Runway lighting

[gsagostinho]

I asked Gilberto at some point whether the problem occurred when the lights changed from disc to single pixel, but we were able to rule that out - that was basically the only viable hypothesis. Without an error message and without access to an ATI architecture, there's nothing really I can do. You can start to selectively comment out blocks of the shader and see what makes the lights re-appear (if anything).

Yes, I think we discussed this in the following post: viewtopic.php?f=37&t=26012

It's a pity though that no one could think of any other hypothesis, this bug makes life much uglier for us ATI users (After experiencing ALS with point sprites it's really painful to go back to without point sprites...)

[MIG29pilot]

Spikes are just greasy lenses.
Proof: if you wear glasses rub some old car oil on them and look at a bright light.

Isn't the windows in a plane some kind of a lens. Especially if there is some kind of film or impurity on the glass surface. It will never or be 100% transparent right!

No. They are not, at least not if we are talking of camera lenses, because when we speak of camera lenses we are also speaking of all the guts inside of the camera which perceive light differently than the eye,especially in the area of exposure and brightness.

[Necolatis]

The FAA average luminosity seem to range from 20 Cd on some taxiway lights to 10000 Cd on some runway lights. No idea how much difference that would be to the eye.

See here for examples that also lists FAA: http://www.adb-air.com/Media/Documents/1045/Photometric%20Data-Inset.pdf

And you can see what light is what on these pages: http://www.adb-air.com/products-solutions/product-center/product/incandescent-in-pavement-taxiway-edge-light,-medium-intensity.aspx

I quite like the current ALS light though.

[MIG29pilot]

Of course, you have to consider the computer screen as well. While the runway light may be so-and-so lumens, the computer screen cannot put out great amounts of light, so even if effects are modelled accurately they may not look right to the end user looking at a screen.

[punkepanda]

The FAA average luminosity seem to range from 20 Cd on some taxiway lights to 10000 Cd on some runway lights. No idea how much difference that would be to the eye..

This is the kind of productive information I think this discussion should include.

Please don't tell me this don't effect how we perceive the light behind a window and under different conditions. It can bloom, it can blind you and it can light up the whole cockpit. If it is wet or icy the whole windshield could become full of colors and distortions.

That could be a big factor in the variations in brightness of runway lights.... Maybe that kind of data is receivable from official airport information?
I would assume that they turn up the "heat" depending on the environment. Ex a runway in a city with many lights need more light power on airport lights to clearly show the runway from all other lights.

So maybe tuning up the brightness in urban airports is a good idea?



Thorsten; You included a blend/distortion effect on cockpit windows for the sun. Why is it that you cannot do the same for bright lights from ground? Isn't it the same kind of logic?

[punkepanda]

Of course, you have to consider the computer screen as well. While the runway light may be so-and-so lumens, the computer screen cannot put out great amounts of light, so even if effects are modelled accurately they may not look right to the end user looking at a screen.

I don't talk about being blinded be my screen. But effects inside the simulator. If I wear sunglasses i will still see the effect behind cockpit window right. Only a little less painful for my eyes and darker.
I believe must of the effects / distortions happens on the windshield. Maybe except for very strong direct light it can blind the pilot a too.

[MIG29pilot]

Well now that would have to go under a topic like glass effects. Also, what if you are outside of the cockpit? No glass then, and fg can be out of the cockpit.

[punkepanda]

Well now that would have to go under a topic like glass effects.

I suggested that too. But I am not moderator here

Also, what if you are outside of the cockpit? No glass then, and fg can be out of the cockpit

I don't think it will have the need for the same effects with the bare eye. Outside the cockpit i suggest less effects on lights as in real world. Except maybe some blind effect viewing directly into a strong light close range. Ex. the landing light in walker mode.

[wkitty42]

i realized the other day, based on someone else's comment, that i didn't have to run ALS like i've been doing... wow what a difference! but then, there's no atmospheric distortions... it is almost like a cartoon where everything is really really clear... i was able to view KSFO at 5nm and ~1500 altitude with no problems... when i backed out to 10 nm with the proper altitude based on the glide scope, things were a little harder to see... enabling ALS made 5nm extremely hard to see... then i had an idea... i've never tried rembrandt... so i did... whoa... the difference is distinct! however, the same trouble with seeing the airport at over 5nms out is still there... even with rembrandt's night vision turned on it was tough...

i've gotta see if i can get out in a real craft and see how it really looks... it just seems wrong that in very clear and dry weather, you cannot see an airport's lights from 5+nm out...

[ludomotico]

In my (little) experience during night flights:

- green lights directly on taxiways are visible only if looked from a distance on tens of meters when you are directly on the line. They are highly directional and not visible from above or the left/right sides.



(the whole disc is about the size of an open hand)

Apparently, some models have a green light on one side and an orange/none light on the other side to prevent taxiing in the wrong direction.

- blue lights on the edge of taxiways are visible from distances of hundred of meters



(in size, they are a bit larger than a regular bulb)

- during a clear night, runway lights are visible from 5nm but you'll probably have problems to see them from 10nm. The pilot must see the lights before the decision heigh. The exact distance depends on many things, but it is somewhere between 1 and 5 nm: https://en.wikipedia.org/wiki/Instrumen ... h_lighting

Remember you must see at least the lights of the touchdown zone, not the other end of the runway that is 1-2nm away.

I really think all these lights must have different sizes. I'd say current lights on runways as shown by ALS are probably OK, or at least not that far from reality. But lights on taxiways, specially the green ones, are far more brilliant than lights in real life.

Thorsten, if I understand the code in simgear/simgear/scene/tgdb/SGTileDetailsCallback.hxx (and I probably don't), all lights have the same size so there is no way any "renderer", default, ALS or Rembrandt, can render lights on taxiways differently from lights on runways. Is it correct? In this case, this issue wouldn't be something that ALS could fix without the support of the C++ code.

(I know this message has plenty of "probably"s and "I'd say"s because I'm expressing my subjective opinion, not my facts )

[Thorsten]

Thorsten, if I understand the code in simgear/simgear/scene/tgdb/SGTileDetailsCallback.hxx (and I probably don't), all lights have the same size so there is no way any "renderer", default, ALS or Rembrandt, can render lights on taxiways differently from lights on runways. Is it correct? In this case, this issue wouldn't be something that ALS could fix without the support of the C++ code.

Yes, the light size/intensity is set by the C++ code for all renderers, and the renderer can't make two lights appear different that get the same size at that stage. If I remember correctly, the taxiway lights aren't set different from the runway lights, which is the reason why the taxiway lights appear too bright.

At this stage it is also possible to give the lights a direction, which would then be used by ALS to fade them out based on viewing angle, but this currently isn't on.

[punkepanda]

it just seems wrong that in very clear and dry weather, you cannot see an airport's lights from 5+nm out.

Amen!

[Thorsten]

I've just tested that I can see the ALS runway lighting in a clear night when it's really dark from up to 20 km distance. At this point, they have the visual brightness equivalent to a magnitude 3.5 star in the sky. You can't see these stars readily without darkness-adapted eyes any more (the stars when you see going out with non-darkness adapted eyes are usually 0-1 magnitudes) - which you're unlikely to have in a cockpit lit by instrument lighting. You're also looking at the much brighter other airport lights.

You see the lights to much lesser distances when it's brighter, just as the mag 3.5 stars fade quickly when the sky brightens.

Sorry - I trust my numbers much more than your intuition. Feel free to compute it yourself.

[abcaster]

I once heard a pilot declaring "field in sight" 50NM out of EKCH.
It was nighttime and pretty clear.

[Thorsten]

I'm wildly guessing he didn't see the taxiway lights first.

(Also note that with completely darkness-adapted eyes you can see Mag 6 stars in clear conditions with bare eye - so there is margin in the one or other odd case - doesn't mean we have to render this regularly.)