Blender Internal and GI???

Hi everybody!

Does anyone know, if it is planned to implement GI in internal renderer in the future?

Looking on all these fancy new features in current development, I think Blender is loosing its integration YafRay and other external renderers. The node system is great, but does not work with YafRay. Also render passes and new shadow buffers can be used only in internal renderer. Ambient occlusion is nice feature, but not very realistic compared to GI.

So there is an problem when deciding how to render the scene… Use the node system to have good control over materials and postprocessing and render passes, which may save hardware requiements to render complex scene, and not being able to render with GI
use external renderer with global illumination, but without new features.

So does anybody know, if there will be GI in Blender?

I don’t think that anybody can answer proberly this question… maybe such plans are in tons head or not, who knows. I haven’t heard of anything like that.

I agree, that a GI-Integration would make the internat renderer complete. (Exept displacement, sss etc.).

I heard, that the changings of the renderer API since 2.42 can better connect external renderer to the main core… so maybe a render pass YafRay-integration (or Pixie…) could be possible.

The hopes, to get someday a GI are right, because if ton wouldn’t see any benefits in the internal renderer, he wouldn’t develope it further and would set all on YafRay. So I think GI and SSS and what else are just a matter of time.



I’ve been wondering this also, with the new render api and all the good that should come from it, i don’t see anything regarding Yafray taking up the render passes features, then blender internal doesn’t need Full GI or Skydome or IBL we could just do yafray passes for that and mix them with the internal passes as well? Could’nt we?

One of global illumination algorithms is called radiosty which is available in Blender…

poison: Yes, you are right, it is one of them, but try to make som caustic effects with it…

but imho the blender internal radiosity is not useful for any kind of production… it’s an old, inefficient system.

There was an old wiki page for an internal GI and SSS project that Ray Wells had started coding for… unfortuantely i can’t find it anymore. the project is on hold now, but i think that it will be implemented eventually

Kind regards,

Yes thats possible, you can render a scene with yafray and the same camera positon ,then load the image with a add/input/image node

Pardon me, caustics is not GI. It has nothing to do with GI. The fact that a GI render as Yafray can calculate caustics, is because you need photons to calculate it, which Yafray already has. POV also has caustics when using regular raytracing, it is calculated as a separate step.

I wonder why caustics isn’t implemented in Blender yet. Everywhere in literature it is stated that the routine is fairly simple.

That is very true. I thought that Ton has Radiosity revision on his TODO list. I think it has something to do with making it multitreathed or multi processor ready or something.

But, you can always render one scene with radiosity, and then use it as reference to render that same scene with regular lamps to achieve that radiosity look. It is possible to make scenes look like it was rendered with GI with regular lamps.

But what is the Render time gain of this ? A lot of focus should be placed on praticalish Render times for animation. Baking passes and all that work around stuff is key and great to have.

Things like texture flickering and such nuances from Yafray are bottle necks in using GI.

GI is a very vague term that can mean a lot of things. At its most simple, it means that it’s taking into account all the relationships between light and objects in the scene, instead of just calculating it locally (i.e. in a non GI renderer, it’s just seeing how bright each pixel should be relative to the light sources, without considering what’s around that pixel).

Henrik Wann Jensen, one of the fathers of GI, offers this definition: “The goal is to compute all possible light interactions in a given scene, and thus obtain a truely photorealistic image. All combinations of diffuse and specular reflections and transmissions must be accounted for. Effects such as colour bleeding and caustics must be included in a global illumination simulation.”

And I agree with simhar, Blender’s radiosity is near useless, especially for animations.

I think you’re partly right. The thing is as computers get more and more powerful the cost is in person time, not in computer time. Therefore getting the computer to do as much heavy lifting as possible is the way to go, rather than people spending their lives tweaking enormously complex direct illumination setups.

Currently the whole thing is in a transition phase, where faking certainly was the way to go, but becomes less so by the day.

The same goes for programming - once to get decent performance people said “assembler” - now it’s C/C++, and in the future . . . ??? The time investment for the fractional gain just wasn’t worth it.

the GI Blender uses is also used by electric image and also by flamingo for example.
there are still situations in which this system isnt that bad. it would be great in case you
could render the GI and bake it into the texture, without the typical triangulation subdivision.

however caustics, IBL and GI would be great to have because it would round up blenders
internal renderer.


and what is it? :slight_smile:

[on hold] - Full GI and SSS in internal renderer - RayWells

Errr, you’re right. I won’t consider radiosity as full GI because it can’t do specular highlights.

You can get quite realistic results with Internal. Check this out.

I know that the tools need to be good enough for the artist to use them, but the artist really needs to do the work. But, I would like GI, tough.

Errr, you’re right. I won’t consider radiosity as full GI because it can’t do specular highlights.

Were specularity is on an object depends on the camera angle.


that image is very good but not realistic.

well it depends on what you want to render. but in some cases blenders internal just doesn’t do it.

yes, this picture is superb, but has nothing to do with GI and nothing to do with that discussion about GI-implementation…

For now you could try to render GI pass in such renderer as Sunflow and then composite it to your other passes rendered in Blender.