Sorry for this beginner like question, but I can't find the answer since I don't know what I need to type in Google search. I have been learning to use Blender for the past few months, and I tried to make some low poly wooden fence for small housesenter image description here

At first I created the standing part (I don't know what i should called it, sorry for my bad English)enter image description here

and then put a plank right through the middle partenter image description here

While it looks good, there are some parts of the plank that driven into the standing part, while it's not visible, I believe it's still rendered by the engine.enter image description here

Since it's only a small object and few in number, I believe in won't effect the performance. However, in the future, I was hoping to develop my own simple and low poly game. If I ever used this as an asset for my game, will it affect the performance altogether?

After some thoughts, I separated the plank into parts, and placed them between the standing parts. Since it's a simple type (cube shaped) fence, it doesn't take much time to connect the vertexes together, but what if the fence is round type? I have tried to make a hole in the standing parts, connecting the vertexes, and it's time consuming. enter image description hereenter image description here

My question is: considering the amount of time needed to work on, and how much it's affecting the render time process, would you rather drive a cylinder type mesh into the middle part, or doing the same thing as I did for the cube shaped fence?

  • $\begingroup$ I would look at the Array Modifier with the Merge Check box. The Array Modifier also has Start & End cap Mesh options. This way you create single post and let Blender do the work. $\endgroup$
    – rob
    Jan 18, 2019 at 9:49

2 Answers 2


When it comes to game assets there is one general simple rule. Use as many polygons as necessary and as little polygons as possible. No exception.

It's not only the amount of polygons but also the amount of vertices that matter. On top of that you also have to take the amount of UV islands for texturing (that means no "Smart UV Projection" or other fancy UV shortcut nonsense).

Also, don't assume the performance hit by the size of the object (in terms of screen space) in your finished game, neither the CPU nor the GPU care about that. They only care about the complexity as I described it above.

Continuous planks and bars (like in your fist example) through other meshes are the way to go, as long as they stay at the same position at their entry and exit points.

  • $\begingroup$ About Smart UV Projection, does the seam automatically created by Blender is really not efficient? I heard most people prefer to mark the seam manually. Since I'm still learning about basic of UV Unwrapping and how it works, I am trying not to rely on anything that automatically or instantly produced, so I have no experience in using Smart UV Projection $\endgroup$
    – Switch88
    Jan 18, 2019 at 17:35
  • $\begingroup$ Box Projection is OK, as long as it works for your model. In most cases it doesn't and manual unwrapping is still more effective in many cases. Smart UV Projection is everything else but efficient, because it generates way too many UV Islands. Every UV island generates a draw call, which means that both your CPU and GPU and most possibly your HDD have to work. There is a short but comprehensible explanation on Stackexchange . In any case, UV unwrapping is a basic and necessary skill in 3D modelling. $\endgroup$ Jan 19, 2019 at 5:39
  • Start with a plane and model a plank;
  • extrude it to give it thickness;
  • add a plane, scale it down, align wih your plank;
  • knife project the plane with the strike through option checked;
  • extrude "holes" horizontally;
  • use Array modifier on everything.

With cylinders you knife project, as above, and extrude circle(s).

Every ime you extrude, remember to remove doubles.

This method doesn't produce i.a. inner edges and, for the sake of performance, it's always better to operate on single meshes rather than (too) many separate ones.

  • 2
    $\begingroup$ They're not asking how to model it, they want to know which one is better for performance. $\endgroup$ Jan 18, 2019 at 14:16

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .