I dont know how many blender users works in game dev, but I think there is many.
I am for example.
I think Blender lacks of some features that is required in gamedev.
First: the exporters. Game dev studios uses other 3d apps, not Blender, so we need to export our models to Maya, XSI etc. With simple models we can use .obj, 3ds etc. But for example with double UV we cant export⌠maybe collada will do it.
And other things:
-correct UV calculation when editing
-face cut (for making destroyed models with good UVs)
-textured mode when in edit mode
With your suggestions Blender would be much more useful for game artists.
Iâd also love to see some lod controls (switch in/switch out distances) that could be previewed via the Game Engine. Maybe the LOD data could be written out via Collada or a simple text file.
The icing on the cake for me would be a texture palette where you could view all associated textures on a model or a scene.
Well there are pretty good exporters for OGRE, and CS. So if you can move your development to one of those engines, then you have very few problems to worry about, if any.
Using Blenders internal engine is also a posibility, but right now itâs somewhat buggy, and the current renderer is a slouch. Although all that is supposed to change with the new OGRE implementation (if it ever happens that is).
Well, anyway, Iâm almost positive that collada will solve most of the problems you noted here.
I use XSI with Blender, with obj exporter. Our engine can load collada, but Blender collada exporter is not perfect (but in work in progress I use it). When I finish a modell, I put it into XSI. All models must be in XSI at my workplace.
Itâs cool to have multiple UV sets now, but it would be great if realtime shaders could represent it in the viewport. For example color map with first uv set multiplied with a shadow/AO/dirt map using second set. At least some game engine examples would be needed showing such features, so I we could have a peek on the result starting the engine.
NGons. Itâs bad that even quad-support is very poor. (If a quad gets concave, it might show up overlaping tris?!)
More, UV-friendly mesh edit tools. Existing ones donât care about UVs or even destroy them. The Blender way of modeling doesnât allow mesh modifications after UV is done.
Talking about UV-friendly tools: I was very happy to see, that vertex snap is available in 2.43. Maybe there should be an extra option for inmediate vertex collapse (and UV change!!!)
Any developer reading this, please donât think Iâm blaming. I really like the way this program changes. Iâm a 3D artist at Black Hole Entertainment, using Blender (and XSI) for UV maping (LSCM/ABF), seam correction (3D paint) and for home enjoyment
The last year I lose a nice work because Blender game-engine canât do âtrack toâ with bones.
I think itâs a must.
We need to move a bone (like another object) without need of an action/ipo.
I reckon getting a really good Mobile 3D exporter would be dead handy
The big headache Iâve run into trying to make a good one is triangle strip generation, which may be a problem other exporters also have. Maybe this should be implemented lower level in Blender and accessible through Python?
I reckon the Collada interchange will become really important. Itâs important to be able to load a Collada file, work on it, export it, and not lose any data in the process, including stuff Blender doesnât yet understand. I donât know if it can do this already, but if it canât it should!
The texture palette sounds like a good idea to me. A natural extension of the current image window.
I think the whole point of blender being a game making software as opposed to a modeling app. is that it is one stop shopping. You dont need to know programming, you dont need to know pythonâŚetcâŚ
There are tools for people who have extensive programming knowledge. Blender I think was ment to be the easy way, as far as game creation goes.
-Plop in a scene
-Plop in a charcter
-Set a few controls
-Game time!
Fast and efficent. This is what it SHOULD be. I think this is what most people tend to see blender as. It has that image of being the easy game creation tool. This is what needs to be expanded and built on. I am no programmer, but all the articles I have read all say that a games limitations is not the designer, but the programmer. These two people are so diffrent they are like apples and oranges. How many art students do you know were taking c++ at the same time in high schoolâŚnot many. There are some great artists here, crappy programmers, and some Great programmers, who are not so good at the drawing thing. Blender needs to find the middle ground. Where good artists can easly use and find the tools they need to make their ideas come alive. âwell learn pythonâ is not a solution. Ogre is not a solution. Blender has started the ball rolling, we just need some people to give it a small push.
Could you be more specific about which editing tools donât properly preserve UVs? I know extrude and make face handle this very poor, but canât think of others.
Show me more than .0001% of PRO game developers who use blender period.
being the exception is not being the ruleâŚ
You are a great modeler and texturer endi. I respect you and your work, but blender is years if not a good decade of becoming this butterfly that everyone wants it to beâŚlets just keep worming our way for now and be happy with what we have. 3dsmax has game dev. by the you know whats. I am talking from education to work place. No way blender is ready to take that on. What we can do is make it better at what it does for now. Talk of turning it into a pro game dev tool is sorta pointless right now.
Why is learning python such an âover the topâ suggestion? I just donât get the reasoning here.
I started using blender with no art skills, and all I could do with python is write a few functions. I took BGE python programming to heart, and thatâs where I found my niche (actually thatâs where I learned most of my general python as well), but when it came time to actually make models and environments for that python code I wrote, I didnât say: Ughh, really, the devs should make modelling easier for programmers, âwell learn how to move vertices, and assign UV texturesâ is not a solution.
Instead I learned from general blender modelling and texturing tutorials, and it was no longer a barrier for me.
Just look in the stickies, and you should find at least 2 decent python starter tutorials (I have one in there, and so does blengine).
Itâs really not as hard as everyone makes it out to be.
If your a professional artist working in a production enviorment you probably dont have the time to sit around learning python to the level needed to adress the limitations you run into with Blenders modeler.
Getting back on topic (I donât think he was talking about the target run time engine) I think the point Endi is making is a very valid one. As wonderful as Blender is itâs current toolset makes things difficult when producing artwork for âprofessionalâ games. For example a car might be remodelled quite drastically many times over a games development period. If you lose your uv coords everytime you slice into a mesh - pretty soon you are going to be looking at other 3D packages which preserve this information.
I think itâs worth pointing out that people have so far been discussing this and that piece of software, which is really only a small part of the picture. In a studio environment (especially a larger studio) any piece of software is part of a much larger pipeline. Here, at the end of the day we need to get assets into the PS3 runtime as efficiently as possible, so although the core tools are provided by Maya, thereâs so much custom stuff bootstrapped on that really you canât just say âwe use Mayaâ.
So essentially, thereâs no reason why Blender canât be used in aspects of game development now - itâs core tools are pretty good - as long as itâs intergrated into the pipeline well.
The art team would not be required to find their own solutions to âwork aroundâ Blenders limitations, those limitations would be considered when setting up the pipeline and appropriate tools and techniques developed by the tools team.
Better UV editing tools would certainly be nice, an equivalent to Mayaâs âalign and stitchâ would be very useful.
Being able to plugin in custom renderers would be nice to, to preview shaders in a standard viewport.
Cheers, dude!
because Blender game-engine canât do âtrack toâ with bones.