Complicated model has become unworkable. Two sided faces, OBJ Export, a million edges, oh my!



I’ve made a to-scale replica of now abandoned school in Highland Park (formerly known as Thompson Elementary). A non profit called Parker Village of Highland Park has purchased it and it working on converting it into an eco village.

Here’s a Google Earth 3d view and a photo of a part of the structure:

Here’s a photo of the model:

Here is a link to model in Google Drive

I think I might have gone into a little too much detail because anytime I try to make an edit the spinning rainbow wheel of death shows up and it takes seconds, sometimes almost a minute, to complete. I have already turned view profiles off, but this only provided temporary relieve.

I’d really like to be able to explode all my groups, but the spinning wheel of death shows up and basically seems to freeze everything up, even if I let it sit for hours.

My goal is to upload the model into SketchFab. The model seems to complicated for the exporter. Basically every time I try, I get an error code that’s produced when the processing takes too long (over 60 minutes).

I did have some luck exporting the model as an OBJ (& MTL - > zip) and uploading directly that way.

Here’s an example of what is produced:

As you can see, the ceiling seems to have just disappeared. So has all of the mortar behind the bricks. There’s just errors galore.

I spend a few hours trying to simplify the model (very slowly) and now when I export as OBJ & upload, I get this:

PVHP Thompson Elementary Test 2 by TysonGersh on Sketchfab

I suspect part of my problem is that I have a lot tiny reversed faces, as can be seen in pink in the image below:

I’ve already spent hours and hours manually correcting the reversed faces that I could, but with how slow this model operated, it would probably take another 20 hours just to finish properly reversing what’s left.

Does anyone have any recommendations for me? I spent WAY WAY WAY too many hours on this model to abandon it


Please put the .SKP file on a file server and post the link in this thread. Then the experts can offer suggestions as to how to put the model on a diet.


If I understand you correctly, I think I already did that. Do you see the hyperlinked sentence below from the OP?

Is that not a link to the model?


To avoid any confusion, here’s the URL to the publicly viewable/downloadable SketchUp file of the model that prompted the post:

I tried uploading to the 3D warehouse, but the file is too large:

I figured a Google Drive link was the next best thing.

Does that address your comment @RLGL?


The model is just short of 160MB
Modeled properly, it might be 1/10 that size.

Modeling individual bricks = Overmodeling
Use a texture image.


I just feel like the brick textures don’t look as good or realistic. That said, I do give them a shot first these days. I initially made this model before I learned that lesson. Is there anyway to save the masonry in this model?


The bricks are a Group, as are a few exterior details.
But the bulk of the building is a mass of interconnected raw geometry.

One can fix most any model, given unlimited time and patience.
I believe starting over would be best.


That is disappointing at best.


Yes, unfortunately you have modelled the bricks to the extreme, there is no need to have that level of detail unless you are viewing the model up close (couple of meters across (even then you would only detail that part rather than the entire model). Ive never had to model to the level of detail of bricks on such a large building, it’s always fine to use a material.

If you are having issues then I would just change your render engine, there are lots that can give a pretty high realism level.

Best thing is to go back and to replace those bricks with textures since they don’t serve a useful purpose only to slow the model down. But the model itself is generally ok except for the points @Geo has already raised.

#10 is a great place to look for brick textures.

The realism of the brick textures depends largely on making them blend together seamlessly. Notice how the mortar is different colors, so that makes it unrealistic.

But I think you’d spend less time, in the long run, perfecting your textures, than modeling every brick.


I think that the standard “brick antique 01” texture is pretty close to the building picture. It can be lightened/darkened of course. There are many others out there too.


I guess my concern is that in SketchFab the performance of the mortar needs to be different than the brick in the rendering. You have to treat the two the same if they are being used as a single material/texture


If you really want to go the extreme levels of material editing then you need to leave SketchUp and move to something else. Substance Painter should be the first stop (as it’s the easiest to pick up for the most part):

I still think you will be fine with what has been provided, can’t see why anybody is going to be inspecting the mortar between bricks in such a giant model?


Hang on a second. If you live close to that building and can get a good sized section of a wall, you could take a photo and then import that into SketchUp/Photoshop and create a texture to get an even better match. If you wanted to be pedantic.


Do you think there’s value in having a model that’s simply accurate? I can’t help but think that that’s the gold standard. No tricks, no illusions, just a literal digital iteration of what exists in the real world. Say I wanted to use my building model for material calculations, or building performance analysis. Having a real representation makes that possible, no?


Displacement maps are 2D textures that can make it look like you modeled every brick by creating detail in the render. Considering the high price of V-Ray, I’ll mention that Blender, a free program, has displacement maps. Although, it’s really hard to learn Blender…


I’m doing that on the house I’m buying (well half, it’s technically a two family flat, so I just have the basement and first floor), but I’m not convinced that the high resolution photo textures really perform as well. You can see what I’m talking about:

Smith t2 by TysonGersh on Sketchfab

I just think it’s noticeably inaccurate. It feels lazy/sloppy. A lot of my work is spent with communities that are very removed from technology and any “tell” that reminds them that what they are looking at isn’t real results in instant rejection. The design demands are very high in my neighborhood.


Yes indeed, but not in SketchUp, you would use something like Catia for that, not a poly modeler such as SketchUp. Mainly as the industry standard for vehicles is NURBS not Poly (and advance engineering).

SketchUp can’t do all the extra things you would need such as stress and material testing, load bearing and wind resistance etc, not even with Plugins. Autodesk can deliver most of these with it’s advanced Pro packages.

In the department I work in, we make complete vehicles in CAD, down to each screw. However this would never be done (technically can’t) in SketchUp. You can make Architectural models in SketchUp/AutoCad but you would not need to have these in the full 3D model for building either, they are fine in the 2D blueprints alone.

3D is primarily used for visualisation, so when something can be faked, it always is (as render time = money).

You also have separate 3D models for different things, and you would never really put them all in the same model. Exterior, interior, electrical, hydraulic…all of those CAD models would not be in a single file it would be far too dense. And if they are in the same file they will be in separate layers and only merged for post production such as infographics or animations.

However in those circumstances we generally transform the NURB geometry into Poly to be used in more visual friendly software such as Max/Maya/Modo/SketchUp.


Spent a few hours manually reversing faces. Progress made, but still needs work:


I think you will still end up with better results swapping all those bricks out, especially if you are just planning on using Sketchfab nobody can tell the difference.