How detailed/when to use subsurface modifier

This is kind of a broad question here, but I am really getting into modeling and want to make sure I am doing it the right way. Right now I am just modeling random objects. I just got done with a Tide detergent bottle.

Lately, I’ve been really concentrating on making sure my meshes are nice and neat. What I did for the Tide bottle was to start off with very basic, blocky polygons and build from there a very blocky shaped bottle. I then applied a subsurface modifier. That way, I could continue to easily model the small number of blocky polygons, while checking constantly checking how the smoothed out version would look.

In the end, it looks great. However, with the subsurface modifier turned off, like I said, it looks quite blocky. If I turn the subsurface modifier back on, but change it from level 1 to level 3, then the mesh looks closer to the finished, high resolution version. Of course there are a lot more polygons then.

I guess my question is, if I ever got to the point where I was doing work for someone else or trying to sell models, is my way ok to do? Is it of any concern that the model definitely needs the subsurface modifier to be applied, otherwise it won’t look like much? I’m just wondering if this is the norm, or if it’s better to try and use detail from the get go and model it without relying on the modifier. I guess my thinking was to keep it simple on the bottom, so that there would be easy manipulation, and then smooth it out for the finished renders.


First, to prevent confusion, adding subdivision surface modifier = putting one in the modifier stack, applying a modifier = pressing the apply button on the modifier.

It depends on how much geometry is needed for the required look. Subdivision surface modifier is used along with smooth shading, and in many cases you don’t need high subdivision levels because shading takes out the blockiness.
When you don’t apply the modifier until you absolutely have to (final result is exported to another format for example) you can adjust the levels and can optimize your polygon count in the scene. Objects closer to camera have higher polygon count.
If you have to apply the modifier and perhaps are selling your model, you might want to do a poly reduction sweap to make it more optimized.

With practice, you’ll learn to take the modifier into account when modeling. Even when you haven’t put down a single vertex, you will know how many vertices are needed for different contours in your reference so that it works with particular subdivision level.
Some people complain that they sometimes see real life objects as a topology problem, keeping it so that they don’t have to use higher than level 2 subdivision. Could be annoying.

That is exactly what the subsurface modifier is for. The extra geometry you would have to create to manually achieve the same smoothed out look would be unwieldy and a waste of computer resources. In most cases you do want to make a slightly simpler mesh and use a subsurf to de-blocky-fy it. You’re doing it right.

Also, to clear up a point of terminology, you’re not “applying” the modifier unless you actually click that button that says “apply.” If you do that with a subsurf modifier, it will add all that extra geometry for real, which you rarely really want to do.

Edit: Curse you JA12, you typed faster than me.