View Full Version : new model, problems with windows

06-24-2004, 12:22 PM
I am a novice at modeling, so i am having some problems getting around my window needs for a new model of the USS Enterprise.

First, a bit about my model:
I am using a combination of blueprint images for reference, and i switch between them as needed, so if my model in the images doesn't match up exactly with the BG images, that is why- because parts were made from other references.

Ok- so i am trying to place windows with nice ledges on them, very slightly inset, upon the upper decks area. There will be more all over but considering this is the worst, most rounded area for windows i decided to start here!

I tried a few different methods and found that the solid drill works great, but i get pinched, strange distortions in my mesh afterwards.

please take a look at the supplied images to see what i mean-

i am not sure if this is what some people call 'open GL errors' that can simply be ignored (i doubt it) or if there is something worse taking place. i cannot find out how to get around it. at first when it happened i realized i had bad joints where the new polys were created, and so i merged the stray, redundant points which helped but not 100% despite getting all of them.

anyone know how i'd go about this?


image 1- just a view of the work so far on the primary hull- not bad for a novice, huh?
the 3 round 'windows' in front are, in fact, not windows so i am not talking about these- they were made with the solid drill tool and worked fine. they don't need to be inset. the insetting of the windows is where most of the problems arise, although some slight pinching appears after the use of the solid drill for the actual windows parts.

more images follow---

06-24-2004, 12:26 PM
Ok, so this is my upper decks part, it was made subpatch to get a nice rounded part. the red arrow points to the windows

06-24-2004, 12:28 PM
ok- still with me!?

this is the same part after it has been set to freeze so it is a poly object now- cannot use a boolean or drill tool on a subpatch model. if i am wrong, show me how!

the level of detail is set high for a nice clean round edge on everything, and i thought at one point that perhaps if i lowered it and then did the solid drill operation that this might reduce the problem but it didn't.

06-24-2004, 12:31 PM
ok, so a tip to those using the post- if you forget to add your image and post it, you cannot add it when you go into edit mode! then you cannot delete the post to hide this error and have to start over!

here is the image for #3

06-24-2004, 12:32 PM
with the box for the solid drill in the BG, we now go ahead...

06-24-2004, 12:33 PM
wow! a nice little window appears!

or is it all that nice?---

06-24-2004, 12:36 PM
well, after making it slightly inset with a smooth shift operation, the puckering appears.

actually this isn't accurate- the puckering is there but very slight before the smooth shift operation- doing the smooth shift just makes it worse.
even if i comprimised and left out the inset windows, the puckering is still there.

i have noticed the puckering is concentrated at the edge where the additional polys created by the solid drill tool came very close to other verticies- is this problem related tho that happening? if so, if i were able to re-arrange my mesh so it's verticies were positioned as not to interfere with the window placement in this manner, would it make a difference?


alright, so how! how do i clean this up, or avoid it?

can i cut clean windows into a subpatch object or no?
is there a way to get this back to being smooth again?

i appreciate any advice on this, kinda going nuts with it for 3 days now

06-24-2004, 01:29 PM
Have u tried to bevel instead of smooth shift?

06-24-2004, 01:49 PM
yes i have.
the problem is that it requires that i blow away the new polys, and make a new one so the bevel tool works the way you'd want it to, instead of it creating a bevel in all the smaller sub portions of the window polys.

so, creating a new poly might seem the way to go but it still screws up- observe:

still puckering once the solid drill is applied.

less problematic with the bevel, but still there.

06-24-2004, 01:51 PM
Maybe I am not understanding what you are trying to do, is it this?

06-24-2004, 02:04 PM
Here is another image

06-24-2004, 02:12 PM
ok, looks like you know what i am up to but your display is wireframe shade.

set it to smooth and you should get puckering- this is a good representation of how a finished render would look, would it not?

or am i imagining things?

i have 2 new images to illustrate:
in this first one, i did a solid drill, then merged stray points, and deleted the polys, and made a solid new ploy.
look at how the normals are- 2 look right, 2 do not, they point in an incorrect direction to the rest of the mesh-
now look at the next image--

06-24-2004, 02:15 PM
the non- puckering area corresponds to the new windows with the normals correct.

the puckering area corresponds to the area with the window normals being incorrect.

so how can i correct this??

06-24-2004, 02:17 PM
The images u are seeing are textured wire frame, and put into texture, it looks the same, as does for smooth shading.

Here is what I did.

I created my window
used Solid Drill>Stencil

Then I took Smooth Shift and Right Clicked only (also tried the 0.0 smooth shift plug I downloaded from flay)

Then I Moved the polygons backward

then I used the stretch tool and shrunk the polygons.

06-24-2004, 02:19 PM
smooth shade

06-24-2004, 02:26 PM
I know what you are talking about now!!!

Hehe, cut the inner edges of the window out and past them. The renderer sees them as a surface and tries to smooth them out causing the false look of the surface. If you cut and repast them, that pucker will go away.

06-24-2004, 02:30 PM
These are the inner edges I am talking about:

06-24-2004, 02:51 PM
yes, those puckers are smoothing errors. cut/paste the windows/sills. that should correct them.

06-24-2004, 03:13 PM
you are my new best friend :D

06-24-2004, 04:24 PM
Originally posted by pdrake
yes, those puckers are smoothing errors. cut/paste the windows/sills. that should correct them.

FYI, if you ever merge points, your problem is going to come back. I hate smooting errors.

The polys with the weird normal actually are a curved surface, right? I wouldn't even be surprised if they are non-planar.

06-24-2004, 04:28 PM
right and right.

i tried the merge points deal after the cut/paste and yes, the problem returns.
but, leaving them seperate works.

and yes, non planar polys all over that object.

i don't know what this means, what is it?
is it going to be a problem?
is there a way to correct it?

06-24-2004, 05:44 PM
the smoothing angle is based on adjoining polys. that means polys that share the same vertices. cut and paste seperates them and makes them different objects. as long as you aren't doing any deforming there will never be a problem with them being seperate. any deformation will show a tear. when you merge them, they once again share the same vertices so the software tries to smooth across them.

06-25-2004, 07:11 AM
I would work on the geometry and try to figure out a way to make it planar.

If you shift it in, cut the window, paste into an empty layer, kill the poly, then use the points as a basis for a flat polygon, that might help.

06-25-2004, 07:56 AM
For Non Planar Poly's, Triple them a triangle can never be non planar.

Also, another thing you might want to try is adding creating your window as a subpatch and use subpatch weights to "stiffen" the window edges.

06-25-2004, 09:56 AM
Try doing a small Supershift first, using only inset, not shift, and then smoothshifting the window in. This will give you a rim of polys which are planar with the surrounding non-window polys and this will act as a buffer for the rounding calculation.

06-25-2004, 10:40 AM
this is all getting interesting but the question is still un-answered:
what is a non-planar polygon, and is it a problem?

06-25-2004, 10:47 AM
a non-planer poly is just that. non-planer. a plane is a perfectly flat specified area. imagine a piece of glass with four sides. that's planer. now what happens if you try to move one corner out of the plane? the glass breaks. this is what happens to non-planer polys. they break. if you triple non-planers they become polys with 3 points. if you try to lift any corner of a piece of triangular glass it won't break, just pivot up on the one edge.

hope this makes sense.

06-26-2004, 04:02 PM
i think i understand that now, thanks.