I think it’s promising and valid.
The problem is that all the technical infos are not organized in an user-friendly way.
There is this for a general infos, this discussion but it’s pretty much impossible to follow and randomly full of tech infos
given on requests.
So a newbie can’t really go into learning of an efficient workflow that goes deeper than
“select Filming Log Encoding, select a contrast preset and add a ASC-CDL node”.
IMO a specific place where everything is explained in a easy to learn organized way, including workflow, would give it a boost of interest.
BTW… is it possible that everytime someone asks if it’s possible to add different GI methods, the discussion turns into Luxrender tonemapping and color management?
Hm, that’s not a bad idea. Maybe I should create a “public” resume (stripped of details that don’t need to be published everywhere like my street address) and link to it from my signature.
True. The simplest way I can explain the filmic blender color management is that it is tone mapping (kinda sort of). lol. It does a better job of mapping the linear values into srgb space than the default blender srgb profile.
Even years ago; this was an option. Also, Lux years ago to today is quite different (Classic LuxRender vs LuxCrender ‘2.0’/LuxCore). In some ways, LuxCore is closer to Cycles than LuxRender was, especially in relation to biased settings, that were impossible to use in Lux ‘Classic’
You guys all missed the point. I was asking more realistic options for Cycles and you said it’s realistic and that one guy even said it’s more realistic than LuxRender. No, it’s not. Also I know the size of sun lamp in Cycles makes sharp shadows, but again it wasn’t the point. And we get back to the original question why Cycles can’t have more realistic rendering options.
Hi, hm, thread title ask for biased Cycles and iirc this means more unrealistic options.
I ask developer for implement Coherent Path Tracing, old and simple technic to double the render speed but it is to biased. :eyebrowlift2:
But there aren’t simple options that make it simply “more realistic”, no button “Make me renderings realistic”. The only difference I see is that LuxRender supports spectral rendering (basically using more than 3 values for the color). The other differences are mainly the materials and the integrators. Bidirectional Path Tracing converges faster for some cases, but that is the only difference.
And biased renderings don’t need to be less realistic than unbiased ones. In fact they can converge to the exact same result. The difference is that if you take two renderings of the same scene, but with different random seeds and mix them (take the average) the unbiased one converges further (the result gets better, less noise), the biased one doesn’t (at least not always). It’s more a mathematical property and says nothing about correctness. Photon Maps are for example biased but converge (with unlimited memory) to the correct result.
Do you mean http://graphics.ucsd.edu/~iman/pdf/[Sadeghi%20et%20al.%202009]%20Coherent%20Path%20Tracing.pdf? Coherent Path Tracing is unbiased, says so right in the abstract.
I was using basic/default materials in both scenes. I suspect it’s not really the materials, because even if you have the basic material it looks better in LuxRender, because the lighting is realistic. The lighting either looks realistic or not. You can’t fix that with better materials or anything else really. The only problem in Lux is of course that being unbiased it takes forever to render away the noise. But then again noise is problem in Cycles as well.
Is Glass in Luxrender a PBR material? Because I have understood that Glass in Cycles is just a shader not entirely corresponding with reality. You have to combine Glass shader with other nodes to get a better result. I am not referring to Caustics which is a well known weak point of Cycles.
Uncolored, absorption-free glass is accurate in Cycles. As soon as you give it a color, to fake absorption, the reflection is also colored whereas it should be (almost) white. Mixing a refraction shader with the desired glass color together with a white glossy shader increases realism. The mixing factor of course then is governed by a Fresnel node.
What I find funny is that LuxRender is really hard core, editing the materials requires MIT degree, but it still has a simple material library with couple of basic material settings. Imagine if we had a material library for Cycles. Yeah, let’s just keep dreaming.
@Krice
Lux has the IOR presets included with shaders, it’s not a library of materials.
Exercise on materials for a week and you to can give yourself an applause for achieving MIT status
A switch to make it thin (aka architectural in some sw).
Unchecked, it should produce current glass.
Checked, refraction is replaced with transparency and normals should be handled automatically (fresnel backfacing IOR issue).
Offset slider (default 0), where thin transparent rays are angled based on viewing angle, to fake thin refraction.
Fresnel power (default 0), where -1 means more reflective and +1 means less reflective (known as anti reflective coating, very much used). It only affects the mixing, not the IOR through the material itself.
Fake shadow color output socket. It handles color and IsShadowRay automatically. It gives a strong indication that Cycles physical glass shadows are crap to the unusable level (needs caustics and diffuse bounces checked if underlying surface shows glass shadow), and user can then input that to a transparency shader and add it to the glass shader for fake shadows. It’s important that it’s the color that is outputted and not the transparent shader itself, because typically users will want to modulate that color somehow (like brighten to create fake caustics).
The thin switch appears to be very important, because the IOR/fresnel/backfaces is kind of unintuitive to understand until you really understand the reasoning.
Another thing that is bothering me about transmissive materials, is that you need to perform some really bizarre tricks to obtain translucency with highlights and bumpy thin glass, ref this thread. I consider myself pretty familiar with how materials work, but I would never have thought about Geometry Incoming -> Refraction Normal to obtain that effect
