How To Get Identical Results with Dome Light and GI Environment

So if I place an HDRI in the GI Environment of the Environment rollout of the Render dialog, I get smooth looking results on my object. If I use that same map in a vraylight set to dome, I get very noisy results. What settings do I have to set the dome light to to get the same results as the Gi Environment? I assume I need to increase the sampling subdivs of the domelight, but if you know exactly how much I need to get identical reuslts that would really help. Thanks for any advice.

- Neil

Its bit more complicated. With dome light you get direct light with override you dont. You also get specular which you dont get from override. So you are missing a lot of real physics when you not have dome light.

As an alternative you can also use vray light with skylight enabled.

Now that’s confusing. What do you mean by you don’t get direct light with override? If I put an HDRI in the GI Environment, I do have to turn GI on to get it to work, but then my surface does receive direct light from the HDRI, not just bounce light. As for specular, I assume you mean that with the GI Environment, it doesn’t transmit specular reflections, whereas it does with a vray light set to dome (unless you mark the vray light as invisible).

- Neil

Generally you just need to raise the subdivs of the dome light until the noise goes away.

Best regards,
Vlado

Umh u lost me…

IBL generates:
Diffuse light
Specular
Reflection
Refraction

Override generates:
GI
Reflection
Refraction

Light using skylight:
Diffuse
Specular
Reflection
Refraction

I’ve rendered examples in zip below.

Hope this helps.

If you want smooth results as usual increase light/mateiral/etc samples… :slight_smile:

Edid, Vlado/chaos is there any reason why when using light+skylight my specular pass appears to have clamped specular?
JPEG.zip (785 KB)

Thanks Vlado.

- Neil

DADAL,

If I make a teapot on a plane with a 100% white color, then set the GI environment to white and render, then I get a white teapot and plane with slight darkening under the teapot.

If I then take the same scene, make a vray light set to dome, make it spherical, set multiplier to 1, make it invisible, and render, I get an identical render (noisier, of course).

So that’s what I mean, they both provide direct light and bounces. It’s just you have to turn “Enable GI” on to see the direct light of the GI environment. Whereas you get direct light with the vray dome light even without turning GI on. So your chart is correct, you get “GI” with the override, but the GI includes Diffuse light and diffuse bounces.

- Neil

Try adding VrayLighting and VrayGlobalIllumination as render elements when doing these tests and you’ll see how Vray treats the different methods. As DADAL says, when doing IBL from override you only get indirect light. This is the reason the lighting is smooth, but also the reason why you won’t get the nice, sharp direct light and shadows. When adding the HDRi into the VrayDomeLight, add more shadow samples and the noise should go away. To get faster, nicer and more balanced lighting with this method, I recommend our own Akin Bilgic’s Vray Optimization Guide which step-by-step shows how to get the best results with these lighting setups (and most others).

Could you post an example where IBL gives you sharp direct light and shadows? Because again, if you render using both methods for me they provide identical results in all the scenerios I’ve tried (solid color, HDR image, etc).

- Neil

Its in the zip file I send I posted early in this topic… Just open it up and go over images you will see the difference between override and Dome…

So for example, do you mean the extra darkening happening between the images “Dome.lighting.jpg” and “Override+GI.GI.jpg”? If so, I didn’t see that in any of my renders. Let me do a few test renders tonight and post em.

- Neil

There is also noticeable difference in shadow sharpness, blotchy GI samples and so on if you look at those 2 renders as well as lacking specular component when using GI override solution and so on…

I didn’t say you’d never get sharp shadows with indirect light, but working with high res hdr’s with bright spots can be a hassle without the domelight. Below are two very simple renders, first one with direct light, the other with indirect light only:

To get the sharper shadows with this hdr and indirect lighting I had major issues with splotchiness regardless of how much I upped the samples. It’s probably some to get the full dynamic range from the hdr as indirect light, but I didn’t get satisfying results. And of course, this will vary from hdr to hdr. :slight_smile:

Well, I get the lack of specular reflections, since it doesn’t provide anything to reflect. And I get the blotchiness since the samples on the vray light are kind of low. But I hadn’t noticed in my tests the shadow sharpness issue. Again, will do some test renders tonight and see if I can replicate them.

- Neil

I have to admit that the more I think of it the more the terms direct and indirect light are too nebulous. For me as an artist, I read direct light as light coming from the light source and hitting a surface (whether that light source is a tranditional cg light or the sun painted on an hdri), and indirect light means light that comes from a surface and hits another surface (or bounce light). From what I remember back in the old Brazil days, the reason a skylight (or in vray, the GI Environment) is considered indirect light is because the light is “bouncing” off the environment and then striking the objects in your scene. But as an artist, an hdri, while technically an environment, isn’t an object and so should be considered a direct light source and not a bounce light source.

My brain hurts :slight_smile:

- Neil

Totally agreeing with you. The interchangeable nature of this software’s terminology can get really confusing. I’m also seeing only indirect light when casting IBL from VrayMeshLights. I have no idea why it’s not producing direct light, even though Store With Irradiance is turned off and the GI override is disabled. An issue for another time, I guess.

Can you show me an example of this? This sound incorrect to me..

The blotchiness comes from GI not from light. Light is raytraced and GI is well it depends which solution u use…

So it must be really hdri dependent. Attached are two different hdris (one is outside, the other inside with light coming through windows). Note they both look pretty identical in terms of shadow sharpness (although the environment one tends to be more noisy) So that’s why they seem the same to me, must just be the particular hdris I’m using.

- Neil

I might be mistaken, but it looks to me like both those hdrs have relatively large light sources. I think the difference between dome/GI-override is more pronounced when using hdr’s with way smaller sources, e.g. the sun. In my example the sun spot was only a few pixels wide, and that seems to be a problem for the GI engine when it’s not cast as primary rays (resulting in lots of noise and splotchiness). I’ll have to continue testing with more samples to get any wiser, I suppose.

Heya

Yes those examples are very close to each other. Both of these approaches do should look similar at they highest settings with the difference of lack of secular component in Override case. The smoothness and noise that dome produces is related to quality of IBL texture. If its a night shot with lots of smart light sources then you need a lot more samples then for example a sunny desert with 1 sun as light source. You always have to add more samples to the dome in order to clean up the noise. Its really up to you how you use. But general practice in industry is that you use dome for direct light/specular and overrides only in special cases. There are drawbacks to GI based lighting solution.

1. It can back fire on render times
2. Relative to GI settigns which can be very high for certain scenarios (as well as Dome approach)
3. Your shadows tend to be more blurry and less accurate, you can see it in ur example.

Try rendering something bigger, a viz/interior and so on you will see how much longer it might take than direct Dome.

Try doing some compelx tests, this teapot might be misleading you..