$12256 / $11500
Hi all,
I wonder, if there is a good 'rule-of-a-thumb' for amount of triangles for models used in indie games? I understand that the answer depends on many different parameters, but still...
When I start modelling art for a game I approximately know:
approximate amount of active objects (characters) on the map
approximate level design idea (from large scale (mountains) to small passive objects (small rocks, furniture))
How can I decide which detail level to give to each of these? Where do I start?
I.e. I don't want to end up remaking characters models because they are too high-detailed and the game is running too slow on average computer, neither I want to make low-detailed models to see in the end that they could have been made much more detailed.
The same problem with level design. E.g. I can make ugly low-poly pillars which could have been made better, or I can make high-detail pillars just to see that 100 of them bring FPS to zero...
Is there any practical 'golden rule' for such considerations? Or a kind of 'real-life' test for the engine?
This website's content is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.
I'm not really a 3D guy, so I'm just talking generally. Hopefully someone experienced will come along with specifics. I'll try and make this useful for other viewers.
What do you mean for use in indie games? What does being indie have to do with how many polygons a model should have? Typically, 'indie' developers use big commercial software if they are making a 3D game (Unreal, Unity, Source, whatever), which are professionally optimised for handing heavy rendering calculations. But they have their limits. Depends too much on what the scene requirements are (how many objects, how detailed).
As a general 'rule of thumb', the poly count should be as low as possible. This does mean that the minimum requirements to play go down and an improved overall experience for those who could already play.
Normal maps are your friend. Get well aquainted with them. Can just skip to the end to see the immediate contrast between using them and not in terms of poly-count http://www.youtube.com/watch?v=ApzVWuPa4A4
Something that is very important to keep in mind while modelling is 'Will the player actually look at this?'. Again, it all depends on the game, but how often do you stop during playing a game just to crouch down and get a good look at some pebbles on the floor, or a fence, or some guy's shoe? If a player isn't going to notice that something is very low-poly, then it might aswel be very low-ploy. Look at existing games and see how they have done things.
Your aim should not be to try and create an actual place, but just to trick the players into thinking that what they see is what is actually there. If you can do that, then you are a good artist.
More important than anything I think is to be creative and resourceful with how you approach your modelling.
Stay a while and listen.
Hello there,althoigh i am new here i would like to suggest something to you.
i personally asked this question to my mentor and teacher about polycount, he told me that I should forget about poly counts before mastering modelling...when you perfectly model a face of a celebrity or even yourself, then make it again, that time you will realize that there wasn't any need for a particular set of edges,you just need to adjust neighboring vertices to that point which gives you almost same result w/o those edges....
I think the polycount thing is gained through experience of modelling many different kinds of assets...
Try this if already have a highpoly model, duplicate it and start deleting the edges whichu think are unnecessary..and will give u exact same look as high poly model...u can start from there maybe..
Thanks for the answers.
I'm going to use Castle Game Engine (Open-source). It is definately not the best 3d engine, but it is 1) free (which is imporatnt for my opensource game) and 2) designed for free pascal. I still haven't gone deep inside it, but roughly it meets all my requirements.
Maybe I may put the question more specific. I'm still new to 3d modelling and therefore path 'from simple to complex' is more prefferable for me. I.e. I cannot start to work as pro designers do - create a general level map, then add details, then test it and add more details until the effeciency is ok. I start from modelling something small and simple like candles, rocks, barrels... and after I gain enough experience I will start modelling the castle/home/cave they are in and then place the already-made models inside. And in the future I want to make a level-generator that would automatically place the objects 'in spawn points' on the generated map.
It is much more easier for me to make a more complex model, than a more complex texture. E.g. it is easier for me to model a high-poly barrel than a low-poly barrel with everything else painted by textures. Another example: should I make models of spoons, forks, plates and other utensils or leave it to texture.
And definately it comes only with experience, but still... I do want to think in a correct way.
Of course, player characters, who he will see constantly must be high-poly, but these are only around 13 models. Other game objects quality may be sacrificed for effeciency to different extent, but I whish I could approximately catch the balance beforehands, so that I won't end up remaking too much of the art that I would have spent few years to make.
its a good and necessary thing what you are teling, and i think we can try to understand the "flow" of edges/vertices, if we understand an objects flow, the reduction in edges comes automatically...hope this help :D
[I am also learning low-poly modeling and i'm not expert in it :) ]