Board index FlightGear The FlightGear project

FlightGear 3.2 released

Questions about the FlightGear organisation, website, wiki etc.

Re: FlightGear 3.2 released

Postby Figaro » Sun Oct 26, 2014 11:20 pm

It's a shame not everyone can afford liquid nitrogen cooled i7-5960Xs and GTX980s in four-way SLI :/

One can only dream...
User avatar
Figaro
 
Posts: 1280
Joined: Fri Feb 25, 2011 9:23 pm
Callsign: 4L-FIG
OS: ElementaryOS/Win10

Re: FlightGear 3.2 released

Postby cossack90 » Mon Oct 27, 2014 5:52 am

I thought this was a simulator not a game...
Pilot of VOXY AIRLINES
WIP: Project Ukraine
User avatar
cossack90
 
Posts: 160
Joined: Mon Dec 27, 2010 10:30 pm
Location: Holland/ Ukraine
Callsign: UR-AOK / VOX0001
Version: 2018
OS: Windows 10

Re: FlightGear 3.2 released

Postby Figaro » Tue Oct 28, 2014 12:39 am

Who said it was a game?

Anyone? Anyone, at all? :/
User avatar
Figaro
 
Posts: 1280
Joined: Fri Feb 25, 2011 9:23 pm
Callsign: 4L-FIG
OS: ElementaryOS/Win10

Re: FlightGear 3.2 released

Postby cossack90 » Tue Oct 28, 2014 2:19 pm

I did. Mine opinion is that everyone need to enjoy flying without bothered about specs. If not than its a game not a simulator.
Pilot of VOXY AIRLINES
WIP: Project Ukraine
User avatar
cossack90
 
Posts: 160
Joined: Mon Dec 27, 2010 10:30 pm
Location: Holland/ Ukraine
Callsign: UR-AOK / VOX0001
Version: 2018
OS: Windows 10

Re: FlightGear 3.2 released

Postby Thorsten » Tue Oct 28, 2014 2:41 pm

Well, we keep having these discussions. The great thing about the vast majority of new features is that they're optional. Which means you can switch them off, and FG 3.2 runs quite as fast (if not faster) than 1.9.1 if you accept the same visuals and the old scenery.

But defaults continue to grow as system performance grows, so if your system doesn't, you eventually need to scale back from defaults. The expectation that you get more of everything with every release without any performance drain is unrealistic.

And maxing out all quality sliders and using all available options on a system that's not a gaming rig is just self-inflicted pain.

So, if you want to see all the latest visual goodies, bring appropriate hardware. If you just want to enjoy simulated flight, scale back on options. It's as simple as that.

We support in fact both high-end and low-end systems. But we do not support high end visuals on low end systems. That can only be done by magic.
Thorsten
 
Posts: 11375
Joined: Mon Nov 02, 2009 8:33 am

Re: FlightGear 3.2 released

Postby Hooray » Tue Oct 28, 2014 2:56 pm

still, the people on lower-end/older hardware can still help us improve FlightGear by providing much more feedback, especially by using the issue tracker - obviously, scenarios like the 777-200 at KSFO are unlikely to ever work well on those systems, but flying the ufo over sea should still be possible with >= 30 fps - if not, something else is going on - and identifying that will be useful regardless of the horsepower available. We've seen issues like the effects/listener leak that went unnoticed for months, if not years, simply because most of us tend to update our systems either way - equally, we've seen PUI related performance issues reported without being in a good position to reproduce those - so there are merits to providing feedback, but please don't expect things to change over night - as long as we're getting enough feedback, we can look at things and try to troubleshoot issues that wouldn't show up on a typical developer's machine.

FlightGear isn't specifically targeting low-end systems admittedly - but that's mainly due to the fact that most of us don't have access to these kinds of systems, so we depend on the amount -and quality- of feedback provided by end-users.

FlightGear is obviously far from perfect, but many of its issues are not as prominent on average hardware as they are on less powerful systems unfortunately - which really means that we cannot easily fix things without having people involved willing to do the corresponding troubleshooting to provide actionable bug reports.
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: 11475
Joined: Tue Mar 25, 2008 8:40 am

Re: FlightGear 3.2 released

Postby cossack90 » Tue Oct 28, 2014 9:28 pm

Is there a way to get good quality graphics and high frame rates on low spec machines? What about other gfx engine (open source/ self made)? Would that create a less consumable flight simulator? So the CPU/ GPU can be low-spec?

Edit: And since we still get this type of discussion, maybe there is a point? Maybe we should take action?
Pilot of VOXY AIRLINES
WIP: Project Ukraine
User avatar
cossack90
 
Posts: 160
Joined: Mon Dec 27, 2010 10:30 pm
Location: Holland/ Ukraine
Callsign: UR-AOK / VOX0001
Version: 2018
OS: Windows 10

Re: FlightGear 3.2 released

Postby flyingfisch » Tue Oct 28, 2014 9:48 pm

@cossack90: Try talking to actual developers about this in the newsgroups. The forums don't have many actual FG core devs on it. ;)

I would argue that asking for somewhat high-quality gfx on a low-end computer should be possible. After all, MSFS2004 did it.
Beechcraft A35 Bonanza
Wiki Page
GitHub Repo
Improving the 737-300
Forum Thread
All my projects
User avatar
flyingfisch
 
Posts: 306
Joined: Mon Oct 07, 2013 1:11 pm
IRC name: ffisch
Version: 3.0
OS: Ubuntu 12.04

Re: FlightGear 3.2 released

Postby cossack90 » Tue Oct 28, 2014 10:42 pm

And how to contact them? Thx you see the point. Our planes look awesome and some fly even very realistic, but to enjoy the sim, surrounding environment and effects should be good to (on low spec systems). To inprove the simulation "feel".
Pilot of VOXY AIRLINES
WIP: Project Ukraine
User avatar
cossack90
 
Posts: 160
Joined: Mon Dec 27, 2010 10:30 pm
Location: Holland/ Ukraine
Callsign: UR-AOK / VOX0001
Version: 2018
OS: Windows 10

Re: FlightGear 3.2 released

Postby VoyR » Tue Oct 28, 2014 10:55 pm

"High" All
I experience a very weird bug (?) . I am sinking beneath the ground (AGL) when landing..no solid runways... just underworld... HUD shows crazy altitudes.
It doesn't happen always (after a long flight) in different airports, on several planes.
Just upgraded to 3.2
Anyone else?

Ubuntu 12.04
Nvidia GT430
8GB RAM

OK, I've got an answer= the problem is : simplified terrain =enabled

FG is great , just enjoy ;-)
Last edited by VoyR on Wed Oct 29, 2014 11:45 am, edited 3 times in total.
VoyR
 
Posts: 17
Joined: Wed Oct 30, 2013 11:22 am
Location: Barry Island and Danzig
Callsign: Voi-R9
Version: 3.6+2016
OS: Ubuntu 15.10

Re: FlightGear 3.2 released

Postby Thorsten » Wed Oct 29, 2014 7:24 am

I would argue that asking for somewhat high-quality gfx on a low-end computer should be possible. After all, MSFS2004 did it.


I would argue that even if you scale quality down quite a bit in FG, you still easily reach this quality for the terrain in FG:

Image
Image
Image

(result of a quick google search for Flight Simulator 2004 screenshots)

The problem is that I don't get the feeling people actually want that. I have awesome memories of Falcon 4.0 - in my memory, the terrain was quite realistic and I thoroughly enjoyed the experience. If I would contrast the actual screenshots with what I can render in FG these days, it'd look like crap. Because you still get to see all the high-end screenshots of FG posted everywhere. So, mentally you compare FG on your screen with FS 2004 in your memory and FG and screenshots, and naturally you wish for more.

Compared with what I can see from MS FS 2004 screenshots, FG at low rendering quality is actually *really* competitive - yet we're having this discussion.



Our planes look awesome and some fly even very realistic, but to enjoy the sim, surrounding environment and effects should be good to (on low spec systems).


Yes, I really like these 'should be' sentences. It's a bit like ordering a meal without specifying who pays the bill. There should be free lunches, but there just aren't.

If you want something to look good, you either throw computing power at it, or if you don't have, you have to invest a ton of work.

Case in point - you can make FG scenery look better by either using procedural texturing schemes which allow you things like overlay texturing, weather-dependent texturing etc. which help with creating a consistent visual impression - or you painstakingly spend a few hundred hours handcrafting and testing a carefully color-balanced texture scheme which renders acceptably under different lighting and fogging conditions.

In many cases, visual improvements do not require any infrastructure or faster computers but just better textures - and hundreds of hours of work to get them right. Yet - somehow, almost nobody seems to create these textures.

Of course someone *should* create them, no?

And since we still get this type of discussion, maybe there is a point? Maybe we should take action?


Yes, the point of them is that the average human being always thinks a free lunch is a good thing. It can't hurt to ask for something - maybe someone else does it?

I mean, you can't seriously be of the opinion that the FG developers are so mean or stupid that if there would be an easy way (i.e. requiring neither hundreds of hours on texturing work nor a high performance GPU) to dramatically improve visuals, we wouldn't do it? I spent perhaps 10 hours a week coding rendering stuff - do you honestly think I wouldn't be interested in getting it faster and better and you need to tell me that this is a good thing so that I know? Get your hands dirty, spend a week with an introduction to 3d real time rendering, look into what our renderig pipeline does and what it can't do easily - and then get back and make suggestions. That'd be useful and can be taken seriously. 'This should be possible on low performance systems' I can't even start to take seriously.

(I'm obviously not claiming that everything is perfect, there are things one could do by throwing another 1000 man-hours into the fray - but they have to come from somewhere).
Thorsten
 
Posts: 11375
Joined: Mon Nov 02, 2009 8:33 am

Re: FlightGear 3.2 released

Postby Hooray » Wed Oct 29, 2014 11:42 am

Actually, I have no doubt that there are a number of potential improvements to speed up performance and maybe even improve visuals here and there - the real issue is that there's dozens of areas in FlightGear, including a number of FG subsystems, separate projects/components - most of us are really only familiar with 2-3 areas - and usually, it has taken us years to arrive at where we are, no matter if that means aircraft/scenery development, scripting, core development or rendering related contributions.
Like Thorsten said, there are a number of contributor who regularly spend 10-15 hrs per week contributing to the project in various shapes and forms - including less-visible areas of the project, i.e. "behind the scenes".
So there's a learning process involved, too - there are a ton of areas where there's a fairly low barrier to entry (if any at all) involved - but usually, large-scale change like improving performance, memory footprint, visuals require years of experience and expertise - not to mention a number of skills/qualifications that the project cannot teach you effectively - such as having a strong background in maths, physics or aerodynamics - as well as being an experienced software developer familiar with different coding concepts and paradigms.
At the end of the day, there's a finite amount of time that we can spend contributing to the project - and usually, we are more interested in "doing" something rather than documenting what we know already - however it is exactly this which is helping newcomers to face a lower barrier to entry - shouldering all the work ahead on our own obviously doesn't scale too well - documenting what we have learned over the course of several years however, scales very well - even if we should happen not to be around here - whenever someone shares, and documents, the very knowledge -and poiners- the he/she accumulated, it means that future contributors will face a more graceful entry - not necessarily requiring years of involvement to make certain changes.
It is obviously easier to request and suggest changes than doing the legwork to at least help others in the long-term - but ultimately, you only get to change something by being the change you want to see - absent that, you need a lot of patience and need to find a way to "mentor" people to work on stuff that you're interested in - ideally, not specific to any particular effort or contributor, but in a generic -and timeless- fashion.
As long as people don't understand how these things are working, the project will continue to depend to "power contributors" who either have a corresponding track record and professional background - or it will continue to depend on growing all the expertise over time, which applies to largest share of long-term contributors, who have often been involved for 5-10 years to get to where they are now.
Flight simulation is a complex, and a huge, field - and FlightGear as a project doesn't do much to acknowledge this - which is the reason why certain changes are taking many years to materialize - the most dedicated long-term contributors have all spent the time required to become "domain experts".
The few occasions where some external contributor provided major changes to the project are obviously rare - and it doesn't help that the project doesn't accommodate new contributors accordingly, no matter the concrete effort (sports-model, radio propagation, osgEarth etc).
It is true that XPlane does perform better than FG on average hardware - but they get to spend more time/resources than we do - and while there are very clever folks involved in the FG project, large-scale changes require familiarity with overlapping components - which are skills that need to be acquired - a process that typically takes several years.
Equally, FSX/XP conclusions cannot be made "as is" because of different approaches and architectures - so some things cannot be easily translated.

The fact that some of our most active contributors also tend to participate in discussions like these on the forum, also isn't helping obviously - because such community interactions are also eating up time.

If you want to see a certain change in FG, be prepared to do some legwork first - even if you are never going to write a single line of code, there is so much other stuff that can be done - including participating in the forum/wiki, to help free up developers from having to provide end user support.
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: 11475
Joined: Tue Mar 25, 2008 8:40 am

Re: FlightGear 3.2 released

Postby cossack90 » Wed Oct 29, 2014 8:33 pm

Yes, the point of them is that the average human being always thinks a free lunch is a good thing. It can't hurt to ask for something - maybe someone else does it?

I mean, you can't seriously be of the opinion that the FG developers are so mean or stupid that if there would be an easy way (i.e. requiring neither hundreds of hours on texturing work nor a high performance GPU) to dramatically improve visuals, we wouldn't do it? I spent perhaps 10 hours a week coding rendering stuff - do you honestly think I wouldn't be interested in getting it faster and better and you need to tell me that this is a good thing so that I know? Get your hands dirty, spend a week with an introduction to 3d real time rendering, look into what our renderig pipeline does and what it can't do easily - and then get back and make suggestions. That'd be useful and can be taken seriously. 'This should be possible on low performance systems' I can't even start to take seriously.


No-one say: that the FG developers are so mean or stupid that if there would be an easy way. And i just suggest/ try to improve the development, if i could and have time i was already working as a core developer.
Pilot of VOXY AIRLINES
WIP: Project Ukraine
User avatar
cossack90
 
Posts: 160
Joined: Mon Dec 27, 2010 10:30 pm
Location: Holland/ Ukraine
Callsign: UR-AOK / VOX0001
Version: 2018
OS: Windows 10

Re: FlightGear 3.2 released

Postby cossack90 » Fri Oct 31, 2014 8:47 am

Note: Since I moved to linux (ubuntu) flightgear performs better, display better graphics and has faster load times. And since 3.2 update.... No better way to fly :)
Pilot of VOXY AIRLINES
WIP: Project Ukraine
User avatar
cossack90
 
Posts: 160
Joined: Mon Dec 27, 2010 10:30 pm
Location: Holland/ Ukraine
Callsign: UR-AOK / VOX0001
Version: 2018
OS: Windows 10

Re: FlightGear 3.2 released

Postby KL-666 » Sat Nov 01, 2014 3:53 pm

Fg still crashes sporadically, often without message. But this time i got one in the log, maybe it can help someone.

There were many of the first two tile loading messages, ending with the last 2 messages:

Code: Select all
terrain:3:..\..\simgear\simgear\scene\tgdb\ReaderWriterSTG.cxx:488:Loading tile 2990208.stg
terrain:3:..\..\simgear\simgear\scene\tgdb\ReaderWriterSTG.cxx:396:  Generating ocean tile: e000n40/e002n40/2990208
opengl:4:..\..\..\flightgear\src\Viewer\fg_os_osgviewer.cxx:193:Warning: deleting still referenced object 00000000674C17B0 of type 'class osg::Referenced * __ptr64'

opengl:4:..\..\..\flightgear\src\Viewer\fg_os_osgviewer.cxx:193:         the final reference count was 1, memory corruption possible.


I am on win 8, i7 8 threads, nvidia 670mx, ALS with detailed weather (no rembrandt) and threading model CullThreadPerCameraDrawThreadPerContext

Kind regards, Vincent
KL-666
 
Posts: 784
Joined: Sat Jan 19, 2013 1:32 pm

PreviousNext

Return to The FlightGear project

Who is online

Users browsing this forum: No registered users and 1 guest