In your opinion, what are the "best" poly count numbers to aim for on your models?
polycounter lvl 5
This is a super general question, and I know there's no real true answer to this, but I think it's important to keep asking this from time to time. Google gives me very outdated information if I try to do research from there, and with any luck we might be able to help people like myself with this thread for a while.
It's 2017. Technology advances and gaming machines become capable of performing better and running better-looking visuals. But at the same time, especially if one is developing for things besides pre-built consoles, you need to remember that not everyone in your audience is going to be running around with cutting edge technology and it can be important to scale down sometimes just for accessibility's sake.
So, I often find myself running into this problem while working on a personal project of mine. I haven't imported anything to an engine yet so I don't know how these things actually perform in real-time yet. The best I can do at this state is to just sort of feel out when it looks like I have too many faces on an object, and understanding when I can use things like normal maps to my advantage. For example, I'm currently working on a prominent large partner to the main character, and while it's not close to finished yet, it's currently coming in at around 40,000 tricount. That sounds like a lot to me...but at the same time I have no idea how high that number actually is for what I'm working on. It can be very confusing. To make things worse, if I do decide to run it in an in-game engine, and let's say it lags the gameplay too much, I have to go back, unskin it, remodel it, reskin it, and repaint skin weights, and reanimate it, and that can take a very long time. Unless I'm missing something here.
And of course, there are also loads of factors that need to be considered when thinking about your ideal poly count for something in your game. Is the game a first-person shooter featuring guns? Usable guns should have a lot more polygons than non-usable guns since it's going to be very prominent. Are you okay with running the game at a locked 30 FPS? This will free you up to make things much more detailed than you might have been able to otherwise. There's also things like dynamic effects, physics, reflections, how many things are on the screen at once...I could go on and on but I probably shouldn't since I don't know too much about CPU/GPU management and bottlecapping.
Anyways, I'd like to hear your thoughts on this. How many tris should you dedicate to building (for modern game engines):
Characters
Props
Environments
Both detailed and not-as-detailed, since many things don't need to be seen for long, or up-close. Anything else would be good to know as well, thanks for reading.
It's 2017. Technology advances and gaming machines become capable of performing better and running better-looking visuals. But at the same time, especially if one is developing for things besides pre-built consoles, you need to remember that not everyone in your audience is going to be running around with cutting edge technology and it can be important to scale down sometimes just for accessibility's sake.
So, I often find myself running into this problem while working on a personal project of mine. I haven't imported anything to an engine yet so I don't know how these things actually perform in real-time yet. The best I can do at this state is to just sort of feel out when it looks like I have too many faces on an object, and understanding when I can use things like normal maps to my advantage. For example, I'm currently working on a prominent large partner to the main character, and while it's not close to finished yet, it's currently coming in at around 40,000 tricount. That sounds like a lot to me...but at the same time I have no idea how high that number actually is for what I'm working on. It can be very confusing. To make things worse, if I do decide to run it in an in-game engine, and let's say it lags the gameplay too much, I have to go back, unskin it, remodel it, reskin it, and repaint skin weights, and reanimate it, and that can take a very long time. Unless I'm missing something here.
And of course, there are also loads of factors that need to be considered when thinking about your ideal poly count for something in your game. Is the game a first-person shooter featuring guns? Usable guns should have a lot more polygons than non-usable guns since it's going to be very prominent. Are you okay with running the game at a locked 30 FPS? This will free you up to make things much more detailed than you might have been able to otherwise. There's also things like dynamic effects, physics, reflections, how many things are on the screen at once...I could go on and on but I probably shouldn't since I don't know too much about CPU/GPU management and bottlecapping.
Anyways, I'd like to hear your thoughts on this. How many tris should you dedicate to building (for modern game engines):
Characters
Props
Environments
Both detailed and not-as-detailed, since many things don't need to be seen for long, or up-close. Anything else would be good to know as well, thanks for reading.
Replies
but make sure you dont waste polys... dont subdivide something only cause its easy... every polygon should ad something to the silhouette or do something for animation or shape... dont create useless poly...
concentrate to make good art... if you can load it into mormoset all is fine... every pipeline does have special cases you will learn if you are in production...
With that said, relative to many other aspects of gameart, coaching someone to be more efficient with their poly budget is pretty low hanging fruit.