Why poor graphics?

Okay this may sound rude and I’m sorry if I’m offending anyone, but, Alot of great BGE Game have been released but have Very poor graphics. Why is that? Is it because BGE simply can’t handle High graphics or is it because not enough people try to make them look good? I’m not saying that they NEED good graphics to be good or fun, I’m just curious because I wanted to make a game myself sometime soon :stuck_out_tongue:

I’ve been using Blender for around 7 years now, and just recently started using BGE as well, Actually that’s why I started using blender, to make games, but back then the game engine was really really bad… But I started again 3 weeks ago and achieved results like this:


This runs perfectly fine, 60 FPS with:

  • GLSL Shading
  • Variance shadows
  • Multiple textures and Normal maps
  • realtime SSAO( Ambient occlusion ) Only pre-baked on the models.
  • SSAA ( Anti aliasing )

But of course this is far from a complete game, this is the only 5 objects in the scene…

Why?
Because no-one has put in the effort.

Most completed games are made by programmers, not by artists, and thus have the programmers idea of good models.
(For some bizarre reason, programmers often seem to have PC’s that can’t handle the graphics either)

Artists rarely complete games, and so only team projects have a hope of good graphics. Unfortunately they also rarely come through, and I can’t think of any completed team games (with the exception of the perpetual WIP Krum)

I think some people limit there graphics because they have thousands of objects in a scene or a big map,
Some people are obsessed with Frames per second,
Also I think I can have terrible graphics, and a working game, I can replace models, not behavior,

also some people have never even messed with GSL… not that GSL is not amazing…
because python and logic bricks and animation keys or translation matrices have already got there plate full just trying to implement a concept, let alone make a pretty looking thingus… and form can be important, but function is fun!!

Also bigger model = harder to rig, unless you rig before a sub divide,

  1. What you have there is something simple and easy for your gpu to render.
    2.Not everyone using blender have a powerful PC.
    3.There are mostly noobs/kids so you can’t expect for any suprising work from them.
    4.Blender has performance issues when it comes to graphics and the artist has to put a lot of effort to optimize the game.
    5.Lazy people.

Mostly free time and the right motivation.

I agree with sdfgeoff.

BTW. A game is more then just logic and/or graphics ;).

As it takes a long time to complete a game, high-detailed models will extent this time a lot. As usually you have to balance between what you want, and what you can achieve with your resources (time, money, skills etc.).

In terms of high-detail, I wonder how it looks like when you bullet add holes to your high-detail models (in game). (and how long it takes to make that)

I could acutally put myself under the Artist calsification.
Most of people who do games are programmers, or don’t have PC’s powerfull enough to support GLSL.
I’m 14, and i’ve been using Blender since the 2.46 days. I’ve been training all this time focusing on the BGE, and i have to tell you, if you want a game to run with a high framerate, you’re gonna have to optimize a lot.
I have actually achieved great results with the BGE, even if sometimes i decide to use UDK. But is not mostly the artists fault, or the programmers fault, it is because of how the BGE is optimized.

Great results CAN be achieved in BGE see link:

But, as socker said much better than me, BGE is NOT especially optimized, or easy to optimize for. I have a simple demo of my game, a barely populated level, and I am already running into optimization issues, where my high-end gaming PC can stutter sometimes. I think this is an issues with BGE and how it calculates collision in the bullet engine… UDK does it much better, much more optimized. Your little scene might run fine at 60FPS, but add in a whole level with collision and it is another matter.

Indeed. Also I think BGE has compatibility issues with some cards.

OP if you’re serious about making games I recommend trying a proper game engine like UDK or Unity. BGE is nice for learning or prototyping but you shouldn’t be considering it for a full game. Unity is good, and I believe it supports Blender.

“Poor graphics”, because a good game doesn’t need a very good graphic.
Take a look at the classic-games like Tetris or Super Mario. => “poor graphic”, but a lot of fun. :slight_smile:

Another reason (like it was mentioned before) is, that there are a lot of people with slow PCs.
Before I’d buyed a new PC, I used my old Notebook for game-creation. The programming and Logic-creation doesn’t need a lot of power, but when I’d pressed the holy “P” everything crashed or works very slow.
The only solution: No GLSL => no shadows and much more…

Since I’ve a new, very powerful Desktop-PC, I’m trying to create Games with better graphics. [Shaders from Martinsh, e.g. “Lensflare Playground” or other very cool stuff. (Thank you Martinsh :slight_smile: )]
Also it is possible to use more lights, the new dynamic variance shadows and a lot more.
This is e.g. a feature which isn’t included in the free Version of Unity.
Unity uses baked shadows, I think => low power consumption.

But a very big problem, I’d found, the games look and work very well, but doesn’t work on older, slower Systems.
So it is necessary to find a good mix between graphic and the Systems where it should works on.

I love this question when it is asked; for not only does it introduce new ideas to the table, it also gives me a chance to say the same thing slightly differently.
The issue with game development in our ecosystem is roughly the following:

  • New users to the engine start over ambitiously and approach their project with a “learn-en-route” workflow. They watch tutorials or practice designing models and artwork as they work on the relevant area of the project.
  • Users wish to “make a game” but the quality of the game actually often comes second as they hide behind the term “noob”, which allows them to excuse poor models. People may not give enough realistic criticism to suggest that the artwork isn’t good enough.
  • Programmers make games, Artists don’t.

First of all, let me be clear; most of us will have fallen into category 1 during the initial phase of our development. When we first become aware of the engine, unless we have prior experience in modelling or game design, we often end up in category 2 very quickly.
My third point / category isn’t one you should quickly jump to scorn. In fact it is the nature of the BGE and most other game engines that you do not need to be able to design 3D models in order to make a game. As a result, that part of the game design process often comes second to the programming components.

Furthermore, people often default to implementing “placeholders”. However, in the nature of hobbyist game design, you rarely have time constraints that would limit you to rushing out a proof of concept, so why bother creating something you’ll end up redesigning later on? Do it right; first time.

^^^^

Fair enough.
The fact that there isnt much feedback and criticism is a problem in the ‘journey’ to get better.
And another issue is that , the wip forum gets very populated by games that barely get finished or they are just simple …made in 1 day with an idea thought in 1 h.

Python is a slower performance language (not by a really noticeable ammount). I know I would rather sacrifice shaders and pixel lights just to have a billion logic bricks running to make the A.I. seem smart. <- Yes, Logic Bricks. I’m cool. :rolleyes: Plus my computler is old. . .

Great results CAN be achieved in BGE see link:
http://www.youtube.com/watch?v=k-LHB6oOdO0

That is a team project probably made on a beast of a machine.

Performance before graphics, as I always say :wink:

Its on a pc custum built : 500$ Not a beast.
And not a team (i did that by myself 80%)

I consider myself to be an amateur but I also can see your point. The BGE has potential for good graphics indeed. But I admit that I don’t know much about realtime SSAO and SSAA. Could you maybe explain how I could use this features for my game?

I have programmed my whole life, and I have always had freetime game projects (mainly C/C++ and Java). Sorry, none of them are currently available for downloading… As I just recently found out Blender and its game engine, I really can’t say at the moment anything about its performance, but I can say something about my game projects. I have currently one test game project to learn to use BGE.

Usually, I tend to work heavily with the game internals - for me, games are their internals :slight_smile: Graphics comes slowly to game (and sound effects even slower, if ever), because usually when making changes to the game internals you need to remake all the graphics. It really does not encourage to put effort to finished outlook until you have a working game, and sadly as a programmer you usually loose your interest to the project at the exact moment you have somewhat working game logic (if no-one is paying you to finish it). All the games I have done were never really finished, and all of them were using the temporal “demonstrative” “placeholders” as graphics at the moment I left them alone.

Making a game that looks finished is not an easy task to do. Usually games have much more content than players ever realize. To make it look finished, when getting your game logic complete, you need to remake just about everything to fit together. The larger your game is, the more laborious task that is, and usually non-commercial game developers just don’t have enough interest to do it; it is enough to know that it could be done. One of the few reasons I could imagine I would polish my game projects would be trying to attract other developers to help to finish the game.

About Python. Yes, it is not the fastest language in the world, but it works in multiple platforms without compilation, unlike e.g. C or C++. It is probably still true, that if you would like to make a game with cutting edge features, you probably would need to build many additional features under scripting level with lower level languages. But still, you probably would have some sort of scripting language at the top of your game engine to fit the pieces together. You may do it by yourself by gradually extending your data file manipulation routines to a programming language, or you may use some existing one.

Implementing game AI can be very difficult with any language. Compiled languages help a little bit, but usually game AI’s are full of algorithmic tricks to make them work in real time. I have done several board game AI’s, and at the moment I would really think twice to change the language to compiled one (I usually start all my projects with Python). AI algorithms often involve lots of such work, that can not be avoided when changing language - they tend to build huge data structures to dynamic memory, and building and managing is slow process even with C++. But I might be wiser about this subject later, as one of my main points of interests in the current game projects is the game AI.

@ AcAnimate
Anything less than photo realistic is poor graphics to you?

Nope, I don’t believe I ever said that.

I never said that you should compare anything to my “photorealistic” result, my question was simply, why alot of finished BGE projects had really poor graphics, when some projects do have stunning graphics. So my guess was that BGE couldn’t handle high end graphics.

Right,it cant handle hq graphics.