Board index FlightGear Development Effects and shaders

The Compositor

An exciting "new" option in FlightGear, that includes reflections, lightmaps, the particle system etc.. A lot is yet to be discovered/implemented!

Re: The Compositor

Postby abassign » Thu Jul 23, 2020 12:38 pm

abassign wrote in Mon Jun 22, 2020 8:34 am:Hi, I have a problem with the compositor (the version of FGFS is 2020.3.0). In both the 2020.3.0 and the 2020.2 versions, I noticed that the transparent surfaces have an anomalous behavior as a function of the angle of incidence of the light.


The problem was solved by modifying the effect using "Effects / glass" For example I built the effect with the following code:

In the property list:

Code: Select all
 <effect>
        <inherits-from>Aircraft/G91-R1B_HD/Models/Effects/exterior/exterior-glass</inherits-from>
        <object-name>cover (Meshed)</object-name>
        ...
 </effect>


And in "Models/Effects" I put this code in the file: exterior-glass.eff

Code: Select all
<?xml version="1.0" encoding="utf-8"?>
<PropertyList>
    <name>exterior-glass</name>
    <inherits-from>Effects/glass</inherits-from>
</PropertyList>
abassign
 
Posts: 844
Joined: Mon Feb 27, 2012 5:09 pm
Location: Italy (living 5 Km from airport LIME)
Callsign: I-BASSY
Version: 2018.3
OS: Linux Mint 19. x

Two or more spotlights always converge in same point

Postby abassign » Sat Jul 25, 2020 4:16 pm

In the tests on the compositor lights that I am carrying out for the interior of the cabin of the G91R1B plane, I need to have various spot-lights to reproduce the real scene (2 + 6 red spot lights + 2 UV spots). A single spot light functions correctly, but two or more spot lights always use the 6 coordinates of the first defined spot light, even if, viewing the geometry, it is observed that both the
positions and the directions are those required. I would like someone to solve this problem in order to better test the lighting effects.

I deduce that the problem is at the level of data transfer at the OpenGL level, or currently the spotlight vector has not been defined.
abassign
 
Posts: 844
Joined: Mon Feb 27, 2012 5:09 pm
Location: Italy (living 5 Km from airport LIME)
Callsign: I-BASSY
Version: 2018.3
OS: Linux Mint 19. x

Re: The Compositor

Postby abassign » Fri Oct 16, 2020 7:05 am

Compositor will certainly be wonderful, but I do not see great evolutions on its development, the most problematic thing is the lack of spot lights, as already reported a few months ago, but none of the developers in this forum have clarified the current problem that prevents the use of multiple lights, in fact the second spot light uses the same orientation as the first that has been defined. It would be helpful for someone to explain if this is a known problem or incorrect information on the Wiki.

For the shadows the result is not bad, even if I ask myself the terrible question: in terms of performance I do not observe substantial differences between Compositor and Rembrandt, in fact the only real advantage is that you can have shadows (for the lights I would say that still we are far) as in Rembrandt and the ALS functionality that Rembrandt did not allow.

I hope no one is offended by what I am writing, but I think it is important that things be said as they are and not as our dreams would like them to be. Obviously, as a developer of the G91R1B, I look forward to the future developments of Compositor, on which I place many hopes.

I would also like to know is it possible to use Compositor for objects on the ground, for example for lighting the interiors of hangars or lights ... ?

An idea, maybe there is but I don't know, it is possible to replace the Sun with the Moon, I mean ... At night having the luminous effect of the sun, which brings back the shadows, but with the ashen light of the Moon, it would be fantastic . Moreover, the light of the Moon is visible only with the sun well below the horizon, at least -15 degrees, so the management algorithms would not interfere with each other.
abassign
 
Posts: 844
Joined: Mon Feb 27, 2012 5:09 pm
Location: Italy (living 5 Km from airport LIME)
Callsign: I-BASSY
Version: 2018.3
OS: Linux Mint 19. x

Re: The Compositor

Postby Icecode GL » Sat Oct 17, 2020 3:05 pm

[...] but none of the developers in this forum have clarified [...]


I haven't been paying much attention to the forum lately, I'm sorry. There are just too many communication channels to keep track of and I honestly prefer to spend my time doing something else.


Most of the issues you mentioned are already fixed/added upstream, although they might take a while to reach to most people as the current development branch is expected to be kind of a mess. The only thing I can tell you is that you'll know when the Compositor is ready to use, no need to get impatient about it. There are plenty of other development projects in FG to get hyped about, feel free to ignore the Compositor for a while.
Icecode GL
 
Posts: 633
Joined: Thu Aug 12, 2010 12:17 pm
Location: Spain
Callsign: icecode
Version: GIT
OS: Fedora

Re: The Compositor

Postby Hooray » Sat Oct 17, 2020 5:36 pm

abassign wrote:I do not see great evolutions on its development


The compositor is not even merged/integrated as the default build/startup option currently, and as far as I remember, there are still merge requests pending.
Still, everybody on the devel list seems to agree that the Compositor is going to become the foundation for all future renderers/pipelines in FlightGear.
So I am not sure what the fuss is all about ?

James stated clearly that there are "blocking" issues that need to be resolved first, specifically see:

https://sourceforge.net/p/flightgear/ma ... /37126793/
James Turner wrote:Looking forward to trying this, although it’s going to be a slightly long road; Compositor doesn’t work for me on macOS yet, and OSG 3.6 has the text display issue, so we’re blocked on 3.4.2 for binary builds at the moment.


If you'd like to help pave the way for the compositor, feel free to get involved in addressing/troubleshooting remaining issues (namely Mac OSX and osgText issues relating to OSG 3.6+):

http://wiki.flightgear.org/Post_FlightG ... ges#Canvas

Finally, do note that these items are added to the roadmap AFTER the ongoing LTS version is released - thus, Compositor/ComposieViewer support might only become relevant in 2021 (and possibly later, depending on how many people get involved in troubleshooting remaining issues)
Please don't send support requests by PM, instead post your questions on the forum so that all users can contribute and benefit
Thanks & all the best,
Hooray
Help write next month's newsletter !
pui2canvas | MapStructure | Canvas Development | Programming resources
Hooray
 
Posts: 11979
Joined: Tue Mar 25, 2008 8:40 am

Re: The Compositor

Postby abassign » Sun Oct 18, 2020 5:41 am

Icecode GL wrote in Sat Oct 17, 2020 3:05 pm:
I haven't been paying much attention to the forum lately, I'm sorry. There are just too many communication channels to keep track of and I honestly prefer to spend my time doing something else.


I thank you for your response as I thank Hooray for a clearer understanding of the state of development of Compositor, in which we who develop aircraft have a lot of hope.
Actually the G91R1B has lost compatibility with Rembrandt in order to better manage the innovations introduced with Compositor and Canvas (which we use for the complete management of liveries also in MP).
Good job guys!
abassign
 
Posts: 844
Joined: Mon Feb 27, 2012 5:09 pm
Location: Italy (living 5 Km from airport LIME)
Callsign: I-BASSY
Version: 2018.3
OS: Linux Mint 19. x

Previous

Return to Effects and shaders

Who is online

Users browsing this forum: No registered users and 1 guest