The landing strip lights are barely visible ...

[abassign]

The landing strip lights are barely visible and are no longer visible beyond 300-500 meters of altitude ...

I'm using the 2019.2 / 3 version of FGFS but I've known this problem for a long time in the track lights that are very weak and not visible at a distance, can anyone tell me what I need to change?

Thanks for collaboration

[Thorsten]

I've known this problem for a long time in the track lights that are very weak and not visible at a distance, can anyone tell me what I need to change?

The lightbulbs - they're very weak in reality as well.

[abassign]

The lightbulbs - they're very weak in reality as well.

Too bad it's not as you say because the reality is very different ... this is a REAL example:

Not only that, but as you can see from the images, even the lighting disappears in the most distant part of the landing strip ... which is absolutely impossible in reality ... The tracks light up to be visible, not invisible

This is an example of a night landing in LIME airport, the PAPI are quite visible at night, while the path of disecsa virtually no, the lights in the center of the track are visible only at the end of the landing, the side lights are visible are in certain angles of view but then disappear. In fact the situation worsens if a 2K or 4K video is used .. in this case the lights are almost completely invisible, try to believe ... I remember that once there were sprites that allowed to see lights even at a acerta distance (as in reality) causes the halo. If the lights are treated as single points of light, for example 2-4K display the situation becomes "dramatic" because their pixel is much smaller. I would like to have the opinion of others.
This video was made with a 1K video for which the situation is certainly better even if it's not exciting ...

[Thorsten]

Too bad it's not as you say because the reality is very different .

So is FG - like many other features, the visibility of lights is calculated taking into account a couple of relevant parameters - under the right conditions, they can be seen from a good 20 km away.

It's quite easy to engineer a screenshot where they disappear in 200 m distance and compare that with a digital image with a long exposure/high CCD gain which is waaay more sensitive than the human eye under the same conditions and then claim that 'something must be wrong because my FG scenario doesn't match what the CCD camera can provide' - however that doesn't prove anything.

If the lights are treated as single points of light, for example 2-4K display the situation becomes "dramatic" because their pixel is much smaller. I would like to have the opinion of others.

If you treat them as a single point of light, that's like your problem - select point sprites instead - the code then obviously does not do that (a fact which you complained about bitterly in another thread when you essentially claimed the light should be confined to just the emitter model - when it was explained to you that this will be completely aliased-away by rasterizer beyond some distance).

I would like to have the opinion of others.

Go searching the forum - there alternating claims that the lights can not be seen far enough or that they can be seen from too far away.

But light propagation isn't a vote - we know the base emission power, we know the apparent brightness under perfect conditions as a function of distance, we know the brightness attenuation by the atmosphere as a function of distance and we know the max. sensitivity of the halfway darkness-adapted human eye, we have a reasonable idea of the eye's brightness adaption state - the rest is just math.

[abassign]

I am full of your comment, but it is a pity that the eye does not work as you say ... and that today the 1K displays are less and less, if you want and time, check with a 2 or 4 K display and you will notice that the individual pixels, even at night, they are hardly visible, so much so that we see even less than the 1K example I sent, in which we see very little. To say that CCDs are more sensitive than the eye may be true, but whenever I happen to land at the same airport in the movie at night ... LlME ... I, who will surely have super powers, I see exactly as in the movie and not as seen on the screen. Therefore, if you have a minimum sense of reality, you simply have to say:

.. "In fact it is not that great, you have to work at it, maybe if someone wants to see if it is possible to do better, we will be happy to welcome it in our big family"..

[Thorsten]

I am full of your comment

I have no idea what this means, but since it doesn't sound pleasant - I'm sorry to hear that.

if you want and time, check with a 2 or 4 K display and you will notice that the individual pixels, even at night, they are hardly visible

Which is why we do not render lights as single pixels but rather extended discs in the first place - unless you select that in the GUI. Did you miss that statement?

(Oh, and I have a 4k display and am quite happy with it...)

but whenever I happen to land at the same airport in the movie at night ... LlME ... I, who will surely have super powers, I see exactly as in the movie and not as seen on the screen

Really? Every time? Same visibility every time, same residual light - every time? Strange place you fly to.

I do have trouble imagining this...

Anyway - in FG we have weather - sometimes I see the lights from 20 km away, sometimes from 2 km - it really depends.

[abassign]

"There are no deaf worse than those who don't want to hear" ... it seems clear to me ... anyway the 1K movie I sent you is clear or better dark ... if it's okay for you ... it looks happy, it seems to me completely out of reality. In the movie the day (or rather the night) was clear and clear, not by chance the lights of the airport area are better seen.

Secondly, if you can find out, where are the lights for the runway and the descent path implemented?

Thank you for your precious collaboration.

[abassign]

This is a detail of what I see in my 4K video ... as you can see the evening is splendid the distance we are around 1-2 miles from the head of the airport runway ... altitude around 600-800 feet. .. and the track is now shrouded in darkness ... does it seem clear? I would be curious to see something similar with his PC because if there are differences it means that it is a parametrization problem. My machine is equipped with an NVidia 1080 and a 4K video.

[Thorsten]

it seems clear to me ... anyway the 1K movie I sent you is clear or better dark ...

I don't do things by eye-balling - I gather data and solve equations.

There's the famous story how they asked policemen in a car simulation to adjust the parameters till it 'felt right' - the simulated cars could accelerate from 0 to 100 in a second in the end. This as a warning - 'gut feeling' and 'eyeballing' are very poor ways of setting up a simulation.

So, in your video I see all lights consistently pretty muted and a relatively close horizon, from which I'd conclude a poor visibility - and I can generally make out features on the ground quite well, from which I conclude that it's not really that dark, there's plenty of residual light.

Whereas in the real video, the ground is really dark and I can't recognize any features.

(Actually, it's not been really dark in FG nights since the faulty automatic moon phase implementation always illuminated the terrain regardless of whether the Moon is in the sky or not)

"There are no deaf worse than those who don't want to hear" ..

Indeed.

This is a detail of what I see in my 4K video ... as you can see the evening is splendid the distance we are around 1-2 miles from the head of the airport runway ..

Yeah, you're like looking directly into the setting sun (about 3.8 x 10^26 Watt of energy output) and expect to see a 20 Watt lightbulb from 2 miles away against that contrast? Seriously - do the numbers... In reality you see jackshit under these conditions, FG shows you far too much since we can't really create blinding output on your screen.

[Johan G]

Slightly off topic.

There's the famous story how they asked policemen in a car simulation to adjust the parameters till it 'felt right' - the simulated cars could accelerate from 0 to 100 in a second in the end. This as a warning - 'gut feeling' and 'eyeballing' are very poor ways of setting up a simulation.

There is also the famous F-22 crash caused by pilot induced oscillation (PIO). When Lockheed Martin developed the flight control laws, the various gains for various flight phases and tasks were tuned by experienced test pilots in a simulator. After the crash and the following review they realized that they had tuned the gains on a fixed base simulator. This resulted in that during flight, when there was not only visual cues but also motion cues, the gains would be too high for the now larger pilot inputs. This in turn caused the pilot to continuously overcompensate, which would cause the PIO. There were of course also some other aspects, IIRC some slightly abrupt gain changes when the gear went up and the control law changed, but the "eyeballed" gains on a fixed base simulator was a big part.

[abassign]

Returning to the subject and to the problems I detected ... I show you a night sequence with the setting of the scenario at maximum visibility and a 2K video (intermediate from that of 1K in the youtube movie and 4K of the previous screen-shot). As you can see something is not working, if you don't see it there is always the story of the 3 monkeys (I don't see, I don't hear, I don't speak) that you could cheerfully comment on, but it doesn't solve the problem of course.

To avoid doubts I put the maximum visibility, but if someone "has eyes to see", looking at the images he will always observe with magnificent quality the images of the various airport buildings and the red lights placed on the buildings next to the runway. Lights that evidently are not in the visibility rules used for the runway.



By day the "fires lit" produced by the PAPI dominate the scene, I had already pointed out that they seem like bright suns at dawn or at sunset rather than simple lights normally visible only on the axis within a certain angle. If anyone remembers the PAPI of the LOWI scenario they were quite another thing ... but how long does it take to insert PAPI models that make sense, we're in 2019 ...!



In this frame it seems to me all ok, not enough shining, but if I am only a few able to see what is not seen ... I am not surprised that this image is magnificent.





Here now I start to worry, I am looking in the rear direction and I observe that the lights are turning rather quickly ... from memory I would say too quickly ... and here nobody can tell me that it is an effect of the fog ... since the visibility is at its maximum and everything else that is not light IS VERY WELL ... therefore it is an effect of an obvious error in the management of this effect. Obviously nobody is perfect, even less who thinks he is perfect! You seem to see a pair of scissors at work cutting the lights away ... how sad.





The scissors continue to cut the landing strip lights and those of the airport that are managed by the magnificent and perfect algorithm ... until they get their complete disappearance .. while all the other lights, correctly, remain beautiful and bright with a obvious reduction with distance.





Therefore it seems obvious to me, gentlemen of the court, that the algorithm I have shown is guilty of violating the rules of optics and physics and those who do not see all this are guilty of perjury!

[abassign]

Using the UFO it is possible to better understand the problem. In the three images with the UFO I am positioned in front of the run way. If the FOV has a high angle the track lights are visible even from afar .. but if we zoom the lights are less and less visible ... until they disappear completely. In reality the lights are point-like, but they also have a halo (due to the lights on the track side due to the reflection of the ground, for the headlights to the parabola) that propagates and remains visible compared to the point light even at high distances. If the point lights are highly directional, the diffusion lights are much less so, and it is these that are seen and make the scene very different and more natural. So the solution is simple ... add to a point light with a rather narrow angle, even a diffused light that is a function of the type of lamp and at this point everything becomes closer to reality.

[abassign]

I made a standard flight, an overflight over the runway (LIME 28) and then a veer with landing. In this image the track is dark, in fact the point lights have a dead angle and here they are all practically invisible.



As we move away, the angle approaches the visible one and the light begins to show its effect.



After the veer the track is quite visible since the angle is correct.





During the approach it is observed that the central lights begin to be visible as the angle is reduced (it is an effect similar to that of the PAPI). In reality the central lights are always visible because they make a very marked halo, being half-set in the concrete.

Thus showing all the lights ... but this does not happen in reality ... since those white lights have a very strong diffuse component. The point part is mainly present during the day if those lights are on. In fact, the point lights are visible during the day while they are diffused at night along with the point lights.











As you can see from the lights of the plane (G91R1B) the lights are always double, that point, which represent the lamps and the diffused ones that disappear by day.

[Thorsten]

You really need to read the explanations given to you.

Seriously.

If you get told it's a multi-parameter problem and you modify a single parameter - then you don't see the full picture. Not on 10 screenshots, not on 100, not on 20 movies.

Here's KTTS, GPS/route manager shows 15 miles distance, map shows you the relative orientation, it's a dark, clear night - and...



... we have a nice visual on the runway lights. Under the same conditions, I see the runway outline from 100.000 ft altitude right above the runway.

So - working fine as far as I'm concerned - maybe the issue is (once again and again) with your settings?

Just read through what I've written, try to understand it, check your settings thoroughly - and you'll figure it out eventually

[www2]

@Thorsten

I think a part of this problem is that we use a low Dynamic Range camera in Flight Gear.[1][2]
And i think a other part of this is that Flight Gear use a display referred workflow in state of a scene referred workflow.[1][2]
And we are not the only one that have fall for this trap.

[1] https://www.youtube.com/watch?v=m9AT7H4GGrA
[2] https://www.blenderguru.com/tutorials/s ... otorealism