OTOY Forums

An idea that has always been in the back of my mind is to have multiple environment maps.

in order to use the lighting passes, you need to have geometry with emitters and set their layers. Wouldn't it be really powerful to have multiple hdr maps to control lights and therefore have more control over in post? The possibilities with using hdr light studio would be endless and fine tuning would be much faster than using geometry.

If a question that comes up is "why would you need that when you can manipulate all of the lights in light studio with one map"... well, in post, you have one pass for it, so its all or nothing.

multiple additives environments node at the same time would be a great option yes.
Currently you can kind of simulate this using several HDR with a mix texture node and plug it in the environment texture.
but you can't have several Sun for example...

calus wrote:multiple additives environments node at the same time would be a great option yes.
Currently you can kind of simulate this using several HDR with a mix texture node and plug it in the environment texture.
but you can't have several Sun for example...

True, I thought of that, however, in post, it would still be one lighting pass so no advantage.

I just threw the suggestion to light studio to add 2 features, grouping of lights and the ability to export multiple files for every light and/or group of lights. They had a great response to it so it sounds like it may happen.

I agree more lighting options is desperately needed in octane (Spot lights, Barndoors, multiple hdr contribution options). The Light passes are very powerful , would love to see this evolve into a way to link lights to certain groups of geo.

yes! linking to geometry or not sure, but since it's physically based, maybe to a render layer and it only effects objects in that render layer.

@sohara20, light linking doesn't make sense for a path tracer, the cheating must be done in post (as Gah suggests)

Why doesnt it make sense?- Its in almost every other Production path tracer out there.
Yes its, "biased" but makes creating art directed lighting much easier. Even if the lights are still contributing to secondary bounces , still having the option to link the direct light contribution to a group of geo would save alot of time in post.
If its truly something the architecture would not allow, then a more robust way of adding arbitrary number of RGB mattes to control light passes in post is needed.

and btw weta has their own unbiased spectral renderer. I would be very surprised if they didnt have a smart way of controlling light contribution in the render or in post. Otoy and weta are both from NZ so that's got to mean something.

The end goal is to make Octane a production renderer is it not? If they don't come up with novel solutions to problems all productions have they will fall to the wayside in a span of a few years just like maxwell render and any other renderer out there that cannot be adopted by studios with broad feature requirements.Its good to have individuals that like your software but alot of people will not waste their time learning software studios do not use as it has no effect of their paycheck. I dont know what otoy's biggest market is but its only a matter of time before the ILM's of the world want to render on gpu, if they are smart they can win a large chunk of this. (my two cents), and octane rocks right now so i have few doubts.

As I see it :
The first enemy for a path tracer is fireflies, every time there's a mathematical mistake, approximation not physical enough , the result is the same : Firefly
Each time you introduce cheat from old practice like additive materials or light linking, what you do is destroy the optimization made at sampling level by the renderer.
And the other side of light linking, you can introduce energy loss, your render doesn't even look physically correct
So light linking doesn't make sens for a path tracer.

I don't share your vision of the industry, what I experienced:
last 10 years "every production renderer" out there has evolved to physically correct,
modern productions avoid this practice, light linking is the devil
This is not a 2016 workflow, this is the old way to do things.
Current workflow is as follow: first make a physically correct render then cheat it in post as the director want (so the picture is cheated in the end but still "feel" correct).

I don't think light linking has anything to do with the small place of Octane in the feature film industry
The Obvious reason is the geometry limitation (and geometry export time), typical production scene have hundred of million of polygons, case closed.
pre-V3 license scheme didn't help as-well for larger adoption.