Light edge falloff additive to other lights

[kevininrussia]

Where two lights outer circumference dim area meet, it looks like the lights become additive on the overlay.

https://i.imgur.com/EnSG4zU.png

[flashgordon]

The fact that lights are additive creates all sorts of problem areas, like this. The luminance of a pixel needs to be the maximum of all light sources, not the sum of them.
46214

[Kelrugem]

The fact that lights are additive creates all sorts of problem areas, like this. The luminance of a pixel needs to be the maximum of all light sources, not the sum of them.
46214

Or the addition should be with respect to their brightness/colour :) It looks like that dim+dim=bright is applied here which creates such jumps to bright areas

[HywelPhillips]

I concur.

Dim + Dim = Bright makes for some ugly-looking jumps, and I don't believe rules systems use Dim + Dim = Bright for system effects either. Two dim light sources shouldn't make a bright light source in the overlap.

Worth a play around with the suggestions to try max luminance, or max luminance and add colours to see what looks nicest.

Cheers, Hywel

[pindercarl]

Where two lights outer circumference dim area meet, it looks like the lights become additive on the overlay.

https://i.imgur.com/EnSG4zU.png

Thanks for sharing, kevininrussia. There is an issue with the lighting that cause the dim regions to render at half the expected brightness (i.e. half rendered at one-quarter). This did not impact the areas of overlap, so light that overlapped had a gradient of bright to half-dim to bright. This should be fixed in the next test channel release.

[Kelrugem]

Thanks for sharing, kevininrussia. There is an issue with the lighting that cause the dim regions to render at half the expected brightness (i.e. half rendered at one-quarter). This did not impact the areas of overlap, so light that overlapped had a gradient of bright to half-dim to bright. This should be fixed in the next test channel release.

nice, thanks :)

[kevininrussia]

Here is the same scene with same settings in new build. There is still an additive to the light overlap.

https://i.imgur.com/w4jPFlg.png

[pindercarl]

Here is the same scene with same settings in new build. There is still an additive to the light overlap.

https://i.imgur.com/w4jPFlg.png

Yes. Lights are additive.

[johnecc]

Carl, is that by design, or an error that is being looked at?

[pindercarl]

Carl, is that by design, or an error that is being looked at?

That is by design.

[Ludd_G]

Is it ok to offer my option that lights really should not be additive?

I don't think it's 5e RaW or RaI that lighting be additive? Although it would not be a surprise to me or my players that I've been ruling stuff wrong all this time! Maybe in other rule sets? I'm also personally not a fan of it aesthetically.

Is it actually a problem to code it to not be additive, or simply a choice between two equally difficult/easy options to code?

Please take this as a data point rather than a critique of your great work on lighting so far.

Cheers,

Simon

[pindercarl]

Is it ok to offer my option that lights really should not be additive?

I don't think it's 5e RaW or RaI that lighting be additive, I've certainly never played that way? Maybe in other rule sets? I'm also personally not a fan of it aesthetically.

Is it actually a problem to code it to not be additive, or simply a choice between two equally difficult/easy options to code?

Please take this as a data point rather than a critique of your great work on lighting so far.

Cheers,

Simon

5E does not have a rule for overlapping lighting. However: 1) In 5E, when darkvision overlaps dim lighting, the character can see as if it is bright. So it is reasonable to infer that overlapping dim lights result in bright. 2) Real world, physical lights are additive. 3) There is no good mathematical model for overlapping lights that is not additive.

(Note on point one: there was a bug with this so the feature is currently disabled.)

[Ludd_G]

No worries and thank you for your clear explanation of the thought process for it being the way it is. Much appreciated!

Cheers,

Simon

[flashgordon]

3) There is no good mathematical model for overlapping lights that is not additive.

Are you applying each light to the image individually, or are you building a lighting map that is applied to the image at the end? With the latter approach, as you apply each light to the light map, you can just have each pixel do a MAX(currentValue, appliedValue) for each color channel for each applied light.

[Lo Zeno]

With the latter approach, as you apply each light to the light map, you can just have each pixel do a MAX(currentValue, appliedValue) for each color channel for each applied light.

Regardless of the approach, I don't think they should choose the MAX value, and instead keep doing additive light, because light... IS additive, as pindercarl said. Two lamps make a room brighter than just one. In real life, you can even see the same effect as the pictures posted by kevininrussia, the only difference will be that the edges won't be as clear-cut because of the scattering caused by particles in the atmosphere, other diffused lights and lack of "perfect" darkness around.
I'd rather see some approximation of light diffusion, to make the gradient of the overlapping light "smoother", than not have additive light.

[Griogre]

Hmm, to play devil's advocate for 5E. Dim Light + Darkvision = Bright light while there are no rules for Dim Light + Dim Light. Since they specifically mentioned Darkvision + Dim light is bright, but didn't say anything about dim light + dim light, I always thought it would be dim light, myself.

What are you doing for Bright light + Darkvison? Are you making that brighter than Bright Light? If not, why not limit dim light + dim light to dim light too?

My two cents, take them for what they are worth.

[Kelrugem]

Regardless of the approach, I don't think they should choose the MAX value, and instead keep doing additive light, because light... IS additive, as pindercarl said. Two lamps make a room brighter than just one. In real life, you can even see the same effect as the pictures posted by kevininrussia, the only difference will be that the edges won't be as clear-cut because of the scattering caused by particles in the atmosphere, other diffused lights and lack of "perfect" darkness around.
I'd rather see some approximation of light diffusion, to make the gradient of the overlapping light "smoother", than not have additive light.

Yeah, precisely :) I also highly prefer additive light, and I just tested the situation in the first/seventh post of this thread: If I change the falloffs to 100, then it looks very natural again :) (though you can actually create such rather sharp edges in real life, too, by the way, like in interference, but usually not with torches next to each other :D) So, it is here probably not anymore about how lights add, it is (again :D) how the standard falloff is defined :) So, again about: Aesthetics or rather mechanics?

As already said, I also strongly vote for that the lights add with each other; due to that this is how it happens in real life one could argue that this provides the best compatibility with all rulesets. If a ruleset tells otherwise, then I'd wave that away and say that I prefer more realistic physics. Such "simplified" rules of certain rulesets only exist to simplify gameplay at a real table, I think. However, here it is about FG, and I think FG should take an advantage of using the computer behind the light calculations :) Hence, I'd ignore any exceptional rules which deny addition of lights; and honestly I have never seen any real table which did not treat all the lighting rules extremely sloppy. I have never seen a DM at a real table who actually cared whether a light is dim or bright :D So, there should be no harm to ignore any separate rules about (the denial of) additions of lights since they are not really important

[Sterno]

As long as it doesn't end up doing this, like it did a build or two ago. It's a pretty bad look.

https://i.imgur.com/nA3SQhI.png

[HywelPhillips]

Lighting *IS* additive, but it also falls off (from a point source) like 1/r**2, not like a step function which is bright for 30 ft, dim for 30 ft beyond that, then to nothing.

Game systems are trying to do something non-physical to simplify the situation, and FG is trying to do something to appease both the physics/aesthetics people AND the people who would like the lighting to represent the rules.

I know it is late in the day to bring this forward as a suggestion, but is this a sufficiently divisive question that it might be worth having "Rules-style lighting" vs "Physical-style lighting" as an option?

You could have two sets of default lights, one with falloffs set to provide maximum clarity for rules system representation, and one which is designed to avoid these jumps in brightness and unphysical artefacts? Using the MAX shortcut might be another way of doing it, or doing the physics right and using 1/r**2 with a cutoff so you don't have to keep calculating lighting out to infinity. Whichever, like Sterno I find the example with dim + dim = bright very unappealing, because it is unphysical and (in my opinion) aesthtically displeasing too.

Physically if those two light sources are point sources of the same intensity, the light at the centre of the "lens shape" where both overlap in the middle, 6 squares away from each of the two lights, should be 1/(6*6) + 1/(6*6) = 2/(36) = 5.5555% as intense as the light one square away from either of the lights. The intensity closer to one of the lights by one square should be 1/(5*5) + 1(7*7) = 6.0% Two squares closer to one of the lights it should be 1/(4*4) + 1(8*8) = 7.8% as bright. Three squares closer it would be 1/(3*3) + 1/(9*9) = 12.3% as bright.

There is no point along the line joining the position of the two lights where the light is dimmer than it is at the midpoint between the two lights. That "lens shape" of bright light in Sterno's example is not physical, and that's why it looks weird.

If that's not possible as an option personally I would like to have something much closer to physical reality and more asthetically pleasing than the dim + dim = bright solution.

Best regards,

Hywel

[Neovirtus]

Yeah, I have to agree with the binary theory folks here. In 5E D&D we have two levels of brightness: Bright light, and Dim light. I think all lighting should be normalized to the two values, as I don't recall any rule in 5E that says "if two or more dim lights illuminated the same square, it becomes bright light." If you have 7 light sources all illuminating the same square with dim light, it should be dimly lit. But that's, just, like, my opinion, man.

[HywelPhillips]

And in case you don't believe my maths, here are two LED candles on a stone floor which I just photographed:

46409

Note that the brightness monotonically falls off from either light to the mid-point, which is where the minimum brightness occurs.

Hywel

[Neovirtus]

I can't view your attachment Hywel, not sure what the issue is.

[HywelPhillips]

Don't know why that's not working, let's try again:

https://www.silksoles.com/temp/YER09021.jpg


46410



Hywel

[Surge]

In 5E D&D we have two levels of brightness: Bright light, and Dim light. I think all lighting should be normalized to the two values, as I don't recall any rule in 5E that says "if two or more dim lights illuminated the same square, it becomes bright light."

You need to stop considering the lighting system only in terms of 5E. It needs to be system agnostic.

[Neovirtus]

You need to stop considering the lighting system only in terms of 5E. It needs to by system agnostic.

5E is what I play, so that's my perspective. Sorry.

[Kelrugem]

And in case you don't believe my maths, here are two LED candles on a stone floor which I just photographed:

46409

Note that the brightness monotonically falls off from either light to the mid-point, which is where the minimum brightness occurs.

Hywel

I think no-one doubts that light is additive :) As I wrote it is more about the default settings of the falloffs :) So, yes, maybe some settings about the default falloffs may be good, but that would be more settings, and one still needs to define the standard value of this setting

I think addition should be kept of course as others, you and I argued before :)

[LordEntrails]

Keep in mind, many RPGs use two light levels (bright and dim) to represent two ranges of light levels. Those without penalty, and those with penalties. They are really not trying to say that all areas with bright or dim really have the same brightness (i.e. number of lumens), but rather that in terms of lightning, they all have the same penalty or game effects.

Most folks don't expect VTT lighting to mimic the aesthetics or the rules, they expect it to appear like they see in the real world. But, to some, they want the VTT to represent the rules and ignore the expectations the real world places upon the players.

Hence why I like the steep fall offs, it helps appease my real world expectations while clearly indicating the "rules".

[chaiwalla]

I agree that lights should not be additive. No game function and jarring and unrealistic visually.

[Moon Wizard]

I've responded here; closing this thread as it is old.
https://www.fantasygrounds.com/forums/showthread.php?68472-Two-colored-lights-placed-nearby-removes-the-color&p=600741&viewfull=1#post600741

JPG