I’ve read the documentation regarding new features and improvements for 2.54 but there seems very little in regards to the game engine (understandably with Sintel it’s mainly modelling, animation and rendering).
I was just wondering, at this stage, whether there’s any significant benefit to migrating my game from 2.49b to 2.54? Are there any significant speed increases, awesome features I’d missed or anything else?
Ah okay, thanks for the info! There’s not really enough of an incentive for me to switch yet, so I think I’ll just stick with 2.49b for the time being Thanks again!
the new blender version is 20x more light than the 2.49 version, the shadow buffers are incomparable between the two engines and the 2d filters works muth better on the 2.54 than on the 2.49, i can tell you, dont waist your time with 2.49, it s old and slow.
I made a test that i posted here a while ago, i used a scene with about 1 000 000 polygons, then I tried to open this file on the both versions of blender on blender 2.49b I got a 1.5 FPS and on the 2.53 33FPS.
Take your conclusions, if you want a better program use 2.5x if you want to keep the fight with 2.49 go ahead, less competition on the game market.
I’d go with 2.53; even though it’s a beta and isn’t quite stable enough yet, like leonn said, it’s quite a bit faster than 2.49. Since (if you plan to use the 2.6 final build when it comes out) you have to migrate, you might as well get used to using and coding in 2.53.
Haha, brutally honest. I’m going to run some tests myself and if the difference is as much as you say then I’ll definitely switch. I’ve been making all sorts of LOD and culling systems in a bid to squeeze the best performance possible out of 2.49, so this could make it a lot more painless.
here is the thread that i made http://blenderartists.org/forum/showthread.php?t=195714
Unfortunably i deleted the 2.49 image, hehe, but you can see that with 1 000 000 polygons i got 33FPS, and all the objects where texturized and there is a spot light buffering shadows
cool, also, it’s not about the quantity of updates, it’s more about the quality. The dynamic loading and freeing won me over, not so much the extra speed because I usually try to keep the engine 30 - 40 frames, 2.5 just makes that a little easier.
Well if its a legend it s a living legend on my pc because my frame rates are very hight on the stand alone games that I made with 2.53, the only problem is that when I make a runtime, i dont know why the corner of my screen get blue, but i think that this ll be no problem on future releases.
But is weird to think that the blender with all those menus and stuff is quicker than its own clean runtime, i thought that it was a small version of the same program.
I still find that the instability of the GE in 2.5 is putting me off from moving to it. Also, and someone please correct me if I’m wrong, but I’ve read that python 3.1 is still a slower than python 2.6. So I was also planning to wait untill pythons caught up a little and all the libraries/modules are updated.
New sound system (one that works and supports ogg, mp3, wav, and probably more)
Real-time external IK targets
Ability to use raw OpenGL code (allows for GUIs, etc)
Dynamic loading/freeing of assets
Python API improvements
There might be more, but I can’t think of them right now.
As far as the speed improvements, here is my hypothesis (it’s untested): Blender 2.5 is better with threading (and some other things), which gives the embedded BGE player more resources to make use of. However, I would expect the 2.5 Blenderplayer to have similar performance to that of the 2.49b Blenderplayer. The Blenderplayer has been known to have better performance to the embedded player. So, maybe the embedded player has just closed the gap now. But, ultimately, there have been no dedicated efforts to improve performance (minor or major) that I know of.
Also, the bug with the blue bars in the Blenderplayer has been fixed, so it should be fine in the next release. You can also grab a build from graphicall.org (if you can find one that includes the Blenderplayer).
well ninja goliah, im not a game guro, but i have brains too.
if the 2.5x isnt is quicker than the 2.49 version how can you explain my framerate?? maybe is some kind of graphicall improvement, i really dont know but i m 96.59% shure that the 2.53 version of blender is indeed mutch light than the 2.49b, to prove it download this file:http://www.mediafire.com/?6t18wwyk9l986g2
the test249 file is the runtime created with 2.49b version of blender and the test file was created with 2.54. The two files are the same, both have some spheres and a spot light with 2484 of buffer size, the poly count on the scenes are 921.000 squares.
My framerate on the 2.54 runtime version is perfect, on the 2.49b i have about 1fps.
You guys ll need a very powerfull computer to run this file, sorry about that, but i was pushing the blender to the limit to show that the new blender s version is mutch better than the old and one.
Im not a guru but i can correct some guys aswell, i hope that you guys find this usefull.
So would the dynamic loading and freeing alow me to switch textures during play?
Or does it allow me to do dynamic linking on the fly? That would be good for future expansion of my game ideas, but not so much right now (not enough assets to be a drain yet).
But it could be very useful if my game gets bigger.
Any one know where I can find some tutorials on using raw OpenGL code? I’ve wanted to do that for a while…
Hmm… I’m still building my games on 2.49, even if it’s slower. I can still use it as a test ground. I tried modeling with the 2.5x and for some reason, when I selected a few faces, others were selected too! I’m not sure if it’s just on my pc or if it was at that time, but that prevented me to use 2.5x. So when the 2nd RC comes out, I’ll migrate!