View previous topic :: View next topic |
Author |
Message |
CMe Member
Joined: 30 Jun 2005 Posts: 72 Location: Ontario Canada
|
Posted: Thu Jan 26, 2006 11:35 pm Post subject: curved surfaces |
|
|
Hi, I thought I would give a preview to the curved surfaces plugin I'm currently working on. By creating a crude shaped model it can be subdivided into a smooth curve. Here's a screenshot. The bottom object is the original rough representation, the top one is automatically generated.
Last edited by CMe on Fri Jan 27, 2006 4:10 am; edited 1 time in total |
|
Back to top |
|
|
trucker2000 DeleD PRO user
Joined: 11 May 2005 Posts: 1839 Location: Sacramento, California
|
Posted: Fri Jan 27, 2006 12:01 am Post subject: |
|
|
Pretty cool idea. _________________ Some day I will grow up and be a real modeler.
"Never give up! Never surrender!!"
Sys specs:
asus
8 gigs ram
Invidia gtx560 video card
Windows 8 (worst op sys Ever) |
|
Back to top |
|
|
Jeroen Site Admin
Joined: 07 Aug 2004 Posts: 5332 Location: The Netherlands
|
Posted: Fri Jan 27, 2006 6:38 am Post subject: |
|
|
Excellent stuff! |
|
Back to top |
|
|
Daaark DeleD PRO user
Joined: 01 Sep 2004 Posts: 2696 Location: Ottawa, Canada
|
Posted: Fri Jan 27, 2006 2:14 pm Post subject: |
|
|
Wow.. That's a lot of polygons in the first example? Was that done with a setting to show off how good it is, or do you really make that many polygons all the time?
I could see using something like this to create some nice static curtains. |
|
Back to top |
|
|
CMe Member
Joined: 30 Jun 2005 Posts: 72 Location: Ontario Canada
|
Posted: Fri Jan 27, 2006 8:43 pm Post subject: |
|
|
Depending on how much you refine the model the more polygons there will be. The first image was done with a subdivision setting of 3, and has 1089 vertices (2048 polygons. The bottom image was done with 2 subdivisions and has 153 vertices (256 polygons). The range for the subdivision is only limited by your computer, so no thats not as good as it can get. I've sent a model with 100000 vertices to deled with no problems. There does come a point though where refining wont make a viewable difference. 2 is usually enough.
To give an idea of whats possible, this is the method used in quake 3 to make all those arches and organic material in the levels. I developed the code base for this some time ago for a quake 3 clone I was writing, so algorithm wise it outputs the exact same as q3 does.
It's also an intuitive way to make terrain as you can start with a large grid of points and just pull some of them up to get the effect of hills. |
|
Back to top |
|
|
Twixn DeleD PRO user
Joined: 01 Dec 2005 Posts: 136 Location: Melbourne Australia
|
Posted: Fri Jan 27, 2006 10:46 pm Post subject: |
|
|
Thats cool.
Yeah, it would be usefull for making terrain.
Does it work well for model smoothing aswell?
-Twixn- _________________ Failure is not an option!....It comes bundled with the software. |
|
Back to top |
|
|
CMe Member
Joined: 30 Jun 2005 Posts: 72 Location: Ontario Canada
|
Posted: Sat Jan 28, 2006 12:08 am Post subject: |
|
|
No, it's not designed to work with any model type. It must be a grid as it only uses the vertices of the grid for control points. There is also the restriction that a subdivided grid must be even numbers. ie 2x2 polygons or 2x4 polygons, 4x4 etc... |
|
Back to top |
|
|
Twixn DeleD PRO user
Joined: 01 Dec 2005 Posts: 136 Location: Melbourne Australia
|
Posted: Sat Jan 28, 2006 1:16 am Post subject: |
|
|
Well, its still pretty cool
It might be very usefull with the landscape plugin.
-Twixn- _________________ Failure is not an option!....It comes bundled with the software. |
|
Back to top |
|
|
Jeroen Site Admin
Joined: 07 Aug 2004 Posts: 5332 Location: The Netherlands
|
Posted: Sat Jan 28, 2006 7:49 am Post subject: |
|
|
What would be a good method to achieve model smoothing? |
|
Back to top |
|
|
CMe Member
Joined: 30 Jun 2005 Posts: 72 Location: Ontario Canada
|
Posted: Sat Jan 28, 2006 9:23 am Post subject: |
|
|
Well visually an object will look smooth when rendered if it has vertex normals and is rendered with gouraud or phong shading for example. As for geometry, thats where subdivision and normalization comes in. It has an interesting effect when working with primitives. If you take a cube, subdivide it a few times and normalize the points you end up with a tesselated sphere. There's different methods for smoothing like Laplacian smoothing, Loop Subdivision, Butterfly Subdivision etc.. They just give different approximations. Loop seems to produce the smoothest, but butterfly preserves the shape better afaik |
|
Back to top |
|
|
Jeroen Site Admin
Joined: 07 Aug 2004 Posts: 5332 Location: The Netherlands
|
Posted: Sat Jan 28, 2006 10:08 am Post subject: |
|
|
Thanks for the tips! I'll try to determine which method would suit DeleD best as I definately want smoothing implemented. Note: 1.4 already has a "poor-man's-subdivision" implemented, being Vertex Bevel. |
|
Back to top |
|
|
CMe Member
Joined: 30 Jun 2005 Posts: 72 Location: Ontario Canada
|
Posted: Sat Jan 28, 2006 10:43 am Post subject: |
|
|
This has actually sparked my interest. I'm going to write a prototype for the loop method and maybe butterfly once I'm done with this curved surface code. |
|
Back to top |
|
|
Jeroen Site Admin
Joined: 07 Aug 2004 Posts: 5332 Location: The Netherlands
|
Posted: Sat Jan 28, 2006 10:56 am Post subject: |
|
|
Great, keep us/me posted, CMe! Do keep in mind that I will be giving it a shot myself later this year. One of our goals for DeleD this year is to have hi-resolution modeling (hence subdivision/smoothing/etc) implemented in the DeleD core (not via plugins) this year. I for one would be very grateful if you would assist me when the time has come. |
|
Back to top |
|
|
CMe Member
Joined: 30 Jun 2005 Posts: 72 Location: Ontario Canada
|
Posted: Sat Jan 28, 2006 11:54 am Post subject: |
|
|
Oh of course. I'd be happy to help. |
|
Back to top |
|
|
CMe Member
Joined: 30 Jun 2005 Posts: 72 Location: Ontario Canada
|
Posted: Sat Jan 28, 2006 9:54 pm Post subject: |
|
|
The plugin is done. It's available on my website. |
|
Back to top |
|
|
|