Board index FlightGear Development Effects and shaders

Landmass problem

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

Re: Landmass problem

Postby enrogue » Thu Aug 17, 2017 7:21 pm

@Thorsten

thanks for the pointer - I'm reading through your GLSL primer to see if I can understand a bit more (I freely admit I know nothing about GLSL!)
User avatar
enrogue
 
Posts: 63
Joined: Mon May 19, 2014 6:40 pm
Location: London (UK)
Callsign: enrogue
OS: OS X, Ubuntu

SCNR ...

Postby Hooray » Thu Aug 17, 2017 8:19 pm

abassign wrote in Thu Aug 17, 2017 7:15 pm:Strange however that with X-Plane 10 this problem I have not noticed ... maybe I see little and I'm aging ... :(

Now that is very interesting, could you do us a favor and also check if the problem persists in Microsoft FSX when you disconnect your yoke ?
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: 11007
Joined: Tue Mar 25, 2008 8:40 am

Re: Landmass problem

Postby wkitty42 » Thu Aug 17, 2017 8:31 pm

ROTFL!!
"You get more air close to the ground," said Angalo. "I read that in a book. You get lots of air low down, and not much when you go up."
"Why not?" said Gurder.
"Dunno. It's frightened of heights, I guess."
User avatar
wkitty42
 
Posts: 3768
Joined: Fri Feb 20, 2015 3:46 pm
Location: central NC, USA
Callsign: wk42
Version: git next
OS: Kubuntu 14.04.5

Re: Landmass problem

Postby Thorsten » Fri Aug 18, 2017 5:17 am

Strange however that with X-Plane 10 this problem I have not noticed ... maybe I see little and I'm aging ...


You're definitely on to something here - I don't see it under KDE Minesweeper either!
Thorsten
 
Posts: 9164
Joined: Mon Nov 02, 2009 8:33 am

Re: Landmass problem

Postby abassign » Fri Aug 18, 2017 7:25 pm

wkitty42 wrote in Thu Aug 17, 2017 8:31 pm:ROTFL!!


Image

After all these inconclusive discussions ... it's best to take it on laughing

P.S:
But can anyone test on an Intel I7 machine with Windows without an external GPU?
abassign
 
Posts: 672
Joined: Mon Feb 27, 2012 5:09 pm
Location: Bergamo - Italy (living 5 Km from airport LIME)
Callsign: I-BASSY
Version: 2016.3
OS: Linux Mint 17. x

Re: SCNR ...

Postby abassign » Fri Aug 18, 2017 7:30 pm

Hooray wrote in Thu Aug 17, 2017 8:19 pm:Now that is very interesting, could you do us a favor and also check if the problem persists in Microsoft FSX when you disconnect your yoke ?


You're right, but it is impossible for me because I have no machine with Windows!
However, I would like someone to try, at least to understand and communicate the operating limits of the ALS.
abassign
 
Posts: 672
Joined: Mon Feb 27, 2012 5:09 pm
Location: Bergamo - Italy (living 5 Km from airport LIME)
Callsign: I-BASSY
Version: 2016.3
OS: Linux Mint 17. x

Re: Landmass problem

Postby Hooray » Fri Aug 18, 2017 7:42 pm

abassign wrote in Fri Aug 18, 2017 7:30 pm:I would like someone to try, at least to understand and communicate the operating limits of the ALS.


Again, by definition, there is nothing platform or OS specific when it comes to OpenGL/GLSL shaders -thus, even if the ALS shaders were entirely broken or incompatible, it would not just show up on a single platform, unless the corresponding drivers involved are somehow flawed, too.

For what it's worth, I don't think Thorsten was condescending at all, but even with my own -rather limited- background knowledge on OpenGL and GLSL, I realize that your line of thinking is flawed, and I think you would agree with this assertion if you were to read up on OpenGL, GLSL and GPUs - even if that just means spending a few hours on youtube to learn more about the technologies and terms involved; you would probably quickly realize that there really is a logic flaw involved here, and I am not just talking about the GPU/driver here, but about logic on your end ;-)

None of this is to say that ALS is perfect or even just bug-free, but troubleshooting this properly, and applying some common sense would get us much farther than what we've seen here in the last couple of days (and in fact, I am a little surprised seeing that this hasn't yet been locked by any of the mods).

I really suggest that you look at OpenGL/GLSL troubleshooting in a standalone fashion, outside the fgfs environment - furthermore, I also suggest that you take into account the impact that legacy GL code may have on more modern GL/GLSL code.

I also believe that if you can come up with a sound hypothesis and a working "workaround" that Thorsten would be willing to get that committed.
Note that effects actually can be told to be platform/back-end specific using so called "predicates".

Finally, I also believe that many of these issues could be fixed if we had a motvated contributor involved with access to Intel GPU hardware, willing and able to regularly test these features - which is a problem not just specific to ALS. but also one that shows up in the Rembrandt/Canvas departments, too.

Thus, if you have access to such hardware and if you are able/willing to read up on OpenGL/GLSL internals to make heads and tails of the effects/shaders involved here, you could provide a great service to this community - and all that may involve for starters would be watching a few OpenGL/GLSL introductions on youtube and installing the corresponding programming tools to dissect shaders and tinker with them.

Even just knowing that these issues are indeed specific to FlightGear would be good to know.
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: 11007
Joined: Tue Mar 25, 2008 8:40 am

Re: Landmass problem

Postby abassign » Sat Aug 19, 2017 3:58 pm

Hooray wrote in Fri Aug 18, 2017 7:42 pm:Thus, if you have access to such hardware and if you are able/willing to read up on OpenGL/GLSL internals to make heads and tails of the effects/shaders involved here, you could provide a great service to this community - and all that may involve for starters would be watching a few OpenGL/GLSL introductions on youtube and installing the corresponding programming tools to dissect shaders and tinker with them.
Even just knowing that these issues are indeed specific to FlightGear would be good to know.


Thanks for your trust in me :)
Absolutely yes. Already in other times I have analyzed the problems I have encountered in FGFS and I always try to do an analysis that can serve others, including me, to understand problems and solve them.

However, I do not think the current road to using ALS to handle the effects is the best way. It seems to work in Assembler ... when a language like C ++, though slightly less efficient, allows better code management by a community of potential programmers. The OpenGL is very powerful, but hard to keep in mind the variety of platforms is really great. I think it's best to split the problem by exploiting, for the 3D part, a game engine (obviously with a compatible license ... unfortunately there are few, but something exists).

Thorsten is exceptional, but not very collaborative and this is reflected in the way in which to write the code. The experience of a programmer gains on the code battlefield, but if the battlefield is full of mine ... it becomes difficult to have some surviving programmer capable of continuing the job!

Even the Rembrandt experience should teach us that this is not the best way to operate. Rembrandt is an excellent rendering engine just waiting to be updated and improved, the effect of the fog you can get is really beautiful and the lights are great and give a deep sense of 3D realism. Of course the results with the static methods used for the Shuttle are exceptional and they always excite me, but they are complex to get and lack the dynamic that Rembrandt allows. I thought so many times that you could make Rembrandt work on two tasks, one to give the effects of slow-moving scene and the other for the effects that require speed, if you work with the vectors except with the bitmaps the thing could be feasible . But Rembrandt is difficult to access, has poor documentation that is useful only to anyone who has developed it ... Perhaps by making new documentation and cleansing his code something is possible on that line of development.

My idea is that I see few to develop on FGFS at this time is just because the access level is too high, the documentation too fragmentary and unmanaged (Often the code documentation occupies at least 30% -50% of the development time ) and that, as in the case which I have analyzed in this post, can lead to big problems.

However, we are trying to improve, to give new energy to the development of FGFS in order to avoid becoming a dried tree.
abassign
 
Posts: 672
Joined: Mon Feb 27, 2012 5:09 pm
Location: Bergamo - Italy (living 5 Km from airport LIME)
Callsign: I-BASSY
Version: 2016.3
OS: Linux Mint 17. x

Re: Landmass problem

Postby Hooray » Sat Aug 19, 2017 4:07 pm

Actually, effects are explicitly designed to better organize/reuse and structure GLSL shaders, using much higher level constructs than C++ code, which does need to be recompiled, whereas effects transparently map to GLSL shaders.

I'd suggest to read up on how effects and GLSL shaders relate to eachother, and how the property tree links these two concepts together - as well as do a little research to see for yourself, that FlightGear does indeed show rendering artifacts even with ALS/Rembrandt entirely disabled, which can often be linked back to certain legacy OpenGL code interfering badly with more modern OSG constructs.

Keep in mind that the original "OSG port" has never been completed entirely, so there really is tons of legacy GL code running in the main loop which is complicating matters for people wanting to use more modern OpenGL/GLSL code
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: 11007
Joined: Tue Mar 25, 2008 8:40 am

Re: Landmass problem

Postby Thorsten » Sat Aug 19, 2017 6:16 pm

Already in other times I have analyzed the problems I have encountered in FGFS and I always try to do an analysis that can serve others, including me, to understand problems and solve them.


Seriously?

If that comes out of the analysis

However, I do not think the current road to using ALS to handle the effects is the best way. It seems to work in Assembler ... when a language like C ++, though slightly less efficient, allows better code management by a community of potential programmers.


then I guess you still have a long road to go to analyze any problem... There's (as every so often) nothing correct in this sentence.

So, just maybe it'd be a good idea to learn anything about 3d rendering before telling everyone else how it should be done? You're like a passenger entering the cockpit of an airliner and explaining to the pilot how the plane really should be flown, pointing out that inflating the gas bags more will help a lot.

as in the case which I have analyzed in this post, can lead to big problems.


... which however are in the GPU driver, which isn't even part of FG. How did you manage to still miss that part???
Thorsten
 
Posts: 9164
Joined: Mon Nov 02, 2009 8:33 am

Previous

Return to Effects and shaders

Who is online

Users browsing this forum: No registered users and 1 guest