I see a lot of talk about getting something coming out of the rubble that was blender. In my humble opinion (and probably worthless too), I think that new effort should be as focused as possible given the state of free/low$ CG now. I think it is a matter of where the free/opensource community has most their eggs right now.
A. there are plenty of free/open source modelers out right now that really rock. Wings 3D, hamapatch etc.
B. there are a few really good free/opensource renderers out there. Virtualight, BMRT, POVRAY.
C. there are a few free apps like blender and Art of illusion that have animation, modeling and rendering.
D. the fact is that these full projects are decent in their feature set but they lack in one or more of their catagories because development is so broad.
I think that a character animation and effects package that was focused would not have to worry about making modeling tools or renderers…it would only have to be able to interface with them.
It can be done…Just look at something like hamapatch as an example of using other renderers. It will let you use 3 different stand alone renderers. Lightwave will let me import my obj’s from wings and animate them as a subpatch object.
A proposed feature set:
Well, basically look at martin hash’s animation master and copy everything you can for character animation.
partice system/effects and spriticles
be able to use sub-d cages so there is no assigning bones to thousands of vertexes nightmare. maybe interface with the hamapatch guy to handle spline objects too…don’t need to do it as much with the other modelers since polys are fairly universal.
Built in texturing/UVmapping/procedural engine
linked to some outside renderer. maybe more than one.
Scripting and open API so that when all the core is good it can be expanded.
anything you can think of.
What do all you guys think about this? I just saw a fierce amount of interest being stirred up and suddenly realized what I am missing most as a consumer of free cg software…and I figured that by focusing more we could have a product that would kick butt in half the time of some O.K. complete package. greedy I know, but are’nt we all.
Gregg
Post what you think…cause something like this could get off the ground quick.
RAYTRACING! The one feature almost everyone wanted.
Actually, we might need to get rid of some features for some time. How about staying away from modelling stuff? Good thing I’m kidding. Seriously, could we get rid of the shading type caused by ALT-Z? It doesn’t chane unless you press ALT-Z–not even moving things will change it.
One of the things that made Blender so nice was the fact that it was a single package that did modeling, rendering, and animation. The rendering and the modeling/animation can be seperated easily, because they are so different in nature (modeling and animating involves interacting heavily with the user via a UI, where as rendering is–more or less–an automated proccess).
I think that, to start with, the project should be an integrated modeling/animation program, and should completely ignore the rendering aspect of things (with the exception of “keeping the door open” for a renderer–or rendering interface–to be added later).
I think that the project should focus on surface-patch based modeling, polygon-based modeling, intuitive texturing and material tools, and a completely non-linear animation system.
However, I still maintain my opinion that we should wait what happens to the existing version of Blender first. Of course, that is, after all, merely my opinion: I don’t claim to be right.
i agree with cessen. the thing i liked about blender was that you could flow smoothly from modelling to animation to rendering with none of this “export to such a format”, “import to such and such an animator”, “export as this other format”, “render with some-odd renderer”
blender is all in one. that makes it nice, quick, and gets rid of export problems.
Madcow-
i see what you are saying, but if you notice I am not suggesting that blender has none of those features…just what I would like to see in any new app.
The rest-
I know what you mean about having an all around proggie…that just takes a lot of time and things tend to lose focus along the way. I have just heard so many people say “i wish they had focused more on character animation and less on the gaming side”.
My only real point was that there are plenty of free apps out to do everything but real character animation and good special effects. I do agree that some type of modeling tools would really need to be included just to make different types of animation feasable.
Also, the render with some odd renderer issue…just look at the way hamapatch does it. just because you don’t write the renderer does not mean you can’t have seamless integration. You just did not waste time programming your own renderer.
Really what I want is an App like animation master…as far as char. anim. goes.
If the guys at hash would focus on a product like this that was actually open to more than splines and it’s own slow renderer…THAT product would rule the CG world. That is why i think a project focused in this direction made with a little more artist friendly interface would easialy make the blender freeware community look like a tupperware party.
I guess all this depends on how expandable the code base is for blender even if they ever do get it released. I would seriously consider rethinking much of the UI as well to make it accessable to the masses and a little less esoteric.
In an animation program i want it to be able to import a model that can contain uv textures, vertex colors, and mehs, spline as well as nurbs. I would also like the animation program to be about to add lighting and render in opengl. Then have a feature to export each frame to a raytracer (would love it to be lightflow) Then have an app inside of the animator to collecte all the output renders from the raytracer and be able to place them together and save as a raw avi…
That to me would be a dream come true although if it imported blender models
NO RAYTRACING! Blender has a good and fast scanline renderer. Instead of replacing it with a raytracer, why not just put some more export formats directly inside blender without using scripts?
NO RAYTRACING! Blender has a good and fast scanline renderer. Instead of replacing it with a raytracer, why not just put some more export formats directly inside blender without using scripts?
Why no raytracing. I would go for selective raytracing and a decent GI rendering engine. You will see the quality of work increase rapidly.
Exporting is an opinion, but very often there are errors, unexpected results, and so on, … If a raytrace renderer is built in in the program, it would be so much easier and fun to use.
I totally agree about optional raytracing, at the moment im using the POVanim export script and it works great and while it does export animation I shudder to think how long it would take to render a decent length animation. I think intergration into an a raytracer like lightflow, povray etc should be done. I also think blender should do sort of a lightwave thing have a modeller just for modelling and a seperate renderer and a seperate game engine bit, I know blenders great because everything is linked but I think people would appreciate less clutter if they only like one thing. Just because it seperate doesnt mean it has to be difficult and maybe help in its deveopment if people concentrated on their own bit. (Im think of a set of programs, Blender modeller, with scan line renderer and anim and effects, BlenderRenderer, and Blender game creator). I doubt if blender wil go down this road but just a thought.
