First off all, I’m from the “more renderers” the better school. Each has there
strong points. The BI is fast - fast for animation. It’s full featured.
If you’ve got one thats 75 percent there - why not Turbo charge it?
Is it a question of age? The Lightwave default renderer was doing Broadcast TV work
10 years ago.
I keep hearing, the code is tangled. Well why not clean it up? Is it ready for
the grave - or can it be Frankensteroided?
The source code is readily available so there is nothing stopping anyone doing work on it if they want to. But I don’t see a great flood of developers wishing to do so.
+1 for not wanting to see BI die off. It is fast for many many things, and very flexible.
With the code within BI full of hacks on top of hacks and bandaids (according to those who tried to work with the code), chances are few would be willing to do the work needed to clean up the code in a way where you can safely add features without breaking others.
It’s probable that the state of BI’s codebase is a major reason why we now have the Cycles engine, the idea that cleaning up BI would be as much work as starting from scratch and getting better results.
Cycles will also be like that when it’s ready to replace BI.
Actually, cycles will turn out far more flexible then BI because of the node based materials…
the idea that cleaning up BI would be as much work as starting from scratch and getting better results.
And that’s Cycles.
The source code is readily available so there is nothing stopping anyone doing work on it if they want to. But I don’t see a great flood of developers wishing to do so.
@Richard Marklew
Well it seems like the group has to be offering money for anyone to get serious from
what Ive been reading. So the group decides, or BF, more than the developers, right?
+1 for not wanting to see BI die off. It is fast for many many things, and very flexible.
@artao
yeah, I definately agree.
and getting better results.
@Ace Dragon
better results in SOME areas, I’ll grant you. But since when do Blender folks
want to be building renderfarms to do their animations? Last time I had
a render that was an hour + a frame - 2002. The Render of one minute took a whole month! + tied up my main machine. Came out nice though.
when it’s ready to replace BI.
@FreeMind
See renderfarm above.
It’s not so much a question of can it be done but who will do it. What kind of improvements are we talking here?
What kind of improvements are we talking here?
I haven’t a clue Jay. But for almost 10 years Production quality renderers have been common: Mental Ray, Brazil, Cebas Final Render, Chaos Vray, Lightwave default, Houdini Mantra.
Where did BI get left behind?
BI can’t be improved because there is no one around to work on it and the one person who was working on it often, Breacht doesn’t want to work on it because he thinks the code is a mess. If he want from the stuff he did for Sintel in the render branch to this it should tell you something.
And seriously BI is the reason most Blender renders look like the belong in the 1990’s. Everytime I look at a BI render I feel like am watching something from that era; look at something like the Diablo 2 cut scenes, tranformers beast wars etc. Anything from that era will look like what most BI renders look like now
I think that the BI could be improved a lot, and easily.
For example the current raytraced AO could be transformed into a good pathtracer with a little effort.
In the last 2 years I sent some bug reports to improve the BI, but most of all where closed without investigating.
Probably there are no developers available.
GPU programming is difficult and time consuming, it seems to me very hard to add in Cycles all the features of the BI in a short time and doing it in CUDA, OpenCL and C++CPU at the same time.
I use Blender for planetarium animation production, at the end I had to write my own rendering engine, just to support that simple Fisheye camera.
BI renderer is outdated not fast and seriously limited.
I am not sure how you can state it is fast and good compared to what is a standard today.
I would not mind having BI besides Cycles because for some renderings it is still faster as Cycles
currently can render but once that is adjusted I am not sure what benefit BI would have.
It would be like Maya Internal cool for quick previews - useless for production rendering.