Cycles refraction seems "poorer" as Vray

I don’t know if it’s a big or what… But cycles refraction seems to have a lack of detail when I compare it to vray. This is the same scene, omni without shadow, pure refractive material with IOR of 1.5. All cycles max bounces to 128.

What do you think? Should I report this?

Cycles (last official and nightly too):


Vray:


Wow that looks bad, could you upload the .blend? I’m curious…

EDIT: I think this is just fine, in ‘real life’ the material always has some reflection. The black parts you see are supposed to be reflections, so this will probably be resolved when you add a glass material or recreate a glass-like material with the reflection and refraction shaders.

Vray just handles it differently

I have the same effect with glass material. And anyway, the issue is not about the black parts, it’s more about the richest effect, if you look at the bottom of the glass, or around Suzanne"s eyebrows, you’ll see more sophistacted refraction effect.

In fact, I’ve got the same result as Vray with Blender Internal. I’ve decided to report it.


Attachments

BI.blend (1.12 MB)

Have you tried raising the transmission bounces in Cycles?

This is the render result of your scene in Cycles with a glass shader…


Cycles’ refraction shader is not meant to be used on its own. What you’re seeing (or rather not seeing) in your tests is the missing (internal) reflection. No material is just “refractive” without any other surface attributes (as richardvdoost already pointed out), that’s why the glass shader works as intended and that’s why the refraction shader is meant purely as a building block for more complex materials.

1). Try enabling caustic bounces if they’re off.
2). If you’re just using the refraction shader with the transparent one, you would need to add a glossy shader for custom fresnal.
3). Use a lightpath trick to get the same type of transparent shadowing as you would in BI, Cycles doesn’t do this by default because of its physical nature (all light going through glass is technically a caustic path).

Not sure that is entirely true - one example off the top of my head would be heat haze in air. However in the OPs application I would agree.

Good call!

I think that will reflect too, it will have a critical ‘reflection’ angle.


Although with hot/cold air the difference in IOR is so small that the critical angle for a reflection will be close to 90deg

Which is the reason people see lakes in the desert :slight_smile:

Only under some circumstances (e.g. a mirage). For general heat haze as depicted here - the effect is pure refraction.


You can’t compare Vray and Cycles. And, if you have Vray why would you even bother?

That’s only due to the viewing angle though. It’s not magically non-reflective.

True - but in terms of blender shaders - we dont have one that can accurately simulate total internal reflection.

If you wanted to recreate the above scene - you wouldn’t use a reflective material (e.g. glass) since there is little if any contribution from reflected light rays.

To close the debate… it could be due to the fact VRay (and BI) pur erefraction are not so pure. Even If you completly disabled reflection, refraction use some kind of intern reflection. If I use a glass shader for Cycles, I can reproduce the cool refraction effect. Everything is ok.

@IkariShinji : you were totally right

Theoldghost : you think I shouldn’t work with Cycles since I own Vray? What a funny idea of evolution.

you think I shouldn’t work with Cycles since I own Vray
vray is awesome, i’ve used it long time, but it doesn’t integrate as well as cycles preview in blender, if vray had the same degree on integration, vray definitely whens.