 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Greetings all
I used Wings3d to make some objects for pov-ray. I
imported the objects without mishap. The objects had
extra information that I did not want, such as:
1) a normal vector for each vertex
2) a texture index for each triangle (this is kind of
annoying because I can't then apply a texture to
occurances of the mesh later on)
What I would like to do is strip off the normal &
texture information and also convert the mesh2 object to a
regular mesh object. Does anyone know of a program or a
.pov script that will do this?
Thanks (in advance).
Pete
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
No direct answers here, but some suggestions...
> 1) a normal vector for each vertex
Perhaps you can just delete the entire normal_vectors and normal_indices
blocks? If there are certain edges that you wanted to remain "hard", that
can be set in wings itself. I'm not sure exactly what you don't like about
having all of the normal vectors.
> 2) a texture index for each triangle (this is kind of
> annoying because I can't then apply a texture to
> occurances of the mesh later on)
I've also found this feature annoying. To fix it, I select the face_indices
block, and do a search-and-replace, replacing ">,0," with ">,". This removes
all of the texture indices, except maybe the very last one.
> What I would like to do is strip off the normal &
> texture information and also convert the mesh2 object to a
> regular mesh object.
Why do you prefer mesh over mesh2? They both contain the same data, but mesh
is (*much*) slower to parse.
- Slime
[ http://www.slimeland.com/ ]
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Slime <fak### [at] email address> wrote:
> Why do you prefer mesh over mesh2? They both contain the same data, but mesh
> is (*much*) slower to parse.
Have you actually measured the parsing speed difference?
I haven't and I would be interested in knowing some percentages...
--
#macro M(A,N,D,L)plane{-z,-9pigment{mandel L*9translate N color_map{[0rgb x]
[1rgb 9]}scale<D,D*3D>*1e3}rotate y*A*8}#end M(-3<1.206434.28623>70,7)M(
-1<.7438.1795>1,20)M(1<.77595.13699>30,20)M(3<.75923.07145>80,99)// - Warp -
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
in news:3f705ed6@news.povray.org Warp wrote:
> I haven't and I would be interested in knowing some percentages...
>
I have, but don't remeber the exact difference, but mesh2 is indeed
faster. But, there is one tricky thing, I also compared a completely non-
optimised mesh2 with a maximal optimised mesh2. With optimised I mean have
aan as small as possible amount of vertices. The more optimised the slower
the parsing and the more it approached the speed of mesh.
Ingo
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> Have you actually measured the parsing speed difference?
> I haven't and I would be interested in knowing some percentages...
Not with the same mesh. However, I commonly use mesh2's, and face parsing
speeds of maybe about 5-10 seconds (long enough to make me a tiny bit
annoyed just as the parsing finished). Recently, I had to use the original
mesh format for a model with a similar amount of detail (similar number of
triangles). I could go to the bathroom and return well before the parsing
had finished. I ended up making a low-res version of it for testing
purposes, which would take a reasonable length of time (5-10 seconds), using
the other version for only overnight renders.
In any case, the difference is large enough that it's worth caring about.
Sorry I can't give any solid numbers.
- Slime
[ http://www.slimeland.com/ ]
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Wasn't it Warp who wrote:
>Slime <fak### [at] emailaddress> wrote:
>> Why do you prefer mesh over mesh2? They both contain the same data, but mesh
>> is (*much*) slower to parse.
>
> Have you actually measured the parsing speed difference?
> I haven't and I would be interested in knowing some percentages...
Here's the results of a quick test that I ran. I took an OBJ file and ran it through a
convertor that can output both mesh and mesh2 formats.
The OBJ file is 16.7Mb, and has 99213 vertices and 170376 faces.
The Mesh2 parsed in 45% of the time and used 72% of the memory. The Mesh1 is shown as
having 86 frame level objects because there's a separate mesh1 object for each part
that
has a different texture.
Scene contains 86 frame level objects; 0 infinite.
Statistics for C:\My Documents\Poser\MyStuff\Spacegirl\AlienM1.pov, Resolution 640 x
480
----------------------------------------------------------------------------
Pixels: 307840 Samples: 363544 Smpls/Pxl: 1.18
Rays: 363544 Saved: 0 Max Level: 1/5
----------------------------------------------------------------------------
Ray->Shape Intersection Tests Succeeded Percentage
----------------------------------------------------------------------------
Mesh 387596 105567 27.24
Bounding Box 11443310 3850760 33.65
Light Buffer 1512203 649452 42.95
Vista Buffer 1346432 769933 57.18
----------------------------------------------------------------------------
Calls to Noise: 0 Calls to DNoise: 10
----------------------------------------------------------------------------
Shadow Ray Tests: 132628 Succeeded: 30319
----------------------------------------------------------------------------
Smallest Alloc: 25 bytes Largest: 1179672
Peak memory used: 46350651 bytes
----------------------------------------------------------------------------
Time For Parse: 0 hours 0 minutes 49.0 seconds (49 seconds)
Time For Trace: 0 hours 0 minutes 6.0 seconds (6 seconds)
Total Time: 0 hours 0 minutes 55.0 seconds (55 seconds)
POV-Ray finished
Scene contains 1 frame level objects; 0 infinite.
Statistics for C:\My Documents\Poser\MyStuff\Spacegirl\AlienM1.pov, Resolution 640 x
480
----------------------------------------------------------------------------
Pixels: 307840 Samples: 363912 Smpls/Pxl: 1.18
Rays: 363912 Saved: 0 Max Level: 1/5
----------------------------------------------------------------------------
Ray->Shape Intersection Tests Succeeded Percentage
----------------------------------------------------------------------------
Mesh 445143 85352 19.17
Bounding Box 14542692 4389902 30.19
----------------------------------------------------------------------------
Calls to Noise: 0 Calls to DNoise: 10
----------------------------------------------------------------------------
Shadow Ray Tests: 81231 Succeeded: 31723
----------------------------------------------------------------------------
Smallest Alloc: 25 bytes Largest: 12267096
Peak memory used: 33609623 bytes
----------------------------------------------------------------------------
Time For Parse: 0 hours 0 minutes 22.0 seconds (22 seconds)
Time For Trace: 0 hours 0 minutes 6.0 seconds (6 seconds)
Total Time: 0 hours 0 minutes 28.0 seconds (28 seconds)
POV-Ray finished
--
Mike Williams
Gentleman of Leisure
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
In article <Xns### [at] netplexaussieorg>,
ingo <ing### [at] tagpovrayorg> wrote:
> I have, but don't remeber the exact difference, but mesh2 is indeed
> faster.
The exact difference will depend on the mesh. The more vertices shared
among multiple triangles, the better.
> But, there is one tricky thing, I also compared a completely non-
> optimised mesh2 with a maximal optimised mesh2. With optimised I mean have
> aan as small as possible amount of vertices. The more optimised the slower
> the parsing and the more it approached the speed of mesh.
I can't think of anything that would cause this. With the same number of
triangles, fewer vertices should be better...duplicates waste space and
need more computations to remove.
--
Christopher James Huff <cja### [at] earthlinknet>
http://home.earthlink.net/~cjameshuff/
POV-Ray TAG: chr### [at] tagpovrayorg
http://tag.povray.org/
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
in news:cja### [at] netplexaussieorg Christopher
James Huff wrote:
> I can't think of anything that would cause this.
What if POV-Ray internally expands the mesh2? If it does, that would take
some parsing time.
Ingo
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
ingo <ing### [at] tagpovrayorg> wrote:
> What if POV-Ray internally expands the mesh2? If it does, that would take
> some parsing time.
Expands to what?
--
#macro M(A,N,D,L)plane{-z,-9pigment{mandel L*9translate N color_map{[0rgb x]
[1rgb 9]}scale<D,D*3D>*1e3}rotate y*A*8}#end M(-3<1.206434.28623>70,7)M(
-1<.7438.1795>1,20)M(1<.77595.13699>30,20)M(3<.75923.07145>80,99)// - Warp -
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On Mon, 22 Sep 2003 23:58:21 -0400, "Slime" <fak### [at] emailaddress>
wrote:
>No direct answers here, but some suggestions...
>
>> 1) a normal vector for each vertex
>
>Perhaps you can just delete the entire normal_vectors and normal_indices
>blocks? If there are certain edges that you wanted to remain "hard", that
>can be set in wings itself. I'm not sure exactly what you don't like about
>having all of the normal vectors.
>
I make "greebly"[1] angular linear shapes. The normals make
everything kind of, umm, plush and rounded ... sort of
melted. It looks really bad.
>> 2) a texture index for each triangle (this is kind of
>> annoying because I can't then apply a texture to
>> occurances of the mesh later on)
>
>I've also found this feature annoying. To fix it, I select the face_indices
>block, and do a search-and-replace, replacing ">,0," with ">,". This removes
>all of the texture indices, except maybe the very last one.
>
This is a very good idea. Thank you!
>> What I would like to do is strip off the normal &
>> texture information and also convert the mesh2 object to a
>> regular mesh object.
>
>Why do you prefer mesh over mesh2? They both contain the same data, but mesh
>is (*much*) slower to parse.
>
The mesh object is much better for hand text editing.
It's hard for me to keep track of points while
editing a mesh2 object.
> - Slime
> [ http://www.slimeland.com/ ]
>
>
--
[1] Example of "greebles": hull plating of the
Millenium Falcon
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
