Board index FlightGear Development Scenery

The landing strip lights are barely visible ...

Questions and discussion about enhancing and populating the FlightGear world.

Re: The landing strip lights are barely visible ...

Postby GinGin » Thu Sep 05, 2019 9:46 pm

Hey Abassign, did you deactivate the point sprites for runway lights?

In many of your screens, it seems activated.
Difference is quite huge, and whitout it, it looks realistic



With Sprites on:


Image

Image


Without it :

Image

Image
GinGin
 
Posts: 1580
Joined: Wed Jul 05, 2017 11:41 am
Location: Paris
Callsign: Gingin

Re: The landing strip lights are barely visible ...

Postby Thorsten » Fri Sep 06, 2019 6:20 am

I think a part of this problem is that we use a low Dynamic Range camera in Flight Gear.


No, we do not - the ALS shaders do something mathematically equivalent to rendering into a HDR buffer and computing pixel color from there - see here for a writeup - it doesn't do everythinga HDR post-processing stage could do, but it definitely does enough to handle all lighting conditions you would typically encounter correctly.

(I actually don't know how often I've had to explain this point... Rendering is just math, in math there's usually more than one way to skin a cat, if you don't solve quantum problems by wave functions you can solve them by matrix mechanics - you get the same solution - and it would be silly for people to insist it has to be done with wave functions and we have a problem if we use 'just' matrix mechanics or path integrals).
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: The landing strip lights are barely visible ...

Postby www2 » Fri Sep 06, 2019 10:44 am

@Thorsten
Thanks for the cleaners.
I think that the current constants be use are the problem not the math (with some exactions that virtual all graphic programs have[1] but this not mean that FG have this problem :mrgreen: ).
And i say that we switch to the values that blender use with there Filmic Blender color profile.[2]
And the last thing i say is that ALS is a display referred workflow in state of a scene referred workflow.[3][4]
I can be wrong that ALS use display referred, that in reality ALS use a scene referred workflow.

[1] https://www.youtube.com/watch?v=LKnqECcg6Gw
[2] https://sobotka.github.io/filmic-blender/
[3] http://www.xdcam-user.com/2013/10/under ... eferenced/
[4] https://ninedegreesbelow.com/photograph ... erred.html
www2
 
Posts: 319
Joined: Thu Apr 16, 2009 2:58 pm
OS: Ubuntu

Re: The landing strip lights are barely visible ...

Postby abassign » Fri Sep 06, 2019 2:03 pm

Thorsten wrote in Fri Sep 06, 2019 6:20 am:I actually don't know how often I've had to explain this point... Rendering is just math, in math there's usually more than one way to skin a cat, if you don't solve quantum problems by wave functions you can solve them by matrix mechanics - you get the same solution - and it would be silly for people to insist it has to be done with wave functions and we have a problem if we use 'just' matrix mechanics or path integrals).


It seems to me that the problem I have set is much simpler and is based banally on a couple of points that already long ago posed for the ALS emissive lights!

NOTE: Th ... If you look at what I've shown, if someone doesn't believe it or wants to have more parametric information, just take the UFO and take a walk at any airport at night. I think that any of you can do this and lose a few minutes of your precious time to understand the problem I reported.

Some (Th) say that there are so many parameters etc ... but instead of complicating things to get nothing, I think it is better to reflect on this specific problem, they let lose the brightness reduction with distance or residual brightness as components that do not they are present (if not in small part) in the images and videos I sent.

The problem seems clear to me that it is geometric in the management of point sprite.
The point sprites are constructed as "brightness" as a function of the angle of view, I think a rule has been constructed according to the observer point and the point where the emissive element is placed, the angle that is formed between the wo points and their references (for example a lateral runway light has an angle of visibility depending on the horizontal plane), a PAPI has a certain inclination etc ... If the observer places himself outside this angle, here he sees that the brightness until it disappears completely (for example above the runway). In my opinion this way of proceeding is correct for the spot light component, but a light emitter for lighting various parts of the runway is not a laser, but a device composed of two parts: a spot emitter and a screen and / or reflector and an area of land that is indirectly illuminated. These components do not seem to be present in the current "point sprite" simulation and therefore perfectly justify the behavior I observe and the effect of darkness that is obtained by observing the runway by placing itself on the vertical. In the non-point sprite mode the track is visible with greater brightness, but it is very unnatural. I deduce that the solution is to couple to the single spot lights disks with more diffused emission (depending on the angle of view only for the shaded areas, for example a directional light has an illuminated frontal area, but it is dark in the rear part ) that allow for more extended lighting as in reality. Personally I use this technique for the rear and side position lights and the G91R1B beacom and it seems to me that the results are good, certainly better than others that I see in the majority of FGFS planes.
I remember that in the old scenario of LOWI the airport lights were beautifully configured and gave the runway a remarkable aspect at night (I also remember the day as the path is often visible as only spot lights even in conditions of high ambient light, according to the conditions of environmental brightness and the choices of the control tower). I remember that the LOWI track lights were positioned, as in reality on posts, etc. I think that by applying a technique similar to that of 3D scenarios it is possible to extend this visual fidelity to all airports depending obviously on the level of detail we are willing to accept.

LOWI old scenary: https://gitorious.org/fgfs-custom-scenery/custom-scenery-overlay/
I remember that now this scenario does not work well and the integration with the old objects, which were many and well made, was not done with the system scenario ... I would say a real shame.

Image

Image

Image

Best regards
Developer of the program https://wiki.flightgear.org/Julia_photoscenery_generator
FDM developer of the G91R1B aircraft https://wiki.flightgear.org/FIAT_G91R1B
JSBSim collaborator
abassign
 
Posts: 947
Joined: Mon Feb 27, 2012 6:09 pm
Location: Italy (living 5 Km from airport LIME)
Callsign: I-BASSY
Version: 2020.4
OS: Ubuntu 20.10

Re: The landing strip lights are barely visible ...

Postby Thorsten » Fri Sep 06, 2019 4:37 pm

The point sprites are constructed as "brightness" as a function of the angle of view,


No, the runway edge lights actually are not.

These components do not seem to be present in the current "point sprite" simulation and therefore perfectly justify the behavior I observe and the effect of darkness that is obtained by observing the runway by placing itself on the vertical.


Since no one apart from you seems to see this, and since it's not how these lights are coded, it stands to reason that there's something wrong on your end.

You're (as often) refusing to listen to what others say and instead go off on your own trajectory.
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: The landing strip lights are barely visible ...

Postby abassign » Fri Sep 06, 2019 8:04 pm

Th ... Since you see things we have never seen in the night before giving others incompetent poor visionaries, you can flay over one airport and show us what you see? If you show correctly lit airports when you fly on their vertical lines, you can tell us where we are wrong in the parameters.

Thanks for collaboration
Developer of the program https://wiki.flightgear.org/Julia_photoscenery_generator
FDM developer of the G91R1B aircraft https://wiki.flightgear.org/FIAT_G91R1B
JSBSim collaborator
abassign
 
Posts: 947
Joined: Mon Feb 27, 2012 6:09 pm
Location: Italy (living 5 Km from airport LIME)
Callsign: I-BASSY
Version: 2020.4
OS: Ubuntu 20.10

Re: The landing strip lights are barely visible ...

Postby Thorsten » Fri Sep 06, 2019 8:14 pm

you can flay over one airport and show us what you see?


I'd rather not, because that would be really cruel. And I told you earlier what I see - I see the airport outline fine from 100.000 ft.

Thanks for collaboration


You're misunderstanding something here - I have no wish at all to collaborate with you in any way. You've talked down on my work many times, you've attacked me personally when I did my job and investigated your merge request when a (false) claim was made that some stuff was stolen - at some point I decided to return your attitude to you.

I am honor-bound to explain some things here, and I will quickly check stuff you claim to be broken, but that's all you will ever get from me with your attitude of 'I know better than the people who coded it how stuff works and I don't need to listen to anyone'.
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: The landing strip lights are barely visible ...

Postby abassign » Fri Sep 06, 2019 10:52 pm

Dear Th ...

Don't be a victim because on that occasion you acted like the executioner, if I didn't mention that ugly and ridiculous story it's because I firmly believe the need to forgive the one who made a mistake because making mistakes is part of human nature and therefore it is always necessary to try to forgive those who hurt you.

Let us return to my question correctly, that is to say that the intensity of the lights placed on the airport landing strips do not work properly.
If you are so good, as you say and you are not a bluf, I think that in 5 minutes you can solve the problem. If you do not feel interested in solving the problem or you are not able to, I assure you that my life will not be upset and even the lives of all FGFS users will be upset, the reason is that I am sure someone will solve the problem and will become FGFS still a better program.

Remember a little advice that working on this project must be a pleasure and not a source of sadness or anger, fly happy with FGFS!

Have fun
Developer of the program https://wiki.flightgear.org/Julia_photoscenery_generator
FDM developer of the G91R1B aircraft https://wiki.flightgear.org/FIAT_G91R1B
JSBSim collaborator
abassign
 
Posts: 947
Joined: Mon Feb 27, 2012 6:09 pm
Location: Italy (living 5 Km from airport LIME)
Callsign: I-BASSY
Version: 2020.4
OS: Ubuntu 20.10

Re: The landing strip lights are barely visible ...

Postby Thorsten » Sat Sep 07, 2019 6:14 am

I think that in 5 minutes you can solve the problem.


Yes, but you also think you understand it really really thoroughly

It seems to me that the problem I have set is much simpler and is based banally on a couple of points that already long ago posed for the ALS emissive lights!


have done so for a long time and that I really have no clue - so I think you ought to be able solve it in much less than that time yourself. In fact, you've posed the solution long ago :D

You can't have it both ways - be the only knowledgeable person in town when it's about what needs to be done, but then plead no knowledge when something needs to be done.

Let us return to my question correctly


Sorry - I'm done with your question - I thought I made that clear. I've explained to you how I originally coded it and why I did so - that's all you'll get from me.
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: The landing strip lights are barely visible ...

Postby legoboyvdlp » Sat Sep 07, 2019 1:49 pm

Image

This looks acceptable to me, especially with reasonably low visibility (12 miles). However when I zoom in:

Image

Image

I do not understand why the apparent size of the point sprites reduces so much making the lights so hard to see at this point when I zoom in. I would be happy to accept any explanation but to my untrained eyes it certainly does not look right!

Image

Image

Also, in tower view AGL the lights are far dimmer than in other views (including tower view itself). I think that may be something introduced by tower view AGL.

Image
My settings and weather. Reasonable visibility, dark night, no smog, smooth fog, and finally maximum possible visibility.

Image
And this shot looks fantastic thanks to the procedural lights :D

I wonder if you could tell me if the first three shots look right to you. Certainly with normal zoom it looks fine - as bright as what I expect, bit the only concerning thing is the sprites getting so tiny as I zoom in. I would maybe expect them to be so small in the third shot (only a very few bulbs in view) but surely you should be able to see some lights a little bit brighter in the second screenshot.

You may wish to view the fourth one in a dark room... I can only make out the PAPI lights myself in that.
User avatar
legoboyvdlp
 
Posts: 7981
Joined: Sat Jul 26, 2014 2:28 am
Location: Northern Ireland
Callsign: G-LEGO
Version: next
OS: Windows 10 HP

Re: The landing strip lights are barely visible ...

Postby Thorsten » Sat Sep 07, 2019 2:33 pm

I do not understand why the apparent size of the point sprites reduces so much


I don't think it actually does, they're just not scaled up with the zoom factor, so while they stay the same everything else grows quite a bit - fundamentally they don't work based on real geometry - which is why they behave different from the aircraft-side implemented lights (which are real geometry).

The issue with a light is that usually the disc size should not equal the light emitter size but the brightness given the rest of the scene - which precludes using real emitter geometry (which would be tiny at large distances, you could not resolve star discs with bare eyes, but you can still see their light for instance) - but also if you would scale the disc size seen from afar when zooming in or getting closer, you'd have 10 m sized lights on the ground, which also is a problem.

So the problem of getting you a reasonable perception of light brightness as a function of distance is not straightforward - you'd want the spot size to grow when you zoom in, but somehow not really with the rest of the scene.

(Admittedly zooming behavior was not on my priority list when coding them, not sure whether the zoom factor is taken into account at all).
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: The landing strip lights are barely visible ...

Postby www2 » Sat Sep 07, 2019 4:16 pm

Thorsten.
What are the values are use for runway/taxiway/airport lighting?

I have found some Real world values:
https://www.genesislamp.com/lampsbulbs.html
www2
 
Posts: 319
Joined: Thu Apr 16, 2009 2:58 pm
OS: Ubuntu

Re: The landing strip lights are barely visible ...

Postby Thorsten » Sat Sep 07, 2019 4:26 pm

You have a somewhat exaggerated idea of what kinds of numbers I can remember across a few years it seems...

I can tell you that I found a similar website (not that one), looked up the wattage of various bulbs, corrected for the actual percentage of that emitted as light (which is not overly high for incandescent bulbs), converted that into apparent stellar magnitude as a function of distance and cut where an eye adapted to a lit instrument panel or the ambient scene light (whatever is lower) would no longer see (i.e. you can't see all the way down to mag 6 as with fully darkness adapted eyes). That's the theoretical max. distance you can still see a light.

On top of that, atmosphere-driven attenuation is considered, so the distance is normally lower than that.

If you're interested in the problem, I highly recommend you do the calculation yourself, I found it pretty instructive.
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: The landing strip lights are barely visible ...

Postby V12 » Sun Sep 08, 2019 5:15 am

Original forum topic is "landing stripes barely visible". I have opposite problem - lights are too bright, distance attenuation doesn't work :

Image

It is EDDF from 52k feets and 15 nm away in very good visibility and high atmospheric pressure.

Similar screenshot 17 nm west of EDDF, altitude 8500 feets :

Image
Fly high, fly fast - fly Concorde !
V12
 
Posts: 2757
Joined: Thu Jan 12, 2017 5:27 pm
Location: LZIB
Callsign: BAWV12

Re: The landing strip lights are barely visible ...

Postby GinGin » Sun Sep 08, 2019 8:50 am

Tick « use point sprites for’runway Lights «  in rendering options
GinGin
 
Posts: 1580
Joined: Wed Jul 05, 2017 11:41 am
Location: Paris
Callsign: Gingin

PreviousNext

Return to Scenery

Who is online

Users browsing this forum: No registered users and 9 guests