subject: |
|
|
| Adar wrote: |
Actually it's enough that I unite 2 objects together and then optimize for the optimization to go crazy.. even if both objects are optimized at first and fairly simple.. |
I very much would like to see some examples of this. If I use CSG on simple objects and then optimize them, everything seems to work as intended. Please provide some examples (screenshots or .dxs files) so I can use them to improve the routines. |
|
| Back to top |
|
|
Adar
DeleD PRO user
Joined: 29 Aug 2007
Posts: 233
|
| Posted: Tue Feb 05, 2008 6:36 pm Post subject: |
|
|
| Jeroen wrote: |
| Adar wrote: |
Actually it's enough that I unite 2 objects together and then optimize for the optimization to go crazy.. even if both objects are optimized at first and fairly simple.. |
I very much would like to see some examples of this. If I use CSG on simple objects and then optimize them, everything seems to work as intended. Please provide some examples (screenshots or .dxs files) so I can use them to improve the routines. |
I've e-mailed you a DXS with the problem.
That's just 1 sample, I'm encountering many issues with the optimize tool. |
|
| Back to top |
|
|
Jeroen
Site Admin
Joined: 07 Aug 2004
Posts: 5332
Location: The Netherlands
|
| Posted: Wed Feb 06, 2008 7:40 pm Post subject: |
|
|
| Adar wrote: |
I've e-mailed you a DXS with the problem.
That's just 1 sample, I'm encountering many issues with the optimize tool. |
I haven't received the DXS yet. Can you resend it? |
|
| Back to top |
|
|
Adar
DeleD PRO user
Joined: 29 Aug 2007
Posts: 233
|
| Posted: Wed Feb 06, 2008 9:31 pm Post subject: |
|
|
| Jeroen wrote: |
| Adar wrote: |
I've e-mailed you a DXS with the problem.
That's just 1 sample, I'm encountering many issues with the optimize tool. |
I haven't received the DXS yet. Can you resend it? |
Done  |
|
| Back to top |
|
|
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
