Re: Making a shadeless render… Just use the render-passes? In BI you’d use the colour pass and in cycles there’s that pass for diffuse, glossy and transmissive sperate. sounds like much less work than changing all materials.
Re: Is BI going to be kicked out? Ton has said on multiple occasions that there’s no plans to get rid of it, and if there were, it would first be discussed with the community.
Kick BI out? With all its UI, ghosting everything around, especially when working in cycles?
On a second thought, not a bad idea at all.
Cycles isn’t ready to replace it. But, not a bad idea at all. Trying to think otherwise.
Re: diffuse color pass - I tried that, but it is not the same, though useful as it is. This is the same case as any other - a stab at a work around to achieve something that does take more time, but yet is not the exact same thing. The whole scene is shadeless then, yes, but how to incorporate shaded objects without some costly work around? My point is that Blender Internal is useful because of its own ease of use. I can check a box and turn a plane into a shadow catcher, and another and make one shadeless - with no work around, no fuss, and I can animate this and composite the work all before lunch with no noise.
Blender has Apples and Oranges both - no point saying that the orange is an apple if I squint my eye tightly and focus on the shape of it 
[QUOTE=michalis;]Two good things happened in blender development, one of them is cycles. The other is the new UI + bmesh.
Everything else looks like projects left unfinished on their baby shoes. BI among them. Ghosts.
abc123 is so wrong…[/QUOTE]
You suprised me with this one. I didn’t think you’d be one to be out with internal. The main thing about internal is, in many ways, cycles is surpassing internal. However (imho) there are certain features from sss that look good bounds ahead of cycles e.g. hair, sss, smoke/volumetrics (which cycles have non of), less noise, etc. I’m honestly not sure that cycles can surpass these. Sometimes faking looks more real than real.
Oh, and for once, I think I might agree with endi.
For the moment I might agree with SSS though that will likely change for the better. But hands down cycles hair even in its unfinished state beats BI, BI hair is usually a self shadowing mess when ever I’ve seen examples even from really good artists.
But old school shading models are kind of dying even Pixar used a lot of physically correct shaders and ray tracing for monsters university.
Internal is not on the way out, it was out. There are some updates lately, to make it work with Freestyle and BEER, and Ton did some updates. SO I guess BI is now usefull only to do non-photorealistic render. Which is sad really, since MODO renderer focus on CPU (at least for now) and is a wonder. I’m still thinking why we didn’t have a BI revamp like we did on the GUI? But being open source, I guess it depend on the who willing to do the work. And Bretch prefer to work on cycles.
End of story I guess.
Brecht’s answer to the need to rebuild the renderer was Cycles - you can look back and read up on it around the time of Sintel and after. Internal was deemed to be too much of a mess to add features to without creating more bugs, so Brecht took to writing a renderer from scratch.
Is BI onits way out ? most certainly it is.
I think its a huge mistake to compare 3ds MAX with blender , or any other commercial 3d app. Blender community wanted several redesigns including GUI , we got them with 2.5 , 3ds max users more than a decade later still waiting to get rid of their awful GUI. Blender is a community driven software with developer that dont hesitate to do their own thing but also listen to the community.
BI has been an abandonware for as long as cycles first appeared. Its still in for 2 reasons a) It works well with old computers b) cycles is not quite there yet. A is not a big reason to keep cycles but B is. So that means its a matter of time before BI is kicked out.
Why cant we keep BI if it just works ? I think a major reason will be that if Cycles come to the position that can do everything that BI can and make you coffee, BI will be remove based on the simple fact that it will simplify blender cause developer will have to worry for less code and BI is no small part of blender. Till now the major weakness of blender has been bugs and lack of features , developer will have to worry for less code and concetrate on new stuff with BI removed for good.
But that will not happen any time soon, and as Ton as said because it will certainly be a very important decision it will be left to the community. But if you take a look at the trend right now and the art produced with blender, its not hard to believe that BI is on its way to becoming a relic.
I love the polarization - either you like Internal and can use it as a tool, or you hate it and wish it dead. As a texture painter, I love it and can’t see it going away without some form of it staying behind for using the GLSL interaction for texture painting and baking. Cycles has a lot going for it, but like has been said, not finished yet.
Can you texture paint in Cycles? I don’t think I’ve ever tried, heh.
Most of the “Blender Internal is going to be cut!” claims seem to be based entirely on speculation. I’ve never seen it stated as a plan by Ton, or any other developer. It’s still even the default renderer.
You can texture paint while using nodes in cycles, but you can’t preview in anything other than the normal texture draw mode in 3d view. (I mean that you can’t paint while render preview is on, so you can’t see the bump or displacemet until yousave the painted texture) I prefer to use internal for that because of the layer manager addon that allows to use the texture layer stack - much easier. After, I can plug them in to the cycles node tree.
texture paint in cycles,yes you can,you add a basic diffuse material,change color to image texture. (I don’t think you can get shadeless material viewport like in internal,though) but i’m sure it could be implemented.
as for claims being based on speculation or not haven’t seen it from any developers,check this post by andrew price of dingto tweet…
https://www.facebook.com/photo.php?fbid=10151098840226548&set=a.322299916547.157552.308429371547&type=1
Edit: to get shadeless viewport in cycles you use MATERIAL viewport as opposed to texture viewport.so yes its completly possible to project paint etc. in cycles.(I think glsl lights still effect it though)
Yeah for texture painting turn into BI mode.
This has nothing to do with the render engines. Am I right?
that “claims seem to be based entirely on speculation” argument is just a big fail. When anyone can go and browse blender source and see exactly how much love BI has been getting all this time -> https://projects.blender.org/scm/browser.php?group_id=9
Facts, zero speculation.
There is nothing stoping BI competing with cycles. Blender could use afterall a “lightweight” render engine. BI could take advantage of CUDA technology and offer something that could be blazing fast but not as “polished” , just for quick previews of final renders with no dithering issues of Cycles.
Question is who is going to do it. Render engines are a tricky area and its even trickier to find developers for it. Things certainly dont look good for BI, and thats a fact.
Like I said, you can, but not with the same feature set or interaction with the textures - just try the texture paint layer addon like shown in BI and you can paint on the bump while you still see the diffuse textures… until that same kind of addon can be made to work with Cycles and we get the same view port interaction with bump, we will need the BI system because that is where the textures are created and stored while working with them (images I mean). Baking as a start of that - in BI, I can bake the vertex paint and Ao as a start to a UV mapped image, and use that to begin my painting as well as bake basic diffuse colors - no baking in Cycles yet, so I still feel we are talking about a half finished project already calling for the instant removal of the previous system.
Don’t get me wrong - I love Cycles and what we can do with it, and many artists are already adapting their workflows to use it for every thing they can - but it isn’t done yet, and it doesn’t do everything. Until it bakes, paints, and runs in the game engine, you got a ways to go for a large number of users.
IIRC BI doesn’t run in the game engine either that is its own separate engine. I remember making a similar remake and been corrected by one of the devs, could be I was imagining all this.
What, they don’t have a maintainer? By that logic, a ton of other areas in Blender are going to be cut as well. Nurbs, simulation, etc.
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.)
Yeah, the viewport isn’t actually BI either, though both are rasterisation renderers. Another reason to keep BI, by the way: there’s a ton of raytracing engines compatible with Blender, but how many rasterisation engines are there?
No , in order to follow my logic you will have to have an alternative to Nurbs that can do what Nurbs can do and more. Then yes Nurbs will be going out of blender. But since this is not the case , NURBS are here to stay for a very long time, even though they are partially at best implemented. My logic is not really that hard to understand , at all.
Except this is not really true of BI and Cycles. They are completely different technologies. It’s like saying Multires sculpt should be thrown out because we now have dynatopo. Or that we should get rid of FK type deformer bones because we have IK deformation, or get rid of Lattice deform because we have mesh deform, or get rid of cloth simulation because softbody simulation is more accurate anyway.
yeap thats exactly what I am saying. If the user workflow is very similar it does not matter if they are completely different technologies. Which are not , actually they are very similar, they are both render engines they fulfill exactly the same roles.
If you implement a new thing that works very similarly to Nurbs and fullfil the role of Nurbs so well that people abandon NURBS for it, why keep it ? So you end up with a graveyard of code that will be a pain to maintain ? that simply is not logical.