I’d like to be able to use UE4 to light cars and other products for producing stills (yes stills) where I can, with no limitation paint a car in certain areas with certain types of light, intensity, falloff and have each light effect only the parts of the car I want them too.
Currently in terms of light channels (the means to exclude/include) a particular light actor from a static mesh/es I have only 3 (well 1 that can be used for global / all pervasive lighting and two additional channels ) that can be used to creatively include or exclude a light from a static mesh or meshes.
Obviously in offline/DCC renderers there are no such limitations so wondering if we will ever see more (unlimited?) Light channels or whether in the roadmap there is proposed means to achieve the same without crazy processing overhead? Haven’t tried to test it but did read recently too that we still can’t use light channels with the sun actor???
This freedom of creativity is a must if UE4 is to be adopted for such use cases and I know that programs like hdr light studio if implemented for UE4 + realtime interaction + raytracing would have me ditching any other offline renderer in a heart beat.
While I’m here;) easily being able to control a light actors diffuse effect would be good too (like is possible with specular scaling).
The ease, interactivity and performance of UE4 is great for automotive and product lighting (like many other tasks;)) I feel we just need to get more of the tools that can aid in the creative process of lighting and which are available in offline renderers such as Vray,/Corona etc.
Can’t new lighting channels be generated and added to the available channels for enabling / disabling? I don’t recall and don’t have the program open currently. That is a major limitation if it’s only the 3. Diffuse can be controlled per light actor via Diffuse Boost property, but that’s all I know about thus far for light actors.
Hey. Wasn’t aware of diffuse boost but does it enable you to reduce the diffuse effect of a light actor too (one accept negative values). Will have a look, thanks.
Re lighting channels there are only 3 and as far as I’m aware no ability to add to this tally.
I wasn’t sure if diffuse boost would suffice. What exactly are you intending to do?
And instead of lighting channels for activating / deactivating certain lights for certain parts of cars and such, trigger volumes or other triggers could be utilized in conjunction with camera switching (view switches to a new camera, triggering its particular perspective and effects).
Control the diffuse effect of a light actor like you can with its specular scale property. With the specular scale you can turn off the specular effect completely. With diffuse boost you can only increase its effect and not turn it off. So like in Vray for example I could like a particular mesh either with the light actors specular or diffuse effect off as needed depending on what I’m trying to achieve.
Secondly, using Vray for a still shot from one camera (or any other offline renderer as example) I can choose for any light to exclude/include only specific actors. So I may create a spotlight that I only want to effect the front left wheel so I include only that group of parts in the include list for that light. Conversly I may create a big overhead rectangular light that I want to affect all of the vehicle accept say the rear window. Then I would create and exclude for the rear window and only the non excluded geometry would be lit by this light.
So basically light channels but not with the limit (as far as I know) of having only 3. Channel 0 by default is what all static meshes use for global lighting so that leaves you with only two further channels Channel 1 + 2 to creatively light the vehicle (exclude and include geometry from light emission). Which may be ok in very basic scenarios where 1 or two/few lights will suffice but in more complex ones leaves you needing more.
If I want more than two parts of a vehicle to have unique lighting applied to them whilst not affecting the others then I’m out of luck it seems.
as mentioned before offline renderers offer unlimited includes/excludes (comparable to UE4 light channels) for lights that offer greater creative freedom with minimal setup necessary.
Not moaning;) just trying to find out if there is a way to somehow extend the number of light channels or whether epic has plans to look at this.
I see. Yeah, triggers could be kinda impractical or inefficient to set up for lighting only specific parts with specific lights. It probably wouldn’t even work. The only other one I can think of is blendables. I have no experience with them though, so I can’t really say, merely going from reading about them. Why use UE4 if the offline renderers are so applicable and flexible?
Because I love the software and what it offers and want to use its other features such as realtime and cinematics and ideally want to do it all in the same place.
Understood. All in the same place is a reason for me too. It has lots of customization and build-from-scratch features, yet it’s technically a really challenging application.
Would anyone else on here, perhaps from epic have any information on light channels and a roadmap for expanding the limit from 2 usable channels (the third being the default for all meshes) to allow creative lighting where it is possible to exclude/include light to specific meshes?
I sometimes work with product visualization, and when we get large products (like boats) or even interiors, I also would very much like to select what objects get affected by what lights (AND REFLECTION PROBES).
This was possibly in Unity, but sadly Unreal doesn’t seem to have these options.
If you need light channels/light linking for automotive visualization then you are most likely doing it wrong. Light transport fakes pretty much always result in worse quality. I’ve yet to see a single person who can use it for “artistic” purpose in any meaningful way. Over my 11 year long career as 3D generalist, I’ve never once seen someone who knew what he was doing using per object light include/exclude.
Most people have usually done that because they did a poor job shading, and then they tried to use fake lights to “pop” the parts of the subject they did the poor shading job on.
I mean the client requests are often valid, like “make this thing more shiny, make that thing a bit darker, that surface has way too strong glint, etc…” but many people, instead of finding the actual source of the issue resort to solve these issues by blunt fakes in form of breaking light transport.
thanks will do. Despite rawalanche’s belief this is very much
not a ‘doing something wrong’ scenario.
If I want a hot gradient map in one light actor to light the windscreen only, another spot just to pick up a highlight in an alloy etc etc where each different light intensity, falloff and spread is different it would be nice to do so.
As you can with hdri light studio and pretty much all render engines and DCC software. 3 usable light channels is not sufficient for such purposes.
Well, since you seem knowledgeable, what about reflection probes?
Say you have a glossy car exterior, AND interior, and you don’t with the interior want to reflect the outside of the car and similarly the exterior to reflect the inside of the car. Especially on a thin roof part, getting the inside to only choose the interior probes and the outside to only choose the exterior probes seems impossible in Unreal.
And yes, I know all about light probe priority through radius, etc. etc, but the fact remains that non-explicit selection of reflection probes in Unreal is a really cumbersome solution.
How’s that related to light channels? Unless you use ray tracing (which you should as long as you’re doing visualization, not something like VR) reflection probes are often mandatory. But what’s being talked about here is a fake of restricting specific lights to affect specific objects.
Generally, ray tracing aside, there’s just no win scenario when it comes to reflections in UE4, or realtime render engines in general. The closest reasonable solution for the car is having all of the car as movable meshes, therefore they don’t get captured, then capturing surroundings using single probe, and relying on SSR for most of the interior details. It’s closest you can get to acceptable results, but again, as I said, unless you are doing VR, ray tracing is the way to go these days.
It simply is the case that Unreal Engine is not sufficiently tooled for those kinds of visualization techniques. Why would anyone end up at Unreal Engine as a substitute or better option for visualization when there’s a bunch of better programs out there? Are those better programs truly that lesser of quality and lacking in functionality in contrast to UE4?
Because not everyone just does one task and therefore needs the other tools and capability available or better implemented in UE4. Because people like many other facets of the engine and realtime in general and because they have the taste for good ole fashioned exploration!
I understand. It’s a rhetorical question getting at how Unreal needs some tuning and such so it becomes such a great option without so much of the frustration and difficulty it currently contains. It’s not a criticism of people.
Offline DCC software is quite different from real-time engines. I would not expect these limitations to drop any time soon… Real-time requires always sacrifice of some flexibility and visual fidelity.
