Low poly or High Poly?

[Tunrax]

This is more than a question rather than a “low poly vs High poly” opinion.

Since the start of the course I always heard about “Low Poly means more performance in games and so on”, and that really rolled over my head to the point that I always check the “tris” count of my model every time just to practice model optimization.

A few years ago just to put an example happened the drama about “Yandere Simulator” drama that I don’t want to discuss but to make an example, his game had a lot of performance issues and people found that had “Unnecessary” High poly models in the game like a tooth brush with a lot of geometry in it so I always thought of that every time I do decisions in my models at the point that when we used “subdivision surface” on the bowling ball I kinda said “This is all wrong, how a bowling ball is getting like 9500 tris?”.

Do I am getting too paranoid on the optimization idea? should I chill out with that?

Mostly of my models gonna be buildings that gonna be shown on a website for an architecture business that probably the Geometry count won’t be much of an issue on there, but Someday if I start a career on video game development maybe I would love to solve this issue I have on mind about performance caused by models.

[FedPete]

Hi, you could use the search option. This question is discussed many times.

Basically, the number of vertices for an object depends on its usage.
You can put billions of vertices in an object, to enhance the details.
But if the object is far, far away as in a tiny dot. It has no real use to have billions of vertices.

But, if you are going to use it for close-ups, then it could be wise to have more details (vertices).
More vertices, slower render times (important for games).
What we do is extract information from the high detailed poly model, onto a less detailed low poly model. Using bitmap baking techniques. This is explained in the Orc course.

But it all depends on what are you going to do with your models! If you haven’t a plan, it is difficult to say where to spend your time.

A good approach would be to work modularly. So when you have a simple front door. You can replace the model with a front door in more detail. And even super high details for close-ups. This is also how game engineers work. They create different levels of details for each model. Far away, close by, and for example, a facial expression. It’s not one model but multiple. And all based on the model of high detail.

How high, depends, on what you are planning to do with the model.

As a Blender starter, I would advise not to think much about this. Learn the basics first, then learn how to optimize your work process.

[Gordon]

@Tunrax

Do I am getting too paranoid on the optimization idea? should I chill out with that?

Yeah, I think you should chill out with that. As @FedPete said, if you’re going to have close ups of stuff, you might want to have more detail and if it’s far away then it doesn’t matter, people won’t see the lack of details anyways. It all depends on what you want to do. I suppose later if you make assets for games and your game is running much slower than you want it to, you’ll need to think, hmmm is there something I can do to make the game a bit faster, and that’s when those details come into consideration. At least that’s what I think.

I’m not super experienced with Blender or anything, but I’ve always had the same concerns with tri counts and topology, but in the learning stage it really doesn’t matter. All that matters is what people can see from your renders. Just keep on learning. But I really think it’s good to be curious about these things. It encourages you to explore all of the resources out there, and trust me, there are loooaads. Keep on doing what you’re doing mate.

[NP5]

The course is rather obsessed with low poly, in part it is for simpler things to show students when learning, and also so many students are game centric. It’s in the name GameDev!

However, Blender is for a lot more than just games. Renders of single images are not much hampered by high polycounts. Quality matters to non game images usually. High poly high detail real geometry is required for 3D printing. No use baked textures there!

So Blender is a magnificent tool and can be used for many different end goals, which have different things that are important to them.

I am thinking of trying to start a new moto! If your scene has not got a million verts you are not trying!

[Gordon]

hahaha I like that moto

[Tunrax]

Thanks for all the feedback! I knew I was being paranoid with thiss stuff of optimization and things but I’ll think with the stuff I’m gonna make at first is not gonna be necessary to matter about extra geometry.

For now while learning I’m gonna Chill Out with this thing about low or high poly and just make the pieces low poly as heck.

[Digitz]

It really does depend on your application of what you are making.

For game assets, you’ll really want to make a high and low poly model. The high poly for getting all the small details and high quality textures, then the low poly for making it more practical for a game engine.

If you’re just rendering an image and want that image to be high quality then your close objects are going to be high poly and your distant objects can be low poly objects. Like trees in the distance can just be planes with a tree image on it as an example.

If you’re going for movie/film vfx, that will have to be high quality and high poly.

If you’re going for a stylised look then low poly is often synonymous with stylised.
In the end it really depends on what you are going for as the end result.

Don’t worry about it while you’re learning, you’ll understand more as you increase your knowledge of the subject you are studying. In this case it is 3D art and design I suppose.