“The rumors of my death are greatly exaggerated …”
Okay, so Blender has (currently) two built-in renderers. Or rather three, if you properly count the Game Engine / OpenGL, which by-the-way is the that I would like to see be developed more aggressively.
Each renderer approaches the same problem in a fundamentally different way, using the same geometry database so that images align perfectly, and it’s all integrated into the same system. (Which also supports external renderers.)
So … “what’s not to love?” Use the best tool(s) for the job at hand. It’s all good.
For instance, in film-photography terms, I find that Cycles is a great “soft-box.” Good for realistic, diffused, scattered-everywhere lighting if you have a decent GPU with which to drive it. But at this point in its development, rather difficult to control. If you want spot lights and accents, I find that BI gets me the results that I want faster. So … use both at once. The compositor is your oyster.
personally I think we split hairs here, whether BI does not get any new development or is thrown out , is the same result. BI is dead anyway, unless something changes. Just take a look at any software that has been frozen in time. In the world of software is either you move forward or you die. Blender is no exception. The only reason why BI is still around is because Cycles is not quite there yet.
Come back to this thread when Cycles is there. Maybe a few years , who know, but it wont be far. Then we can talk about it in a whole new perspective. Blender developers speak about the current the situation and only the current situation. I highly doubt that at least in the their back of their minds they dont consider the removal of BI.
GI is what it is by definition, light bounces, etc., and it is often referred to as indirect lighting, unfortunately BI’s indirect only works with approximate gather and the emitters cast no shadows. Easily faked though.
What you’re doing there is called speculation. You’re speculating, that Blender Internal is on it’s way out because of lack of activity in the source code. Here is DingTo stating it hasn’t been dropped. The only people I’ve ever heard say Blender Internal is going to be cut are users on this board assuming it will be.
Tyrant, I mean that the output of the materials and textures run inthe game engine, not the renderer itself - because if I remember correctly, cycles materials and textures cannot be used in the game engine while I have made walk throughs with objects made with blender internal materials and textures that were derived from baked vertex paint and hand painted textures. I don’t mean that the Internal is in the game engine, maybe that is my lack of verbage there.
And yes, just because it isn’t being added to doesn’t mean it is broken - geebus, there is so much we can do with internal, this idea that it should be ripped out seems to be the spite behind cutting off our own nose.
But all you have to do is use nodal materials in Internal, and you can access them at the same time you add cycles material outputs to the same material. That way you use a scene with Internal and a scene with Cycles and you can composite them together, especially if you use render layer masking
Huh. It seems there’s more division than I originally thought.
I also like how everyone seems to think that BI being most commonly used for NPR in the future is something to scoff at. It isn’t that cycles as a raytracer won’t ever be able to do NPR, but it’s still far off from the control you can get with BI. But then, if it isn’t nonbiased, it seems to be a non-renderer here, cycles included.
(No seriously, have you ever consider how terrible a nonbiased renderer is for rendering things that rely on highly emotional aesthetics? Yeah, you can get abstract stuff out of it, or even mimick a style if you cheat enough with the model, but hardly anything pipeline worthy.)
This really speaks to the point I was trying to make. Honestly, I wouldn’t mind cycles replacing BI. The only thing is, it needs a lot of stuff for it to do so that I’m sure it will ever get. See, internal can manage photo-realism and stylistic renders. Cycles can handle photo-realism and… it really doesn’t work with NPR, which I think is a very valuable style. Cycles just doesn’t have nearly the amount of control that BI does. Also, take Alex Glawion’s work. It wouldn’t look even half as beautiful with cycles. Would it possible to implement biased methods in an overall unbiased engine? Than, it would just be a matter of working out the little chinks, and BI would be away with. That, and working with glsl.
Stylization is more an artistic technique than something a renderer has as a feature. Cycles is versatile enough that people can use it for stylized things too. Poked Studios gets similar stylized results with Cycles or Blender Internal. There are examples of cartoon characters done with Cycles as well. DingTo even made a Toon shader for Cycles for his GSoC project which is in trunk already.
Again though, if you prefer Blender Internal why not just use Blender Internal? I don’t really understand what the actual debate here is.
I’ve been pretty much sticking with Blender internal and mainly the reason is because I’m used to using it and it’s perfect for Cartoony style projects whereas Cycles is good for photorealism. And I mostly do cartoons instead of realism.
Any renderer has its pros and cons. Things it makes easy and things ut makes hard…
People still use renderman…
Having both is extremely useful. It allows me to make choice and pick the tool for yhe job…
Approximate ao is fantastic but has its tradeoffs…
Cycles is great too but doesnt scale well enough for me… not by a long way.
Bi has had development recently… interactive viewport render fir a start…
It may be dead to some people but it only really dies when its no longer useful… and its so very usefull fir all styles… not just npr.
Cycles can be slow to resolve at 720p… when you switch to full hd it can become seriously un-tenable!! especially if you have lots of polys and lots of glossy…
it perfoems great for me with just a few objects but complex scenes are only really viable for stills… in my experience at least!
Lots of rough glossy (mixed with diffuse by fresnel) is my main reason for using cycles… In BI i get teh poor mans version using speculars…/environment maps… or even occatioanlly I;ll raytrace it too…
There may be more commits to the Cycles branch these days, but BI is obviously not dead. After rendering three Blender Open Movies and contributing special effects to the fourth, BI has matured into a reliable workhorse with an impressive track record. Cycles on the other hand is still under construction, and since path tracing has its own set of known limitations, there is no way to tell whether it will be able to fully replace BI and how messy its code will look at that point. The only thing we can say with reasonable confidence is that BI will continue outperforming Cycles on noise-free NPR jobs. And that, my friend, is reason enough to keep it around!
By the way, there are still BI patches in the tracker waiting for reviews. (nudge, wink)
1024x2048, 96000 faces (3 levels of subsurf, unnecesary but useful for this example).
Render times with 2.68.2 r59080:
Blender Internal + Freestyle: 30.21
Cycles CPU + Compositor: 27.12
Cycles GPU + Compositor: 12.31
Why do I get the feeling that most people making this kind of claims have never even tried to use Cycles as an NPR renderer?
By the way, if I’m not mistaken, what Freestyle does is some kind of postprocessing on the final render. If that is so, a direct integration of Freestyle and Cycles should be possible.
Cycles has not got a real shadeless shader that can just be turned on with a switch - you have to jump through hoops to set it up. I can render a 1920x1080 frame of animation with shadeless materials and blend textures with edge rendering at 5 seconds a frame with Internal. Cycles would take me forever on the set up for that, not to mention that that frames would take longer to render.
Just the toon shader itself doesn’t lead me to think that Cycles is better, since smoothing is still an issue in Cycles for shading and affects the shaders in certain instances. Brecht still has plenty of work, as well as the rest of the devs working on Cycles, and until we have baking in Cycles, I don’t see myself giving up on internal any time soon.
