YASim vs. JSBSim

[helijah]

Split from B-17 issues.

Hi,

At the same time, now I refuse categorically FDM JSBSim on my models. Sorry but as JSBSim bring more problems than solutions, I would refuse.

Place a plane on the back or on the water are unacceptable for me. A flight simulator must not only simulate the aircraft in the sky but also on the ground (with the differences between sand, soil, bitumen, etc. ..) and many other thing that Yasim done rather well.

Example : Landing with the C172 on the water does not give a good impression of FG.

Maybe in the future, but for now, JSBSim is not the solution suitable for a flight simulator in the quality of FG.

Best regards. Emmanuel

P.S. This does not alter the fantastic work that is done around JSBSim. But frankly I think it is better suited to games (like IL2, or CSP). YASim is what closest to what a flight simulator should have. (for me)

[simbabeat]

My understanding is that jsbsim and yasim can both be very realistic. Just depends what you do with it.

[jonsberndt]

Hi,

At the same time, now I refuse categorically FDM JSBSim on my models. Sorry but as JSBSim bring more problems than solutions, I would refuse.

Place a plane on the back or on the water are unacceptable for me. A flight simulator must not only simulate the aircraft in the sky but also on the ground (with the differences between sand, soil, bitumen, etc. ..) and many other thing that Yasim done rather well.

Example : Landing with the C172 on the water does not give a good impression of FG.

Maybe in the future, but for now, JSBSim is not the solution suitable for a flight simulator in the quality of FG.

Best regards. Emmanuel

P.S. This does not alter the fantastic work that is done around JSBSim. But frankly I think it is better suited to games (like IL2, or CSP). YASim is what closest to what a flight simulator should have. (for me)

Well. It's hard to know where to begin in responding to this one. For both YASim (which I have heard very good things about, but not used, myself) and JSBSim, the quality of the aircraft model depends on the modeler. In other words, "garbage in, garbage out." If you had read about JSBSim, you would know that JSBSim can model anything from a ball, to a runway-landing aircraft, to an amphibious aircraft, to a multi-stage launch vehicle climbing into orbit, and orbiting the planet. You would know that JSBSim has been used for hardware in the loop testing for a UAV autopilot. You would know that JSBSim is the flight model for OpenEaagles, and Outerra, as well as FlightGear. You would know that JSBSim has been used as a tool for researchers, mentioned in theses, and referenced in technical papers - most recently in an AIAA paper entitled, "A Simulation-based Design Model for Analysis and Optimization of Multi-State Aircraft Performance," co-authored by researchers from Draper Labs and MIT.

JSBSim has grown to be a seriously capable flight dynamics model, made by people who are serious about modeling flight dynamics. Just like FlightGear.

Jon

[helijah]

Hey Jon.

the answer to say: "Machin uses JSBSim, other uses JSBSim etc. .." is not an answer. NASA could use JSBSim it would not change anything. I do not belong to the people who follow blindly and foolishly. I Experimented, I trying, and I created my own conclusion.

JSBSim is a fantastic tool. I am the first to admit. Unfortunately, in my view, it brings more problems than solutions in an application as FG. And I do not mean that I do not know. I used JSBSim 2 years before opting for Yasim.

I am convinced that the use of tables is a much more precise, it's true, but currently in FG, this is an inappropriate method. For the moment.....just for the moment.

Best regards. Emmanuel

[jonsberndt]

Hey Jon.

the answer to say: "Machin uses JSBSim, other uses JSBSim etc. .." is not an answer. NASA could use JSBSim it would not change anything. I do not belong to the people who follow blindly and foolishly. I Experimented, I trying, and I created my own conclusion.

JSBSim is a fantastic tool. I am the first to admit. Unfortunately, in my view, it brings more problems than solutions in an application as FG. And I do not mean that I do not know. I used JSBSim 2 years before opting for Yasim.

I am convinced that the use of tables is a much more precise, it's true, but currently in FG, this is an inappropriate method. For the moment.....just for the moment.

Best regards. Emmanuel

Then please list the reasons why you believe it is an inappropriate method. I'll add that JSBSim does not merely use table lookups. You can specify any arbitrary function, value, or table, and any combination of those. Creating flight models is not easy. If you don't want to expend the effort, you are of course free to use whichever FDM you want - or to write your own.

Jon

[ZzerozZ]

Well, I am very new to the whole idea of flight simulators and the unfolding discussion here is teaching me a lot about the differences. Thanks guys.

This forum is a wonderful place to be able to interact with the creators of this fantastic thing, as well as with those people who keep it alive with their determined dedication. And one of the things that makes it so good (for me) is that, just as in "real life", there appear to be many ways to achieve something.

It is great to think that within this thing we call FlightGear, we can find examples of leading technology at use. It is also great to think that de Havilland made their 'Mosquito' from wood, while the world around them used metal. They say 'variation is the spice of life'. - And surely it is these differences which drive the creative force to make something successful.

I offer my most sincere thanks to those few heroic, creative people; who make FlightGear such a pleasure to be involved with, for so many unthanking people who take it all for granted; and that different ideas never become a barrier to the common goal.

[helijah]

Then please list the reasons why you believe it is an inappropriate method. I'll add that JSBSim does not merely use table lookups. You can specify any arbitrary function, value, or table, and any combination of those. Creating flight models is not easy. If you don't want to expend the effort, you are of course free to use whichever FDM you want - or to write your own.

Jon

Oh no! Just read my various writing here and on IRC. I will not repeat again.

For cons, I repeat, this does not diminish the high quality of work done on JSBSim.

[Timaman]

It is great to think that within this thing we call FlightGear, we can find examples of leading technology at use. It is also great to think that de Havilland made their 'Mosquito' from wood, while the world around them used metal. They say 'variation is the spice of life'. - And surely it is these differences which drive the creative force to make something successful.

I'll drink to that!

[jonsberndt]

Then please list the reasons why you believe it is an inappropriate method. I'll add that JSBSim does not merely use table lookups. You can specify any arbitrary function, value, or table, and any combination of those. Creating flight models is not easy. If you don't want to expend the effort, you are of course free to use whichever FDM you want - or to write your own.

Jon

Oh no! Just read my various writing here and on IRC. I will not repeat again.

For cons, I repeat, this does not diminish the high quality of work done on JSBSim.

I understand, but I would like some specifics on what you feel is not appropraite for FlightGear at this time. Of course, I already searched your posts on this forum and found nothing. Normally, when someone finds something amiss with JSBSim, they will send me a note, or post it to the developers list, or file a bug report. That really helps to improve the usability for all of us. I'm asking, because I'm sure your comments will be useful in improving JSBSim for use within FlightGear.

Thanks,

Jon

[simbabeat]

He does work for JSBSim... So maybe he would like to know to make it better for us...

[biffbutkus1]

Hi,

At the same time, now I refuse categorically FDM JSBSim on my models. Sorry but as JSBSim bring more problems than solutions, I would refuse.

Place a plane on the back or on the water are unacceptable for me. A flight simulator must not only simulate the aircraft in the sky but also on the ground (with the differences between sand, soil, bitumen, etc. ..) and many other thing that Yasim done rather well.

Example : Landing with the C172 on the water does not give a good impression of FG.

Maybe in the future, but for now, JSBSim is not the solution suitable for a flight simulator in the quality of FG.

Best regards. Emmanuel

P.S. This does not alter the fantastic work that is done around JSBSim. But frankly I think it is better suited to games (like IL2, or CSP). YASim is what closest to what a flight simulator should have. (for me)

Don't blame the FDM when the blame lies with the modeler or the data. As Jon said Garbage In...Garbage Out. Flight models such as JSBSim require ACCURATE data that is correctly put into the flight model data file....this is true of any FDM that uses actual wind-tunnel and flight test data. It is a painstaking process, but when done correctly it results in what are probably the most accurate flight dynamic models. FDMs like JSBSim are the type used in commercial, fixed-wing FAA level-D sims (the Level-D sims that I have worked on all use coefficient buildup for the FDM, just as JSBSim does.) While blade-element theory type models like X-Plane are used in commercial sims, I have only heard of them used on helicopter sims.

IMHO its FDMs like YASim and XPlane that, though good, are more suitable for games rather than commercial sims- they are good when we don't have access to the real aerodynamic data, but currently do not have the accuracy of an FDM that uses manufacturer provided aerodynamic data.

[Thorsten]

Actually, is there a document (or several) which outlines the internal workings of JSBSim vs. YaSim? I'd be rather interested in some insight what's behind the flight experience, and I find the description in the Wiki rather short.

[ZzerozZ]

This whole discussion is fascinating to me. I am pleased there are so many points of view being presented. And I would also be particularly interested to know which of the aircraft currently available, make best use of JSBSim. (Obviously I am already aware of the default Cessna). But if JSBSim requires such accuracy, there must be some pretty good models 'out there' which I have not yet discovered. I confess to have been working on a livery, which has taken my FG time recently, leaving me no time to investigate other models. So, please folks: I would be grateful to be pointed in the direction of those aircraft which are the best examples of JSBSim.

Also, I feel that I need to make another important comment about this discussion: I hope that Emmanuel does not feel that people are against him. He has provided FlightGear with so many aircraft and has considerably increased the pleasure of the experience for us all. And clearly there are many, very good examples of his preference for YASim, being put to very good use. In fact, when looking through the list of available aircraft, I get the impression that he has done as much work to make FlightGear a success, as just about any other person.

Please keep up the good work Emmanuel. I would soon like to show you my work on one of your aircraft; the B-25 Mitchell.
And for anyone who can help me find the best examples of JSBSim: I'm ready and waiting for some suggestions!

[helijah]

I understand, but I would like some specifics on what you feel is not appropraite for FlightGear at this time. Of course, I already searched your posts on this forum and found nothing. Normally, when someone finds something amiss with JSBSim, they will send me a note, or post it to the developers list, or file a bug report. That really helps to improve the usability for all of us. I'm asking, because I'm sure your comments will be useful in improving JSBSim for use within FlightGear.

Thanks,

Jon

I'm sorry Jon, but my English is too bad that I can properly express my point of view. So I'll do it in French. For you to find someone to translate. After all, why am I the only one still have to translate what I do not understand.

- C'est le principe même de JSBSim qui est rédhibitoire pour moi. Le simple fait d'utiliser des mesures effectuées dans le réel est une absurdité à mon sens. Car aucune mesure ne sera jamais correcte et le croire c'est croire à l'impossible. Quitte à avoir un système approximatif je préfère, et de très loin, un système qui utilise les valeurs dimensionnelles de l'avion pour résoudre ensuite le comportement. De ce point de vue, YASim se rapproche beaucoup de X Plane (même si beaucoup de travail reste à faire) et JSBSim lui se rapproche de FS. Et bien, je suis désolé, mais entre X Plane et FS il n'y a pas de discussion à avoir. Seul X Plane dispose d'un modèle physique digne de se nom.

Mais ceci est un détail. Un simulateur de vol tel que FG se doit d'être global. Le vol d'un avion aussi parfait soit il n'est pas le seul but à atteindre. Les interactions avec le sol par exemple sont tout aussi importantes. Tout les avions JSBSim actuels peuvent se poser sur l'eau, ou sur le dos. Voir même, dans les pires des cas traverser le sol pour continuer à voler sous la terre. Faire la différence entre le sable, l'herbe, la terre, le bitume etc... sont des aspects OBLIGATOIRES dans un simulateur tel que FG. Car les avions ne sont pas tout le temps dans le ciel en train de voler dans des conditions idéales.

Même si je reconnais bien volontiers que le travail réalisé sur JSBSim est tout simplement fantastique. Je trouve simplement dommage de voir autant de connaissances et de compétences perdues pour un projet annexe alors que, tous autant que vous êtes, vous auriez pu apporter à YASim un savoir et des améliorations qui aurait pu faire passer X Plane pour un simulateur amateur. Au lieu de cela vous préférez perdre votre temps à travailler sur un système qui est plus proche de l'émulation que de la simulation. Alors de deux choses l'une. Ou vous n'avez pas assez réfléchi au départ du projet, ou vous avez l'intention de faire de l'argent avec dans un futur plus ou moins proche. Dans les deux cas, cela sera, au final, néfaste pour FG.

Et l'une des plus belles illustrations de tout cela reste la volonté de certains d'entre vous à utiliser YASim pour générer les tables qui serviront à créer des FDM JSBSim ensuite. Ne voyez vous pas là une absurdité totale. Si le FDM YASIm est assez correcte pour cela alors le FDM JSBSim n'a aucune utilité car le FDM YASim se suffira à lui même.

Oh bien sur, je ne me leurre pas. Depuis que j'ai commencé des avions dans FG, toutes les idées que j'ai pu avoir, les sensations qui m'ont envahies, n'ont jamais été prise en compte. Même si au finale dans 100% des cas la route prise fut bien celle que je pressentait. C'est un problème qui me poursuit depuis l'école . j'ai toujours réussi à ressentir les choses et savoir que j'avais raison. Mais l'expliquer est une autre histoire. Alors, en général, j'attends. Le temps fait son oeuvre et me donne toujours raison Et là c'est mon ego sur dimensionné qui parle diront certains. Mais que puis je y faire. C'est plus une tare qu'un avantage d'avoir
raison sans vraiment pouvoir l'expliquer.

Alors je vais continuer à faire mes petits avions dans mon coin et attendre....

Amicalement Emmanuel

[f-ojac]

I will - very quickly and dirty - translate (I translate this for you to understand the core of it, not the eventual flames around :

.../...

JSBSim's principle in itself is a pain in the neck for me. The only fact to use data coming from the real life is - IMHO - absurd. Because no data will ever be correct and to believe otherwise is nonsense. So, when the choice comes to prefer an approximative system I prefer - from far - a system using the true dimensions from the aircraft and which then calculates its behavior. From this point of view, YASim is far much closer to XPlane (even if a lot of work still remains) and JSBSim is closer to FS. So, I'm sorry, between XPlane and FS there is no discussion to have. Only XPlane owns a physical model worth of it.

But this is just a detail. A flight simulator like FG has to be global. An aircraft flight as perfect as it can be is not the only goal to achieve. The interactions with the ground, for instance, are also important. All JSBSim aircraft can land on water or even upside down. Or even cross the land to continue to fly under the land... To be able to make the difference between sand, herb, ground, asphalt are mandatory aspects in a simulator like FG. Because aircrafts are not always flying in the air in ideal conditions.

Even if I do recognize that the work done on JSBSim is truly amazing, I found it sad to see so much knowledge and skills lost for another project while you should have all given to YaSim a lot of enhancements that could have made XPlane look like an amateur sim. Instead of this, you prefer to spend time on a system which is, in fact, closer to emulation than simulation. So, or you haven't been thinking enough at the start of the project, or you have the intention to make money in a close future. In both cases this will be harmful for FG.

One of the best examples of all this is the willing of some of you to use YASim to generate tables which will then be used to create JSBSim FDMs later. Don't think this is completely absurd. If the YASim FDM is correct enough for this, then the JSBSim FDM is useless because it means the YASim FDM is sufficient to itself.

(...)

So I will go on and continue my little aircrafts in my own world and wait...