Board index Other Hangar talk

Help needed - market research for FG

Talk about (almost) anything, as long as it is no serious FlightGear talk and does not fit in the other subforums.
Forum rules
Please refrain from discussing politics.

Re: Help needed - market research for FG

Postby i4dnf » Mon Dec 17, 2012 12:53 pm

Thorsten wrote in Mon Dec 17, 2012 11:49 am:
As for resolution, the default textures are I think usually 512x512 spread to 2000 x 2000 m, which gives a pixel size of 4x4 m. I'm not sure how large the dds base size is - probably similar. They're mapped to a variety of sizes, but 1000 x 1000 m seems the usual. Which gives a pixel size of 2 x 2 m. The pricetag is that cropgrass1.png is 617K , cropgrass.dds is 2.7 M, so the dds textures take ~4.5 times the diskspace (and download bandwidth) of their png counterparts (imagine how a widespread switch to dds would blow up the size of the GIT repository...)

The better pixel size comes at increased tiling artefacts, which the dds set counters by having very low contrast inside the texture so as to create no visual cues to tiling. However, contrast inside the texture is an important device to hide the hard landclass boundaries (see the long version of this argument here, so from higher altitude, the dds set to first oder looks like the mapserver image of the landclass distribution painted with almost monochromatic colors. If you fly low dds is better, but if you're high up dds becomes more dull


Just to debunk some myths:
Most .dds textures are 2048^2. The size they are mapped to varies according to features visible in them, -it ranges roughly from 500 to 1600- so you don't end up with a house the size of a stadium. The size increase is offset by the fact that they remain compressed in VideoRAM... as opposed to the .png ones. Also they don't need any processing at load time (no mipmap generation is necessary since that is already done).
So you get 4x the disk size for at least 16x the detail, and roughly the same memory used at runtime. It's for you to decide which is more efficient...

P.S.: @btw, they are one and the same ;)
i4dnf
Retired
 
Posts: 745
Joined: Wed Sep 09, 2009 7:17 am
Location: LRBS
Callsign: YR-I4D
Version: GIT
OS: Gentoo Linux ~amd64

Re: Help needed - market research for FG

Postby Thorsten » Mon Dec 17, 2012 1:08 pm

Just to debunk some myths:


Thanks for the details. Given that I tried to be honest and wrote I'm not sure how large the dds base size is... your phrasing seems unnecessarily harsh, but... I acknowledge your point - at 2048x2048, the disk space usage ratio is much better. I will edit my post accordingly.
Thorsten
 
Posts: 11191
Joined: Mon Nov 02, 2009 8:33 am

Re: Help needed - market research for FG

Postby Thorsten » Mon Dec 17, 2012 1:19 pm

I think you would be what I loosely refer to as the "Linux expert" (at least compared to me), so I understand your comments, my point is simply that most of us do not have the intimate knowledge and familiarity that you clearly do. I was relieved to find the following comrade with similar problems: http://www.mail-archive.com/flightgear- ... 38817.html


Yeah, it's a small planet, that comrade writing the post happens to be... me. If you read on, you can discover that I get in the following educated about build servers and other neat developments.

So I'm quite aware that the situation is far from optimal - nevertheless I think what you states is simply not true and a bit over the top. One needs to be neither expert nor needs to have the most recent Linux distribution in order to install FG under Linux. You do need to be fairly knowledgeable if you want the most recent FG on an old distribution - that situation isn't covered by the Linux release policy. But one might equally blame the folks at Ubuntu who could package it.
Thorsten
 
Posts: 11191
Joined: Mon Nov 02, 2009 8:33 am

Re: Help needed - market research for FG

Postby btw » Mon Dec 17, 2012 2:47 pm

Your points are fair Thorsten, but If you guys want to be serious about "market research" then although a comment like mine may be naive, you could consider me the typical idiot in the focus group you're trying to sell to, so maybe not a good idea to right off the focus group opinion to quick. Let me add that I think flightgear really is marvellous project and I have been an enthusiast for some time, I won't be switching to fsx.
btw
 
Posts: 63
Joined: Sun Nov 06, 2011 8:53 pm

Re: Help needed - market research for FG

Postby Hooray » Mon Dec 17, 2012 3:30 pm

Obviously, things are MUCH easier on Windows, where you don't have to support dozens of architectures, hundreds of different distributions and different versions (kernels). But that's a general Linux "issue", which is also why we get to see fewer viruses, trojan horses and worms successfully deployed for Linux.
Cross platform development is obvously a huge challenge and deploying portable software for Windows/Mac is much easier, because of the heterogenous nature of Linux deployments.

The linux build situation will improve automatically once "suse studio" is used to also streamline the release process across other distributions.
But obviously that requires people to actually set up the required scripts first, all the infrastructure is already there and provided for free.

We really don't generally disagree with your points, but to change something for the better, there's more needed than just "talking" and voicing "opinions". As you could see, Thorsten made very similar points on the devel list just a couple of weeks ago.
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: 11376
Joined: Tue Mar 25, 2008 8:40 am

Re: Help needed - market research for FG

Postby Thorsten » Mon Dec 17, 2012 4:13 pm

Your points are fair Thorsten, but If you guys want to be serious about "market research" then although a comment like mine may be naive, you could consider me the typical idiot in the focus group you're trying to sell to, so maybe not a good idea to right off the focus group opinion to quick.


You're assuming two things here which probably aren't true. One is the 'you guys' - there is no coherent front of 'us' here, this is an ongoing discussion out of which something may emerge or not. The second is 'market research' - if you read Stuart's initial post carefully, his intention is to find out in which areas FG as it currently is is superior to FSX or X-plane so that we can advertize these areas. Stuart is not asking users to turn this into (yet another) feature request story or asking users for their opinion how things should be.

As for my role, I'm trying to explain why things are they way they are, and I'm trying to separate the possible and useful from the impossible and useless feedback. "I think FG should have the user support of a commercial product but remain free". is a valid opinion, but the devel community is going to disregard it without further ado for obvious reasons. "I think more care should be taken to make the San Francisco area more spectacular as this is our showcase scenery and first impression for many." is a useful suggestion on which we can realistically follow up.

I speak only for myself, but personally I see no sense turning FG into something which 'sells' to a user-support oriented gaming community by for instance 'dumbing down' realism to make the user experience less frustrating for the beginner. Nor would I want that a large share of manpower gets diverted from coding new features to writing better documentation.

In general, it's quite easy to demand X or state that FG isn't possibly competitive without X, however on the devel side there are very limited resources to be allocated, and someone actually working on X means that Y doesn't get done. So you should rather not see this as a wishlist but as an either/or choice.
Thorsten
 
Posts: 11191
Joined: Mon Nov 02, 2009 8:33 am

Re: Help needed - market research for FG

Postby i4dnf » Mon Dec 17, 2012 4:21 pm

Thorsten wrote in Mon Dec 17, 2012 1:08 pm:
Just to debunk some myths:


Thanks for the details. Given that I tried to be honest and wrote I'm not sure how large the dds base size is... your phrasing seems unnecessarily harsh, but... I acknowledge your point - at 2048x2048, the disk space usage ratio is much better. I will edit my post accordingly.


Thanks for the edit. I was referring to a more general perception of the .dds textures, part of which is represented in the quote.

BTW: [in the Canaima National Park] a fair visual comparison would be after you've applied a similar regional texturing scheme to that area.
If there would be a forest mapped for the Everest area, I don't think it's reasonable to expect it to show rock/glacier by being prescient (although by using materials-dds you'd get a high chance of seeing that ;) due to other things).
Worst case corner cases are a dime a dozen, can be found everywhere, and in every texturing/shading scheme ;), and are a consequence of the low quality/resolution of the vmap0 data.
i4dnf
Retired
 
Posts: 745
Joined: Wed Sep 09, 2009 7:17 am
Location: LRBS
Callsign: YR-I4D
Version: GIT
OS: Gentoo Linux ~amd64

Re: Help needed - market research for FG

Postby Bjoern » Mon Dec 17, 2012 4:37 pm

Actually taking a look at the developer portal, I realize that my envisioned "short term goals" article would look something like a mixture between the "Project Infrastructure Enhancement" article (http://wiki.flightgear.org/Project_Infr ... hancements) and the "Coding Help Needed!" box on the portal site (http://wiki.flightgear.org/Portal:Developer).

Also, once the basic article is set up, I'd shoot the developer mailing list a note about it (I take everyone more or less involved in the project is using it) and see how it turns out.

Let's see when I get around to it.



Hooray wrote in Mon Dec 17, 2012 3:30 pm:Cross platform development is obvously a huge challenge and deploying portable software for Windows/Mac is much easier, because of the heterogenous nature of Linux deployments.


Must...suppress...voicing...new...idea...for...distribution...nnnnnnnngh!

WhatAboutABashScriptForDownloadingTheCurrentFGReleaseToEliminateThe WaitingPeriodUntilItPopsUpInTheDistribution'sPackageManagerAndToUnifyFG'sDistributionMethodForLinuxBuilds?

Ah, much better...
Bjoern
 
Posts: 386
Joined: Fri Jan 06, 2012 10:00 pm
Location: TXL or so
Version: Next
OS: ArchLinux, Win 10

Re: Help needed - market research for FG

Postby Thorsten » Mon Dec 17, 2012 4:47 pm

BTW: [in the Canaima National Park] a fair visual comparison would be after you've applied a similar regional texturing scheme to that area.


Yes and no. The point I was trying to make is that a 'dull' low contrast scheme (as the default texture scheme is) has an edge where landcover data is bad (unfortunately in many areas) because it never shows strong contrasts at landcover boundaries. I think for illustrating the problems with bad landcover and a high contrast scheme the Canaima shot is okay. I intentionally did not show the regionalized comparison here since this is unfair, but I can make a Canaima shot in the default scheme and edit it in if you feel that's needed to appreciate the point.

As a side note, my wrong perception of dds file sizes comes from an afternoon converting cloud textures to dds to improve loading time back when building a weather tile took a minute. The dds sheet I got did load considerably faster than png but had a larger file size. However, an rgb texture sheet loaded fastest, so I did not pursue dds further. I distinctly remember png to have the smallest file size, so clearly I must have somehow left the compression out of the dds generation process (although I tried many settings of the conversion program), I never reached your size-to-texture data ratio.
Thorsten
 
Posts: 11191
Joined: Mon Nov 02, 2009 8:33 am

Re: Help needed - market research for FG

Postby Hooray » Mon Dec 17, 2012 4:49 pm

Also, once the basic article is set up, I'd shoot the developer mailing list a note about it (I take everyone more or less involved in the project is using it) and see how it turns out.


please keep in mind that all contributors are currently busy preparing the upcoming release, so don't take lack of feedback as a general lack of interest.
It's just a particularly bad time to ask for additional involvement, because people are already very busy.

WhatAboutABashScriptForDownloadingTheCurrentFGReleaseToEliminateThe WaitingPeriodUntilItPopsUpInTheDistribution'sPackageManagerAndToUnifyFG'sDistributionMethodForLinuxBuilds?

Ah, much better...


We do have that already for debian-based distributions via the "download_and_compile.sh" script, contributed and maintained by brisa: http://wiki.flightgear.org/Scripted_Com ... ian/Ubuntu
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: 11376
Joined: Tue Mar 25, 2008 8:40 am

Re: Help needed - market research for FG

Postby Bjoern » Tue Dec 18, 2012 1:12 pm

Hooray wrote in Mon Dec 17, 2012 4:49 pm:please keep in mind that all contributors are currently busy preparing the upcoming release, so don't take lack of feedback as a general lack of interest.
It's just a particularly bad time to ask for additional involvement, because people are already very busy.


Excellent, I'll wait for the 2.9 release plus a week or so then. Didn't want to do it now anyway.

We do have that already for debian-based distributions via the "download_and_compile.sh" script, contributed and maintained by brisa: http://wiki.flightgear.org/Scripted_Com ... ian/Ubuntu


Yeah, something like this, but for more distributions.
All that'd be needed would be the respective package manager commands and the package names of the dependencies.
Bjoern
 
Posts: 386
Joined: Fri Jan 06, 2012 10:00 pm
Location: TXL or so
Version: Next
OS: ArchLinux, Win 10

Re: Help needed - market research for FG

Postby EigerSA » Sat Jan 05, 2013 1:04 pm

I recently bought FSX from our local computer store, only because it was bundled with another 2 titles at an extremely low price. After loading it onto my windows laptop and spending some time flying about I have the following comment.

Right out the box, FSX is far easier to get started going than FG. When I originally installed FG, it took me a little while to get the hang of things, how to load aircraft and scenery, how to change the default controller setup to better suit me. With FSX, it was so much easier than expected! Everything was right there in the loader and was also pretty self-explanatory.

When it comes to the default scenery, I think FG wins hands down, no comparison at all! Although once you start adding the various "add-on's", FG definitely suffers quite drastically. One thing I did appreciate is that at every airport I flew to, FSX had a default airport of some sort, which made it pretty interesting even if it wasn't accurate.

Interestingly, I found for the most part that FG cockpits were far more interesting and advanced out the box than FSX. Again, provided you're busy to part with a cash (though there are free models about), you can get amazingly detailed cockpits in FSX. One thing I did like was when you hovered over an instrument, a small box popped came up with the name of the instrument and the value, really handy when sometimes a "rev-counter" for example is too small to read.

I can definitely see the appeal of FSX over FG, on start up all the manuals you need are right there, there's no posting on forums and waiting in vain for help. There's even a really fantastic tutorial section based (according to the box) on real life flight school lessons.

I will say that FG will remain my favorite, but perhaps only because I'm used to it(?), though I will fly FSX as part of my local virtual airline.

On the whole I think that FG really is one of the better simulators "out-of-the-box" but there definitely needs to be more work done on the "visual" (eye-candy) part of the simulator to keep people involved in the long term.
EigerSA
 
Posts: 88
Joined: Thu Sep 08, 2011 6:26 pm
Location: Durban, South Africa
Callsign: EigerSA
OS: Vista

Re: Help needed - market research for FG

Postby Thorsten » Sun Jan 06, 2013 6:19 pm

Although once you start adding the various "add-on's", FG definitely suffers quite drastically.


Until you start installing the FG 'addons' (custom scenery, current devel features,...).

I can definitely see the appeal of FSX over FG, on start up all the manuals you need are right there, there's no posting on forums and waiting in vain for help. There's even a really fantastic tutorial section based (according to the box) on real life flight school lessons.


FG has a working tutorial system, and I think lots of stuff is implemented for the c172p (never was a fan of tutorials myself). If you think something is missing there, feel free to add.

I am more of an old-style person, believing in manuals and written text (Falcon 4.0 was a terrific simulator - just without reading the 350 page manual, you would never figure out how to operate the radar of the F-16 properly, or what all these bullseye position calls meant, or just how to direct the wingmen into a meaningful attack pattern,...). In your book, that wouldn't have been user-friendly, but beyond some complexity a tutorial just doesn't make any sense).

On the whole I think that FG really is one of the better simulators "out-of-the-box" but there definitely needs to be more work done on the "visual" (eye-candy) part of the simulator to keep people involved in the long term.


Seems we're already better as far as weather, light and sky is concerned. I would wager a bet that the new procedural texturing techniques also beat FSX for close-up visuals of the terrain. In a fair comparison against our hires, regionally textured scenery, I don't think we're too bad. Of course, you can compare scenery somewhere in Indonesia which was created with really poor geodata against FSX (which probably doesn't have any scenery in Indonesia out of the box) Hawaii addon or so, and then that's not fair. Also, photo-scenery is nice if you view it at the time, season and conditions it was taken - however doesn't compare so well if conditions are different. So we could make an autum comparison in which we'd find that FSX probably can't do anything at all, but we can do autumn colors.

So it's not exactly clear to me what you actually want.
Thorsten
 
Posts: 11191
Joined: Mon Nov 02, 2009 8:33 am

Re: Help needed - market research for FG

Postby OldFlyGuy » Mon Jan 07, 2013 12:16 am

As a retired career pilot, active GA pilot, avionics engineer and one interested in simulators strictly for what they provide insofar as maintenance of flying skills, instruction and situational experience, FG wins "hands down".

Why - it all comes down to one thing - PERFORMANCE.

People interested in a simulator for the purposes that I am, as opposed to gaming, hobby development or other possible motives, seek to enter an experience (you could call correctly call it a fantasy) wherein the simulation is sufficiently faithful to the real world experience that the awareness of it NOT being real fades sufficiently into the background of consciousness to present a valid experience to the user.

Thus, anything that jerks the awareness that it is a simulation back to conscious thought is an impedement to a sucessful experience. Things like visual pauses due to graphic card or computer resource demands, unrealistic controls (keyboard only controls, for example), simulated things that react differently than in the real word, etc.

FG delivers a smooth, seamless simulation basis with far fewer demands on the resources that cause the sorts of breaks in the simulation that are described above. Further, the add-ons for FG are written to a common and much narrower standard of performance than those in X-Plane and FSX, both of which are into a lot of "gimmicks" for game play. {FSX even refers to itself as a game.) The result is performance with fewer breaks in the experience, and with predictable (to a trained pilot) outcomes. In maintaing the "fantasy" this is far, far more important than the resolution of the scenery, simulation of surface traffic, or generation of air traffic. If FG is weaker in any area related to this type of operation, it is the same area that the others struggle with - that of a realistic single session ATC. Generated ATC simulations that simply vector someone around the sky or occasionally change altitudes do not properly present the increasingly complex ATC environment. An ATC simulation needs speed restrictions, climb restrictions, holds, route deviations for weather, modified clearances and many other "wildcard" instructions that teach flexibility in coping with a changing ATC environment while still flying the aircraft. Simulation developers often become absorbed in the reasons for these deviations (terrain, traffic, weather, etc) and introduce them into the simulation, but fail to introduce the consequential and essential changes to the flight environment that are a result of the (terrain, traffic, weather, etc).

Just my 2 cents,
RFB
Military Command Pilot (Ret.)
ATP (Ret.)
GA Pilot (Active)
AOPA, ALPA (inactive), AEA, EAA
Avionics Developer (Real World, Ret.)
OldFlyGuy
 
Posts: 11
Joined: Sun Dec 23, 2012 10:37 pm
Location: Bend, OR, USA
Callsign: N6779N
Version: 2.8
OS: Win 7 (64, OS-X

Re: Help needed - market research for FG

Postby Bjoern » Mon Jan 07, 2013 2:55 pm

Thorsten wrote in Sun Jan 06, 2013 6:19 pm:(Falcon 4.0 was a terrific simulator - just without reading the 350 page manual, you would never figure out how to operate the radar of the F-16 properly, or what all these bullseye position calls meant, or just how to direct the wingmen into a meaningful attack pattern,...).


If F4 hadn't crashed after every odd mission, it would have been the best simulator ever released.

The manual was awesome beyond everything and I am still proud of owning it in its printed form.

(The paper map also included in the box decorated my room for a few years. Talk about a weirdly obsessed 14-year old back in the day...)


I would wager a bet that the new procedural texturing techniques also beat FSX for close-up visuals of the terrain.


Actually, FSX just uses a noise map to give an idea of grass and/or dirt when close to the ground.
You've definately got them beat now.

Also, photo-scenery is nice if you view it at the time, season and conditions it was taken - however doesn't compare so well if conditions are different. So we could make an autum comparison in which we'd find that FSX probably can't do anything at all, but we can do autumn colors.


If you're talking about procedural autumn coloring on phototextures then yes, FSX can't do that.




OldFlyGuy wrote in Mon Jan 07, 2013 12:16 am:An ATC simulation needs speed restrictions, climb restrictions, holds, route deviations for weather, modified clearances and many other "wildcard" instructions that teach flexibility in coping with a changing ATC environment while still flying the aircraft.


There's an ATC add-on for MSFS that improves the airplane and traffic handling in that simgame tenfold (Radar Contact). It imposes speed and climb restrictions on the pilot, implements holding patterns and supports SID and STAR, weather deviations, emergencies, alternates, etc...
Granted, it's (maybe) still a far cry from real ATC operations, but it's close enough for a home user flight simulator.
Bjoern
 
Posts: 386
Joined: Fri Jan 06, 2012 10:00 pm
Location: TXL or so
Version: Next
OS: ArchLinux, Win 10

PreviousNext

Return to Hangar talk

Who is online

Users browsing this forum: No registered users and 1 guest